ABSTRACT. Background and rationale: Many high-technology and high-growth industries, including clean technology industries, have seen major geographical shifts in which companies expand or move manufacturing or R&D operations along their supply chain to new countries. These manufacturing shifts raise issues that are important for public policy and economic competitiveness, evident in recent US efforts to spur and incentivize local manufacturing and the related trade wars with China. Policy makers and academics alike are interested in why firms alter their geographic strategies and whether doing so depends on the inherent characteristics of the technologies they manufacture, such as the ‘technology complexity’ (which can affect the skills or costs associated with manufacturing). The relationship between the internationalization of manufacturing along the supply chain and technology complexity is particularly important for clean energy technologies, where government policy still plays an important role and often tries to simultaneously enhance economic competitiveness and technology innovation to meet climate and broader sustainable development goals.

This paper focuses on the global value chain of wind energy to explore the relationship between internationalization of the suppliers of different components and the characteristics of component technologies. Over the past decade, demand, manufacturing, and R&D activities related to wind turbines and their components have increased their geographic scope from Europe and the US to include emerging economies, notably China and India. Additional expansion of these activities to new countries is expected given ambitious policy targets for deployment in latecomer countries for wind, such as Egypt and Argentina. Extant research on innovation in the wind industry has primarily studied one part of the value chain: the original equipment manufacturers (OEMs)—i.e., the 10-15 manufacturers that assemble wind turbines. Little attention, however, has been paid to the evolution and emergence of hundreds of specialized firms that supply components to OEMs. The components of a wind turbine (e.g., blades, generators, gearboxes, towers, bearings) span a range of competences, internal sub-components, and costs. The ability of suppliers to manufacture certain components in new locations could depend on their complexity, which we define as the knowledge intensity or the difficulty of manufacturing individual components. By leaving out the component suppliers—which are often small and medium enterprises (SMEs) from Europe, the US, China, and more recently other countries—the literature on the internationalization and offshoring of manufacturing provides a limited understanding of the drivers of siting decisions. In addition, it neglects a critical entry point for developing and emerging economies to develop competences and become competitive in global markets for high technology industries. 

Methods: The paper relies on novel data of the wind supply chain at the firm level parsed from industry reports [1]. We develop and analyze this new dataset of 336 component manufacturing firms manufacturing up to 9 major components such as towers, blades, or generators that supply to 19 major OEMs (2006 to 2017). Our dataset details the OEM strategy used to procure each component (i.e., in-house or outsource), over 3,400 supplier-OEM relationships with granularity on the component-level, location of suppliers (i.e., home location and manufacturing location), patenting activity of each supplier (international and home-country), firm characteristics for each of the component suppliers and OEMs (size, founding year), etc. We also obtained the information on international vs. home-country activities at the firm level from industry reports [1], Our main independent variable is technology complexity, proxied using the product complexity index (PCI) at the component level [2]. The PCI captures the knowledge intensity of a product based on the knowledge intensity of its exporters, which the literature assumes is related to the countries that export a certain product and the other products that those countries export. We estimate the PCI by mapping each component of the wind turbine with standardized Harmonized System (HS) codes that classify traded products. We also calculated a network complexity variable as an alternative metric for complexity to serve as a robustness for the PCI one using natural language processing of patent text data to assess the interconnectedness of components [3]. Our main dependent variable is the shift in ‘internationalization’ activities of the suppliers, measured by the annual increase (or decrease) in international activities of the respective supplier. 

We then test whether the internationalization of suppliers was influenced by the complexity of components to understand different strategies for location. We also test whether the internationalization of manufacturing was influenced by location of patenting in supplier firms (home country vs. international patenting strategy). We include several control variables such as the age of the supplier (time difference to founding year), the size of the supplier (number of employees), the diversification of the supplier (dependence on wind vs. operation in more sectors) and the current manufacturing location of the supplier. 

Initial results: Our initial analysis of the internationalization of suppliers and their relationship with technology complexity offers three main findings: (1) The value chain for wind energy is globally distributed and the manufacturing locations have increasingly diversified in the time period of our analysis (2006-2017). Our descriptive results suggest that, over time, component manufacturers that were originally from developed countries (e.g. Germany or the US) expanded to new manufacturing locations in emerging economies (most notably China, but also India, Mexico, etc.), generally without closing down their home country manufacturing operations. In contrast, new suppliers appeared only in these emerging economies. (2) The mapping of components to the PCI shows that wind components vary by technology complexity – for example blades are high complexity products with a higher PCI vs. towers that have low complexity and PCI. (3) The results from the econometric analysis indicate a negative relationship between technology complexity (measured by the PCI) in the wind supply chain and the internationalization of manufacturing. For example, we find that the suppliers of high complexity components such as blades had lower shifts in internationalization than towers.

Significance: Our findings are among the first to present a detailed data-driven analysis of the GVC in a growing industry. These insights can shape public policy decisions related to attempts to meet climate goals and to drive ‘green growth’. This research suggests that local suppliers continue to work with home-country OEMs for manufacturing more complex components and that retaining manufacturing of lower complexity components in first mover countries may be dependent on additional factors or require additional policy or market drivers. Countries that are not home to OEMs, often developing countries, could in some cases develop domestic manufacturing in low complexity components with smaller policy interventions related to training, R&D and procurement. Upgrading GVC participation more quickly and developing SMEs for higher complexity components may require more efforts and time. 

[1] Navigant Research. Supply Chain Assessment Reports (2006-2014)

[2] Hausmann, Ricardo, César A. Hidalgo, Sebastián Bustos, Michele Coscia, Alexander Simoes, and Muhammed A. Yildirim. The atlas of economic complexity: Mapping paths to prosperity. MIT Press, 2014.

[3] Broekel, Tom. "Using structural diversity to measure the complexity of technologies." PloS one 14.5 (2019): e0216856.

ABSTRACT. PROJECT SUMMARY

The objective of this project is to determine strategic and policy ramifications of state-capitalist investments in U.S. shale for technology, companies, governments, the natural environment and society. China is currently the largest foreign investor in the U.S. shale-gas sector.

The central hypothesis is that China’s investments in the U.S. shale-gas sector will lead to technology development to suit that country’s unique institutional and geological characteristics affecting variety and selection of technology (Metcalfe, 1994). We further hypothesize that subsequent shale-gas-production scale will alter technology trajectories and investments in U.S. shale gas with strategic, environmental and national-security repercussions.

We propose: Aim 1: To determine the effects of China’s investments in the U.S. shale-gas sector on technology development. We will focus on environmentally-friendly production. Aim 2: To determine the effects of China’s investments on U.S. small and medium-sized enterprises that have historically pioneered exploration and development of new energy resources and technology. SMEs also lead exports and job creation. Aim 3: To determine the effects of Chinese subsidies and investments on U.S. national-security goals including energy self-sufficiency.

Institutions shape technology development (Mokyr, 2005). Policy locus matters as domestic and foreign demands shape technology development (Peters et al., 2012). Historically, public interest permitted the U.S. federal government to shape shale R&D. When governmental stakes in technology evolution are not recognized, its ability to fund R&D efficiently faces constraints: government investments and contracts have influenced companies’ profitability and survival through defining technological targets according to its own criteria; allowing governments to develop expertise and metrics through producing technologies; and legitimizing for publics governmental attempts to stimulate and to guide technology evolution (Nelson, 1983).

Our research uses quantitative and qualitative data and theoretical sampling including: a) a survey of businesses operating in the shale-gas sector b) patent data on innovations in the shale-gas sector c) personal interviews and site visits to 6 shale-gas companies in the continental United States with Chinese and U.S. investments and, d) analysis of U.S. environmental regulations and responses through innovations from U.S. shale-gas corporations. The methodology incorporates complex interactions between human and natural systems in China and the United States, where most of the technology-development is taking place.

Intellectual Merits: Our project contributes to the inter-disciplinary literature on technology-development policies and metrics, government investments, environmental effects and social returns (Small et al., 2014). Some research has reported on state-sponsored investments in technology, e.g., DARPA’s investments in basic research on energy and materials technology. Other research has shown how China’s investments altered technology-development trajectories in the global, solar PV sector. Yet, no research has looked at how state-capitalist investment alters selection and varieties of technology with social and environmental effects. China’s state-capitalist regimes favor industrial growth, expanding production, and subsidizing key industries for immediate returns rather than economic efficiencies or cutting-edge technologies (Haley & Haley, 2013). What remains unknown, are the long-term effects on natural and human systems of China’s state capitalism monopolistically championing state-sponsored and developed technology in strategic U.S. sectors.

Broader Impacts: At the conclusion of the proposed research, it is our expectation that we will have developed complex measures of how state-capitalist investments are shaping global technology development, natural and human systems. Preliminary findings support that Chinese state-backed investors will focus on immediate, subsidized production with established technology, thereby altering present trajectories of technology development essential to U.S. national security, trade, employment and environmental protection.

This project is being funded by National Science Foundation Grant (SciSIP) #1661733 & #1911289

ABSTRACT. We separate and directly measure the labor-demand effects of two simultaneous forms of technological change—automation and parts consolidation. We collect detailed shop-floor data from four semiconductor firms with different levels of automation and parts consolidation. For each process step, we collect task data and measure operator skill requirements, including operations and control, near vision, and dexterity requirements using the O*NET survey instrument. We then use an engineering process model to separate the effects of the distinct technological changes on these process tasks and operator skill requirements. Within an occupation we show that aggregate measures of technological change can mask the opposing skill biases of multiple simultaneous technological changes. In our empirical context, automation polarizes skill demand as routine, codifiable tasks requiring low and medium skills are executed by machines instead of humans, while the remaining and newly created human tasks tend to require low and high skills. Parts consolidation converges skill demand as formerly divisible low and high skill tasks are transformed into a single indivisible task with medium skill requirements and higher cost of failure. We propose a new theory for the differential labor effects of technological changes on tasks, and hence jobs. Understanding these differential effects of technologies on labor outcomes is a critical first step toward analyzing the impact of emerging technological changes on labor demand, and eventually markets.

ABSTRACT. Background: For decades, scientists have investigated the possible causes of and solutions to the so-called ``productivity puzzle," the phenomenon that female scientists tend to produce fewer papers on average than male scientists. Early work observed that adjusting for age, doctoral prestige and institutional prestige alone did not explain the difference. Subsequent studies increased the number of covariates considered, but the puzzle remains unsolved insofar as we still don't understand the mechanisms driving these longstanding gender differences. A commonly hypothesized explanation of this gap is the differential impact of childbearing on women. Prior work on this topic rarely considers longitudinal data on publications or has precise measurements of child birth, and as such, cannot measure the causal effect of having a child on productivity across time and complicates finding policies which increase the career satisfaction of parents.

Data: Using survey responses from 1,147 tenure-track faculty in 205 CS departments across the U.S. and Canada about the years of birth of their children, and linking these responses to nearly 250,000 publications over time, we investigate two research questions. First, we test whether there are productivity differences between male and female faculty with and without children by employing a difference-in-differences approach among parents and non-parents. Second, we examine if women's productivity is more sensitive to having children than men's by considering an event study, adjusting for career age and general productivity trends.

Results: In response to our first question, we find mothers publish more slowly than fathers, accumulating twelve fewer publications on average a decade after having a child. This is despite men and women following very similar trends in cumulative productivity in the five years before the birth of their first child. To measure the within-gender impact of parenthood, we constructed a control group by assigning ``placebo births'' to faculty without children based on covariates that predicted the age of first child among faculty with children. The covariates include birth year, the prestige of their PhD-granting and hiring institutions, pre-hire productivity, and gender. After lining up the productivity trends of faculty with and without children with respect to these births, we measure and compare the differences in productivity before and after having a child among male and female faculty.

In response to our second question, we observe that the productivity of female faculty is more strongly affected by having children than that of male faculty. Though the short-term effects of parenthood on productivity are somewhat mixed, we see that a decade after the birth of a first child, the average yearly productivity of women is around 10% below their pre-parenthood productivity levels. In contrast, the productivity of men has begun to slightly exceed their pre-parenthood levels. Here we model productivity of an individual five years before and ten years after becoming a parent, as a multiple linear regression. We adjust for age to account for lifetime trends in productivity, and include indicator variables for year in order to consider inflation in publication rates. We then estimate a ``child penalty'' -- the percentage difference in women's productivity compared to men's at a particular event time.

We conclude our analysis by considering how the availability of parental leave policies correlates with the recruitment and retention of women in science. About half of female faculty indicated that support for parental leave was somewhat or very important for choosing their current position, and this was compared to only a quarter of male faculty. Access to paid parental leave that is longer than four weeks was also suggestive of more willingness among women to entertain competing offers. We plan to expand upon these results further by considering responses from faculty at business and history departments. Altogether, these analyses will highlight new policy recommendations for evaluating the productivity of men and women after becoming parents.

ABSTRACT. BACKGROUND AND RATIONALE

There is a demonstrable gender gap in academe, observed in terms of academic production, impact, rank advancement, funding levels, and prizes. Several mechanisms have been investigated, from the social to organizational. A common explanation for the lack in productivity has been the disproportionate weight of domestic care on academic women. As a consequence this restricts their ability to engage fully in the demands of the academy, directing influencing their academic productivity both in the short and long term. The effect on productivity then has a cascading effect: leading to a subsequent Matilda effect in promotion, advancement, and prizes. This has led policy makers to focus on issues of parental leave and other initiatives focused on women as caregivers. However, this conceptualization as a women as the primary caregiver is challenged by contemporary changes in the role of women in society, the increasingly engaged role of men in parenting, and the constellation of caregivers involved in child rearing. These social change as well as the widely acknowledged demands of the academy allows families new options in childcare. Further it sees both parents, irrespective of gender and marital status, but also a range of non-parental figures taking an active role in maintaining the work-life balance in academia. What is missing is an understanding of the cost of different forms of parenting has on academic productivity and performance. In contrast to a range of studies investigating the relationship between gender and academic production and impact, this study focuses on time invested in parenting as the lead factor underpinning productivity differences for both men and women. It does not take parenting as a binary variable, nor does it focus exclusively on the experience of mothers. Using an international survey of academic parents (n=17,519 respondents), this research in progress uses the term “parent” as gender neutral, acknowledging that modern parenting is a joint, or multiple-partner endeavour. As such, this research aims to avoid ascribing loss of productivity on a single individual “parent” alone, in order to further investigate the parenting cost on academic productivity.

METHODS

Web of Science was used to sample all first and last authors who had published at least one article between 2007 and 2017. All authors were invited to complete an online survey. The first survey question used skip logic to eliminate all potential respondents who were not parents. In total, 17,519 individuals who met this initial filtering requirement responded, with nearly equal responses from male and females. Data cleaning was done to exclude unfinished responses and erroneous responses (e.g., doctoral degrees obtained before birth) and to account for missing data resulted in a final sample of 10,444. In addition a free text box in the survey allowed for the analysis of 5976 qualitative responses about participant’s experiences balancing parenting with their partners, and academic careers. ANOVA was used to test the null hypothesis that the mean productivity (i.e., number of papers) and impact (i.e., proportion of published papers that are considered highly cited relative to field and year (PPTop)), is the same for parenting type relative to gender (gender/parenting type categorisation). Permutation tests were used as a post hoc test comparison to further test the relationship between the gender/parenting type categorisation. Here, two test statistics were used; one that measures the square distance between each observation and the group mean; and the other measuring the difference between the group median. To analyse the qualitative data, a random sample of 500 responses were selected and coded thematically using a grounded theory-informed approach. 59 themed categories were developed and then collapsed into 8 overarching codes capable of coding the large number of responses. The remainder of responses were manually coded into these themes.

RESULTS

31.8% of female respondents indicated that they were the ‘primary caregiver’ to their children, compared with 4.2 of male respondents. In contrast, 33.2% of men indicated that their ‘partner is the primary caregiver’ to their children. Slightly more male (55.1%) than female (46.6%) respondents indicated that they ‘share equal roles with my partner’. Although both genders report engaging in dual-parenting arrangements and/or as a satellite parent; 24.1% of women are acting as the Lead-parent whereas less than 3% (2.9%) of men act the same.

A one-way ANOVA test was performed across all gender/parenting type categorization showing a significant difference (F(leadparent) F=25.31 p<0.001), and a further two way test, showed interactional effects indicating that the effects of taking on the Lead parenting role on academic productivity are different for men and women F(gender:leadparent) F=3.34, p=0.01. These results demonstrate that there is a significant cost to the role of lead parenting.

The proportion highly cited papers relative to field an year (PPTop) for each gender/parenting type categorisation was also investigated. The single factor analysis shows how the expected percentage is different for a least on gender/parenting type categorisation (F=3.40, p=0.001). The two-factor interaction ANOVA indicates that while there is a different in the expected percentage between males and females (F(gender:leadparent) F=0.15, p=0.92), there is no different for the lead parent role regardless of gender. Permutation tests verified this finding by showing a probability that the means and medians were the same for each gender/lead parent categorisation of close to 0 (=10-4), p=0.0016.

Results show a significant difference across all types of parenting relative to gender for the number of papers produced, as well as for the proportion of papers published in top journals. For men and women who take on dual parenting roles, the productivity cost is higher than for women. Further qualitative investigation highlights the incidence of an ‘invisible burden’ in self-identified dual parenting families, wherein there is a significant amount of unacknowledged labor that is undertaken by females. As described by one female respondent:

The mental labor of researching and remembering EVERYTHING related to kids activities and school falls to me - including selecting locations, remembering deadlines for sign-ups, getting proper equipment: summer camps, swimming lessons, dance, after school care, parties at school (bringing snacks/valentines etc.), field trips. It is constant, exhausting, and under-appreciated.

Men who were in dual-parenting arrangements acknowledged the existence of this invisible burden on women; “Although I try to be active in child care and share responsibilities equally, my wife still takes care of more child care tasks than I do”. There were, however, some respondents who noted the benefits of a functioning Dual-parenting relationship:

The system was not perfectly equal in all regards, but he made every effort to make it as fair to both of us as possible. That is a big reason why I have had a successful career in science.

SIGNIFICANCE

The results should that there is a connection between the amount of parental responsibility assumed by an individual and research productivity as measure by the number of papers, and the proportion of highly-cited papers. The model also show that there is a significant interaction with gender, suggesting that the link between parenting arrangements and productivity differs is different for men than it is for women. This study demonstrates how the level of parental responsibility is a powerful variable to explain academic productivity differences between men and women. Further research is currently underway to investigate these effects in more detail.

ABSTRACT. Scientists may cite articles for various reasons (Bornmann, 2008; Cozzens, 1989). Yet, the extant literatures from science studies show that citations can largely be understood and explained by the two following schools of thought. First, citations can be used as a means to allocate undue credits to their rightful owners, thus, reflecting the normative structure of science (Merton, 1973). Citations can also be viewed as a rhetorical device employed by the scientists to advance their claims and convince or refute their peers (Gilbert, 1977; Latour, 1987). Meanwhile, citations became an institutionalized practice in the academic publications.  Citations, just as any other institutionalized practices, can be susceptible to become a myth-like behavior (Meyer and Rowan, 1977). For example, scientists could hope to enhance legitimacies of their articles by citing unrelated but “influential” articles. Therefore, it may be not too uncommon to observe these perfunctory citations, in which citations are employed even when they have “no bearing whatsoever on their claims and might be there just for display” (Latour, 1987).

Understanding the extent to which perfunctory citations take place in science may provide implications for the users of bibliometric indicators. However, identifying perfunctory citations is an inherently difficult task since there are no systematic ways from using bibliometric information to find out whether scientists have read citing articles. Meanwhile, the retraction in academic publishing can provide interesting cases on which systematic investigation of citation behaviors can take place. Citations to retracted articles are absurd because retracted articles are considered invalid source of knowledge. Therefore, this empirical setting allows us to observe various circumstances under which citations are perfunctory. Furthermore, I argue that investigating post-retraction citations can itself be valuable because their non-trivial existences are potentially detrimental to the public integrity of science. This study addresses the following research questions. Does knowledge distance between the citing and the cited article explain post-retraction citations? If it does, does the distance effect indicate naïve ignorance or intentional citation out of “argument from authority”? Furthermore, I also investigate extent to which the field size and scope can explain post-retraction citation.

A population data of retracted articles was obtained from the Retraction Watch database. A subset of retracted articles (5,717 articles) indexed by the Web of Science was extracted. These retracted articles were matched to their citing articles, leading to a total of 134,514 citing-cited article pairs where each article pairs would be assigned 1 if cited after retraction and 0 if cited before retraction. The knowledge distance between the pair is calculated first by transforming each document into a continuous vector space using word embedding method (Le and Mikolov, 2014; Mikolov et al., 2013), and then by calculating the cosine distance between the pairs of vectors. The results indicate that on average, a unit increase in the knowledge distance leads to 5.8 percentage points decrease in the likelihood of post-retraction citations. After taking account of its differential effects with respect to the “prestige” of the retracted articles, the distance effect was only relevant under a subsample of the retracted articles published in the prestigious journals. Therefore, the post-retraction citations are less likely to be driven by the topical ignorance of the citers or the visibility of the retracted articles. The post-retraction citations rather reveal the prevalent incidences of the perfunctory citations.

The second analysis involves identifying the field-level explanations for the post-retraction citations. The field size and scope are calculated with the population of the PubMed database. Therefore, this analysis only uses a subsample of the retracted articles indexed by the PubMed database. The field size is calculated by taking average of the number of papers published using each MESH terms from a focal publication. The field scope was calculated by taking the average of the unique number of MESH terms that are co-assigned with each MESH terms from a focal publication. The result shows that the field size is positively correlated with the post-retraction citations, which indicate the difficulty (ease) of gatekeeping in a large field (small field). Interestingly, not only the small fields experience less post-retraction citations, but when coupled with a larger field scope, they were much less likely to experience post-retraction citations. This result indicates that a small field with diverse topics are better at accessing the proper uses of citations, through which scientists are socially compelled to allocate credits to the rightful owners.  

This research is the first to systematically study the existence of and the extent to which perfunctory citations take place in science. The extant studies tend to blame the lack of visibility as the main culprits behind the post-retraction citations (Campanario, 2000; Pfeifer and Snodgrass, 1990; Unger and Couzin, 2006), which tends to view the post-retraction citations as a phenomenon resulting from naïve and ignorant scientists making honest mistakes.  However, the evidence from this analysis rather supports a dim aspects of citation behaviors as the main culprit. Apparently, a non-trivial number of scientists cite for reasons irrelevant to the explanations from the Mertonian norms and from the social constructivists. Some of these scientists seem to cite articles from the “prestigious” journal just to increase the perceived legitimacy of their articles. When citations are used in this way, they represent myth and ceremony rather than credit allocation nor rhetorical device. The policy implication of this paper further supports the line of research that criticized the uses of citation as indicator of the performance (Hicks et al., 2015; Stephan et al., 2017; Wang et al., 2017). Articles published in the prestigious journals can garner citations for reasons irrelevant to the content of their research. Lastly, the field size analysis underscores the inability of a large field to effectively monitor science and that the increasing peer review burden among scientists could lead to the inefficient allocations of credits.

ABSTRACT. Nowadays, innovation and translational research concepts are commonly used in science policies. They often refer to the need to contribute to the diffusion of new scientific knowledge toward socio-economic impact. However, novelty is not exclusively related to discoveries: work can remain ignored for a long time, before being recognized. The aim of this work is to examine the mechanisms that could explain the resistance to the circulation of scientific knowledge and its transformation into innovation in the society. It was B. Barber (1961) who was the first in sociology of science to propose an explanation of resistances to discoveries. However, it was S. Cole (1970) who retranslated this question in terms of "delayed recognition" (DR) [1]. In the early 2000s, the problematic of the DR knows a renewed interest in scientometrics with the analysis of "sleeping beauties" (SB), a concept introduced by Van Raan (2004). It refers to an article that goes unnoticed (“sleep”) for more than 10 years ("sleep" period), and then, almost suddenly receives many citations (the "awakening" period) by a ‘‘Prince’’ (PR, another article), attracting a lot of attention from there on in terms of citations (the "Kiss of the Prince") [2]. SBs have been identified in numerous research fields such as biology, chemistry, medicine and physics, and estimated to occur between, 1 and 7.6% [3]. Therefore, study of SB or DR is very interesting in the exploration of mechanisms of circulation of knowledge [4]. The present work aims to explore, through systemic approaches and several case studies, the reasons for DR pattern of citations and what mechanisms are in play in the “awakening” of some scientific discoveries. Our methodology associates quantitative approach, scientometric analysis, and qualitative approaches, historical and sociological analysis, with semi-structured interviews. The first step of our research focused on the scientometric evaluation of the DR phenomena in relation to either (i) the scientific work of Nobel Prize (in Medicine, Physics or Economics), (ii) a concept: “tumor angiogenesis” and "mesenchymal stem cell" (MSC), (iii) a research domain: “astronomy and astrophysics”, (iv) some clinical fields, "oncology" and “drug safety” or (v) applied research: “solid state chemistry”. According to the cases, we extracted from the WoS® database a corpus made up of specific articles (between hundred and a million documents), looked for landmark papers by a historical analysis of article references (RPYS method) [5] and/or filtered DR papers by calculating for all publications the "Beauty coefficient" (B) using the non-parametric criteria of Ke et al. (2015) [3]. For each case studies, we documented the sleeping period and the awakening period by measuring the trends of publications or citations linked to the case, concept and field as well the echo in public media by measuring the press releases in various databases (Google Trends, Factiva). The second step focused on the identification of the Prince, ie the author(s) at the origin of the citation awakening, mainly by co-citations networks analysis. For some cases, we combined this scientometric approach with (i) a historical approach to produce a biography of the author and/or the Prince(s) scientific work to develop a chronology of the scientific discovery using second-hand data, (ii) several semi-structured interviews with researchers who started citing DR papers around (and since), the awakening year in order to grasp the reasons of the delayed recognition. Our scientometric analysis show that DR phenomena could affect article, book or concept, single publication or bundle of publications related to the same discovery, different scientific work related to the same author, single author or group of authors, highly cited scientists or publications as well average H-index author or average cited papers, sometime published in niche scientific journals. The sleeping period could extend form few years to more than half-century, affecting papers published between 1927 and 1988, awakened between 1978 and 2010 and ranging between less than 500 to more than 10.000 total citations. Many reasons have been suggested in the literature as to why the phenomenon of delayed recognition in terms of citations occurs and what mechanisms are in play in the “rediscovery” or the “awakening” of DR papers. Our results show that some publications “sleep” for a long time because they contain scientific claims that steer scientific or ethical controversy in a given field, or because they are at the boundary between two discipline. This “sleeping” period only ends when the controversy is closed, and when consensus is reached. Other publications don’t get enough citations in the beginning because they contain unproven hypotheses, and the citational “awakening” occurs when an experimental proof of the hypothesis is presented. In some applied fields, an industrial application is needed to make a forgotten publication relevant again . In some cases, the DR papers were simply not available to a wide enough scientific community to lack of visibility due to its publication in unknown journals. However, DR cannot be necessarily attributed to the publication of papers in languages other than English. But it could be attributed to the geopolitical situation of some researchers such as Eastern scientists lacking of communication and collaboration with Western scientists that hampered the diffusion of their scientific ideas. In conclusion, the occurrence of delayed recognition raises questions about the relevance of short-term citation-based metrics for the evaluation of scientific impact on which many research public policies are base worldwide. Constancy and continuity in a research field are important components that ensure development of new research subject area. In contrast, lack of continuous financing and research, could hamper the development and growth of new research area.

[1] Cole S. (1970), Professional Standing and the Reception of Scientific Discoveries. American Journal of Sociology, 976, 286-306. [2] Van Raan A.F., (2004), Sleeping Beauties in Science. Scientometrics, 59(3), 461-466. [3] Ke Q., et al. (2015). Defining and identifying sleeping beauties in science. PNAS, 112, 7431–7626. [4] Braun T., et al. (2010), On Sleeping Beauties, princes and other tales of citation distributions…. Research Evaluation, 19, 195–202. [5] Bornmann L. et al., (2018), Reference publication year spectroscopy (RPYS) of Eugene Garfield’s publications. Scientometrics, 114, 439–448.

ABSTRACT. We use retractions as a measurement to show how the ability to correct errors in science varies over time in various academic fields and in different parts of the world, therefore measuring the strength of underlying scientific institutions. Since science is cumulative, the ability to recognize past mistakes and correct errors is crucial for the continuous advance of sciences. Retraction is a highly visible event in which an error in science (caused by scientific misconduct or genuine mistakes) is revealed and the scientific enterprise responds by not only punishing the offender(s) through tarnished reputations, but more importantly, redirecting future researches towards correct directions. Such a correction in the direction of science is observed as a drastic reduction in future citations of retracted papers, and the varying degrees of reductions in various institutional settings reflect the strength of underlying scientific institutions. Building on this notion, we construct a measure of correction effect (CE) in science, ranging from 0 to 1, where, CE=1 if retraction eliminates all future citations; and CE=0 if no future citation is affected. We hypothesize that the size of CE depends on technological conditions, work organization, and institutional underpinnings in various science systems.

We analyze correction effects in all 4072 retracted articles recorded in Web of Science between 1978 and 2017. We implement a difference-in-difference research design in which the net impact of retraction on citations is estimated through comparing retracted articles (n=4072) to a control group of nearest neighbor articles (n=8009) published in the same journal, volume, and issue. We analyze the whole sample and subsamples divided by time, disciplines, and countries to obtain the average treatment effects of retractions, or the measures of CE. Consistent with previous studies, we estimate an average of 68% post-retraction reduction in citations, or CE=0.68, globally. Yet correction effects vary significantly by time, places, and academic disciplines.

We find that there is a long-term trend towards more scrutiny in science globally. Driven by the adoption of information technology, the global correction effect strengthened in the late twentieth century, with CE growing from 0.45 in early 1990s to 0.93 in early 2000s. However, with the rise of emerging science powers such as China, India, and Korea, the global correction effect weakened in the 2000s, dropping to 0.53 by the end of the decade. In the most decade, we find CE is slowly returning to its historical trend, increasing to 0.72 in 2017.

We find that the correction effect varies by academic disciplines, where correction effect is strongest in social science (CE=0.73), followed by applied technology (CE=0.66) and natural science (CE=0.60), and weakest in life science (CE=0.56). Consistent with previous studies on causes of retractions, this finding shows that as academic fields, especially life science, grow in sizes of scientific teams and increase in complexity in the division of labor, not only these fields are more vulnerable to scientific misconduct, but the correction effect also weakens in these fields.

We also find that there are significant differences in correction effects in different parts of the world. Among the largest five contributors to retracted papers, correction effect is strongest for articles with authors from Germany (CE=0.67), followed by the U.S. (CE=0.63) and Japan (CE=0.63), and weakest in India (0.59) and China (CE=0.58). Yet, the number of retracted papers with German authors (n=251) is much smaller than the U.S. (n=1144) and China (n=982), and is largely driven by outliers. A time-series analysis shows that it is the U.S. that has a significant lead in correction effects over the rest of the world in most of the sample period (1978-2017) with on average of 0.29 higher in CE. This strength of correction effect in the U.S. science system can be explained by two factors: a “Competition Effect” in which intense competition among best people in the U.S. science system increases efficiency, and a “Spotlight Effect” in which more attention is paid to articles with U.S. authors because of the central stage occupied by the U.S. in science.

Through a quantitative study of correction effects following retractions, this study sheds light on the error correction mechanisms and institutions in the scientific enterprise. Existing literature has measured the negative impacts of retraction on future citations, yet none has examined the other facet of retractions: that is, the varying impacts of retraction in different institutional settings reflect the strength of scientific institutions. Our study fills this gap by highlighting the varying correction effects across time, places, and fields, and we explore the potential causes in technology, organization, and institution accounting for these variations.

There are policy implications from this study. To allow science to flourish in the long run, science policy should enhance correction effects and strengthen scientific institutions. To do so, science systems should adopt technological advances (especially in information technology), manage scientific workplace (through improved team science and division of labor), and strengthen national institutions (via community self-policing and increased researcher capabilities).

ABSTRACT. Objective：By its nature, academic fraud is hard to detect. Yet whistleblowing, as oversight power combating research misconducts has not captured deserved attention from Chinese scholars and research managers. This study aims to explore the determinants of whistleblowing intention from two perspectives: individual perception and institutional factors. Method：The study uses a mixed-method research design. Based on the survey data collected from China’s 985 universities in Yangtz River, this paper investigate the factors impacting scholars’ whistleblowing intention. The interviews with university ethics committee further provide a more comprehensive picture on how institutional factors influence whistleblowing behavior. Results：Our analysis reveals that both individual perceptions on research misconducts and institutional factors have significant yet heterogeneous impact on whistleblowing intention. Institutions' reluctance to tackle misconducts push whistleblowers resorts to the mass media. Disclosure intention varies by the social distance of whistleblower and the wrong misconductors.

Conclusion: We argue that it is time to establish a formal institution for research integrity management in Chinese universities. To cultivate research integrity, clarified responsible entities, standardized whistleblowing procedures, and protection mechanism for both whistleblowers and those wrongly accused should be embedded in the system.

ABSTRACT. The process of innovation is intricately linked to the application of knowledge to generate new or improved products or devise new production and management processes (Fagerberg et al., 2010). Knowledge production in public research institutions and absorptive capacity in firms therefore play an important role in determining the rate at which creative innovations can diffuse across value chains (Lämsa, 2008). In many developing countries, sizable knowledge asymmetry between innovation producers and target users impede innovation adoptions and deprives communities of the benefits of technology diffusion (Jacobs et al; 2019). Quite often, application of technology-push paradigm (based on the assumption that new technological advances resulting from R&D and scientific discovery, preceded and ‘pushed’ technological innovation via applied research, engineering, manufacturing and marketing towards successful products or inventions as outputs) has proved incapable of gaining the support of target users because of failure to appropriately take their views into account at the design phase (Kline and Rosenberg, 1986). Those who develop technological innovations to push them to the market run the risk of misidentifying or misarticulating the need they purport to address and may end up with a non-existent market and or the wrong product (Wolfson 2010).

Overcoming the challenge of knowledge asymmetry and limited technical competence in underdeveloped areas requires a cooperative learning approach between knowledge producers and target users because the production and diffusion of knowledge is embedded in a social system of knowledge production (Lundvall 1996; Parrilli et al. 2010). Knowledge exchange with users in such cases where capabilities are weak often tends to require intermediaries (knowledge brokers) to play an important role when the involved technology is fairly simple, such as in Theodorakopoulos et al. (2012). When knowledge exchange is hampered by local capacity constraints due to resource poverty and absence of potent institutional arrangements, the use of knowledge brokers may however become unworkable for economically marginalised rural communities. This often results in tensions between the existence of a strong desire for immediate reaping of benefits of the technology to be adopted and the lack of financial and technical resources needed for a successful adoption. Such tensions create an expectation gap when the proposed technological solutions require learning before they can produce the expected benefits (Theodorakopoulos et al. 2012).

The question this study seeks to answer is therefore: -How can the necessary knowledge exchange between universities and recipient communities be mediated in the absence of local financial resources to pay for knowledge brokers?

This question is particularly relevant when the knowledge required to solve the problem at hand is fairly complex and the knowledge recipients are in marginalised rural communities lacking the ability to pay for costly brokerage transactions and extensive training. To overcome these diffusion challenges, technology recipients must have ongoing interactions with producers and other actors involved in the innovation value chain to deepen their knowledge and technical expertise. The effectiveness of the transfer of technological knowledge will ultimately depend on how well the recipients of technology engage with knowledge producers and with each other to achieve common goals (Chakrabarti and Rice, 2003; Chesbrough et al., 2006, Boardman and Gray, 2010). Innovation scholars have purposed various alternative mechanisms aimed at facilitating knowledge mediation between producers and recipients in order to cut through these obstacle to diffusion. Nonaka (1994), Lundvall and Johnson (1994), Nahapiet and Ghoshal (1998) and Lawson and Lorenz (1999), for example, have argued that he process of diffusion is embedded in and shaped by learning capabilities of recipients and suggested learning models that facilitate the building of these capabilities for knowledge absorption. On the conceptual level, Nonaka’s (1994) dynamic model of knowledge creation provides an example of a co-learning process based on the assumption that human knowledge is created and expanded through social interaction, and a mixture of tacit and explicit knowledge (Lämsa, 2008). For Lundvall and Johnson (1994), overcoming the challenge of limited technical competence in underdeveloped areas requires a co-learning approach between the innovating unit (e.g. a university) and users (community). In Lundvall and Johnson’s (1994) user-producer model, the innovating unit needs access to information about the needs of the user to successfully innovate. For their part, Nahapiet and Ghoshal (1998) have proposed a knowledge exchange mechanism that enables such a learning to take place support innovation diffusion. This mechanism consists of networks of strong, crosscutting personal relationships developed over time that provide the basis for trust, cooperation and collective action.

As for Lawson and Lorenz (1999), they propose collective learning in clusters and networks as necessary to overcome the limits of individual firm’s learning imposed by the difficult transmission of tacit knowledge from one firm to another. This co-learning process also helps overcome the constraints of knowledge asymmetry inherent in the linear transfer of technological know-how involving codified and tacit knowledge. It acts as an ignition phase in the process of knowledge co-production between researchers and other and key stakeholders, which is crucial for the successful development of new ideas and innovative solutions, as pointed out by Pohl et al. (2010). The effectiveness of the interactive learning processes in enhancing the technology adoption capacity of target recipients is empirically shown by Petersen et al. (2016). Organised in network structure in which all forms of knowledge, whether formal or traditional, were valued for their contribution in achieving the desired outcomes performs better them the linear technology transfer approach. Organizational learning and knowledge creation based on a continuous and dynamic interaction between tacit and explicit knowledge can thus be a potent tool to overcome the constraints of the linear innovation and technology transfer models (Lämsa, 2008).

This paper aims to explore the use of co-learning approach as an alternative to knowledge brokerage in order to help rural communities and knowledge producers overcome the difficulties posed by traditional interaction methods when dealing with knowledge transmission associated with bringing new production methods targeted to addressing local challenges. Drawing from Petersen et al. (2016) who underscore the crucial role of interactive learning spaces in the success of community engagement, we posit that knowledge exchange for societal benefit occurs through intensive mutual learning that facilitates knowledge sharing within interactive learning spaces. University-based living labs projects in South Africa provide a suitable setting to probe the effectiveness of knowledge exchange between users and producers in these interactive learning spaces. Our analysis of the co-learning and innovation co-creation process in those living labs sheds light on which knowledge exchange processes can be used to facilitate the contribution of science for society in resource-poor communities.

This study contributes to the debates on knowledge transfer to communities by shedding light on the practical applications of cooperative learning at the interface of knowledge exchange between universities and resource constrained rural communities in developing countries. It uses the case of co-learning and knowledge co-creation experiences in South African living labs projects to provide additional evidence of the use of co-learning to overcome capacity constraints to the absorption of technological knowledge intended to address local community challenges.

ABSTRACT. The concept of “Open Science” is not well-defined and is used in various ways by different authors, so that no consensus on its meaning exists. A recent literature review therefore proposes to define Open Science as “transparent and accessible knowledge that is shared and developed through collaborative networks” (Vicente-Saez and Martinez-Fuentes 2018, 433). This approach focuses on knowledge as materialized in artefacts and research output. Smith and Seward (2017) take an alternative tack which involves shifting focus to openness as a dimension of social practice. This is particularly relevant in the context of the engagement of researchers with open science, which denotes after all a social practice. For example, the share of openly available knowledge artefacts may increase, but if these are not included in practice, this increase is not relevant. When it comes to developing indicators for researchers’ engagement with Open Science and Scholarship, we therefore propose to shift the focus from indicators of outputs (i.e. whether publications, data, materials, etc. are available somewhere, somehow) to open knowledge practices and the actions to support and enhance these practices (i.e. whether actual use of knowledge, data, materials etc. is fostered).

Open Science and Scholarship is not about simply making knowledge unconditionally available to all. In fact, elevating an undefined "openness" to an absolute moral good might inflict serious damage on the knowledge capacities of both research communities and societies at large or restrict itself to mobilizing science for private companies in the framework of the new "platform capitalism" (Mirowski 2018; Van Dijck 2018). We define Open Science and Scholarship as the ambition to democratize the scientific and scholarly system while maintaining and further developing its critical capacities for all stakeholders in knowledge (“Towards a Learning Economy. Synopsis of WRR-Report No. 90” 2014). We thus treat the term ‘Open Science’ as shorthand for the notion of institutionalizing more socially robust – and less hierarchical – knowledge practices, which include not only science, but also other forms of research in the arts, humanities and social sciences, as well as teaching, technology, and infrastructure. This means that research assessment (and indicators) also need to take into account the uneven distribution of resources and capabilities over scientific communities across the world. Open Science should not only benefit the resource-rich communities (Leonelli 2013; Treadway et al. 2016; Hormia-Poutanen et al. 2017).

Science, and by implication Open Science, is intrinsically diverse. The objects, concepts, methods and practices of research differ greatly from astronomy to nursing, from agronomy to history. Open Science practices also differ across fields, countries and social contexts. This means that Open Science policies need to address generic issues such as intellectual property and infrastructures while they also need to be sensitive to these specific contexts. Given the diversity of open knowledge practices that need to be developed and supported as part of Open Science, different indicators are needed to describe these different research practices and even the same indicators are most often not comparable across contexts.

Indicators constitute a form of social technology that has specific effects on institutions and behavior. Unless we account for these effects before introducing and applying these social technologies, perverse effects may follow. If a monitoring or evaluation system focuses on one type of indicator only, it favors the types of research that are captured by this particular indicator at the expense of not rewarding other modes of research. This often has steering effects on science, as researchers may choose to focus on the types of activities that the indicator tracks. It is understandably tempting for research managers (at various levels, from national funding systems, to universities, to research groups within universities) to attempt to harness the steering effect of various metrics and indicators. However, the fact that indicators may be used to steer or nudge researchers toward certain behaviors does not exhaust the considerations that should go into choosing indicators. Indeed, one might suggest that researchers themselves should have a leading role in developing the indicators that managers will use to monitor them. The purpose and context of monitoring or evaluation also matter. It is important to seek an alignment between the open knowledge practices to be ‘indicated’ (e.g. data sharing), the indicators that should describe these activities, and the uses of the indicators in policy or evaluation.

In order to meet these demands, we propose to use ‘indicator frameworks’ to guide the appropriate and responsible use of indicators to assess or monitor specific types of research for a particular purpose (be it accountability, resource allocation, learning exercises, public understanding, scholarly analysis and debate, advocacy, or giving an account of one’s own research practices). Indicator frameworks aim to capture the main relevant knowledge practices of similar research environments – therefore, they allow a certain degree of contextualization and a certain degree of comparison across similar contexts. They also allow researchers to analyze reflexively their open science practices.

In addition, we propose to understand the transformation of the scientific system towards an open framework as the development of novel open knowledge practices by researchers as well as users of scientific and scholarly knowledge. We distinguish three levels of use of indicator frameworks:

1.the scientific system as a whole, including the infrastructures that are required for open science; 2.the research performing organization and research funding organization; and 3.the individual researcher or research group.

Because indicators are heavily dependent on the specific contexts assessed and the particular conditions of their use, there cannot be a single set of universally applicable indicators. However, it should be possible to use the same set of indicators for similar research contexts under the same evaluative purpose, i.e., indicator frameworks. In this presentation, we will lay out examples of indicator frameworks at each level and discuss how they can be used by various stakeholders and for various purposes, including not only top-down monitoring, but also bottom-up attempts to communicate the value of one’s research.

Acknowledgements: This presentation is based on the work of the EC Expert Group on Indicators for Researchers’ Engagement with Open Science and its Impacts. See: https://ec.europa.eu/research/openscience/index.cfm?pg=altmetrics_eg

ABSTRACT. Background and rationale Imagine a world of open science, where not only is research information openly available, but also where researchers are accurately credited for their contributions, where research institutions can be assured that their name is accurately cited, where funding organizations can easily collect works that were created using specific funding or resources, where publishers can streamline submission and review processes, where innovative and accurate search and discovery tools are available, where researchers can easily share electronic information about their contributions, and where evaluators have access to open data to study knowledge flows.

In the world we live in, research is hampered by lack of openness and transparency. Similarly, research evaluation is hampered by a lack of data that clearly connect a research program with its outcomes and, in particular, by ambiguity about who has participated in the program and their contributions. Making these connections manually is very labor-intensive; algorithmic matching introduces errors and assumptions that can distort results; and both methods exclude the people involved in performing research, who have deep knowledge about their research process and outputs.

To address these challenges, the research community is making major investments in open science, including the use of open digital identifiers for the people who do research, the contributions they make, the resources used and grants awarded, and the organizations involved. These identifiers make it possible to accurately map information between systems and discern the graph of connections between people, places, and things, that describes the research and innovation ecosystem.

For the open science vision to become a reality, each part of the community must participate in building information infrastructure that enables sharing of information about research outputs and evaluation of their impact. In this paper, we discuss the development and testing of a three-dimensional evaluation framework to assess community readiness and adoption of open science infrastructure, using ORCID as a case study. ORCID, a non-profit community-based organization, is one component of the open science infrastructure, providing identifiers for researchers. Along with sister organizations Crossref and Datacite and emerging organization identifier providers, we make it possible to create open and transparent connections between researchers, their works, and the places they do their work.

Methods Our evaluation framework brings together Technology Readiness Levels developed by NASA with awareness criteria developed in the communications field. These two sets of criteria help us to gauge the level of open science engagement in a community. We then apply a context criteria set. Context is best understood as the community’s relationship to open science objectives and their specific political, economic, and/or social/cultural circumstances -- as well as their willingness to engage and take action. We combine TRL, awareness, and context to identify the kinds of actions or intervention that would be appropriate to increase the maturity of open science practices in that community.

We are testing this framework in two ORCID communities of practice: (i) national ORCID consortia and (ii) research information platforms that have built-in ORCID technology. Each target group is assessed on TRL and awareness,, and the intersection of these creates a matrix, each cell of which contains a model set of interventions. Context is then applied in a third layer to add precision and priority to the interventions proposed from the initial analysis.

Anticipated results This evaluation framework will enable us to determine the nature and maturity of ORCID’s embeddedness in a variety of contexts, and to define clear goals, evaluate progress towards them, and articulate and test a tailored action plan for each target community.

In addition to the ORCID community mappings, we will assess the applicability of the model to the open science community, including key strategies and barriers that surface during our work. We seek feedback from participants at the Atlanta meeting to understand the framework’s broader relevance and utility for the technology and innovation policy community.

Significance ORCID has a complex, diverse, global community of communities, operating at differing levels of ‘readiness’. Everything we do at ORCID is in service of our underlying mission, to enable transparent and trustworthy connections between researchers, their contributions, and affiliations. This involves collaborating with all sectors of the global research community -- disparate as their priorities and cultures often are -- to facilitate interventions that help increase the openness and reliability of research information. To support and sustain this community effectively, we need to be able to assess the maturity and extent of ORCID engagement and adoption and to do so in a consistent, adaptable, and action-oriented way. The evaluation framework we describe in this paper is a part of our work toward bringing global-scale coherency to our effort.

ABSTRACT. Background: There is a powerful desire among government and non-profit scientific funding agencies to support research and development (R&D) that has broad impacts, generates responsible innovation, and positively impacts society (Bush, 1945; European Commission, 2018). Since the late 1990’s the National Science Foundation (NSF) has required scholars to discuss the broader impacts of their research, and consequently, there are studies assessing the broader impact activities (BIA). These studies classify the type and rate of BIA in NSF grants (Kamenetzky, 2013; Roberts, 2009; Wiley, 2014). However, there are several gaps in the literature. First, it is not clear the extent the NSF is funding research that helps poor and marginalized communities. Second, scholars did not measure the extent BIAs are integrated into the research. Are the BIAs a core part of the project or added as a side activity to the research? To better assess the impacts that research has on marginalized communities, this article outlines a new framework, called the Inclusive-Immediacy Criterion (IIC). The IIC extends other scholarship on broader impacts to include an immediacy dimension and an inclusion dimension. Immediacy refers to the integration of the broader impacts and the research. There are three types of immediacy: intrinsic, direct, and extrinsic. The first type of immediacy, intrinsic, means that the BIAs “are accomplished through the research itself (Jacobson et al., 2016).” For instance, if a PI is developing a new malaria vaccine, the broader impacts are intrinsic to the grant. The second type of immediacy, direct, are broader impacts that are achieved while conducting the research. An example of direct immediacy is training graduate students while doing research. Training the student is not the goal of the research project, but it would impossible to finish the project unless graduate students are trained and work on the project. The third type of immediacy, extrinsic/far reaching, are “accomplished through activities supported by but complementary to the project (Jacobson et al., 2016)”. If a PI visits a middle school to discuss the career path of a scientist, then the BIA is separate from the research and has extrinsic immediacy. The second dimension, inclusivity, determines the target recipients for the BIAs. Inclusive research occurs at a variety of levels ranging from developing a medicine that treats a disease of poverty, to constructing institutions that give a political voice to marginalized communities (Heeks et al., 2014). The team condensed inclusive research and innovation to three levels: advantaged, universal and inclusive. By combining immediacy and inclusive, we created a three by three grid of differently types of BIAs.

Methods: To test the framework, the study analyzes NSF sponsored nanotechnology grant abstracts from 2013 to 2017. Out of the 6,854 nanotechnology grants, the team randomly selected 300 grants to code for the broader impact activities and the IIC classification.

Results: On the immediacy dimension the most common category was direct (232) followed by intrinsic (200) and then extrinsic (110) grants. None of the grants only proposed broader impacts that were extrinsic to the grant. In the other dimension, the most common type of inclusion was advantaged (235), followed by universal (213) and inclusive (109). Within inclusive grants 94% of them are about broadening STEM participation. When comparing the IIC to the traditional BIAs, there are obvious overlaps. 160 out of 162 grants that are universal/intrinsic have potential societal benefit and K-12 outreach and Universal/Extrinsic are closely linked together.

Significance: Overall, most of grants in the sample research will help everyone (universal), 213 grants, or advantage groups, 235 grants. 109 of the 300 grants proposed broader impacts that were inclusive and only 9 grants propose research that is intrinsic and inclusive. This means that most of the grants propose research that was not specifically directed towards marginalized groups. Depending on the policymaker’s perspective these ratios can be viewed positively or negatively. This study cannot determine the ideal distribution of science funding, but by reexamining broader impacts activities through the lens of inclusivity and immediacy, scholars can better inform science policy decision makers about the impact NSF sponsored research has on marginalized communities.

Bibliography: Bush V (1945) Science The Endless Frontier A Report to the President by Vannevar Bush , Director of the. Washington D.C. European Commission (2018) Science with and for Society - European Commission. Available at: https://ec.europa.eu/programmes/horizon2020/en/h2020-section/science-and-society (accessed 19 November 2018). Heeks R, Foster C and Nugroho Y (2014) New models of inclusive innovation for development. Innovation and Development 0(February 2015): 1–11. DOI: 10.1080/2157930X.2014.928982. Jacobson S, Hajjar J, Tilbury D, et al. (2016) Workshop: Setting a Broader Impacts Innovation Roadmap. Arlington, Virginia. Kamenetzky JR (2013) Opportunities for impact: Statistical analysis of the national science foundation’s broader impacts criterion. Science and Public Policy 40(1): 72–84. DOI: 10.1093/scipol/scs059. Roberts MR (2009) Realizing Societal Benefit from Academic Research: Analysis of the National Science Foundation’s Broader Impacts Criterion. Social Epistemology 23(3–4): 199–219. DOI: 10.1080/02691720903364035. Wiley SL (2014) Doing broader impacts? the National Science Foundation (NSF) broader impacts criterion and communication-based activities. Iowa State University Digital Repository. Iowa State University. DOI: 10.1017/CBO9781107415324.004.

ABSTRACT. Abstract Little or no success has been recorded by academic and research institutions in the developing countries in terms of generation of technologies in the globalized world-economy.This deficiencies and non-productive tendencies of the institutions can be linked to ineffective academic –linkage approach and drive of such institutions.Individual academic is either directly or indirectly involved in knowledge utilization for effective teaching and learning. However, in a research-based institutions, where research seems to be their key mandate different professionals and experts are concerned with generation of technologies.A low generation of technologies will result to a small number of IPs in any country. Academic and research institutions can generate a number of technologies through identification of societal and industrial problems, and a good number of IPs can in turn be meaningfully generated and managed.This requires a handshake with the industries using a well-developed framework.This research work therefore develops a novel framework for academic-industry linkage for Intellectual Property (IP) generation and management using a TVET institution in Lagos, Nigeria as a case study.This study uses a conceptual and systematic approach in the development of framework between the academic and industry. This involves the use of literature, existing models, interviews and content analysis.The result of this study is a well-developed framework adoptable and usable in any HEIs and in particular a TVET institution for generation and management of IPs. Owing to the different management structure of institutions,it is recommended that each institution develops IP policy with a special inclusion on academic-industry outputs. . Background and Rationale The TVET in most of the developing countries is expected to play two crucial roles in the national sustainable development (social, economical & environmental development).The first role is to provide training opportunities and career advancement avenues for the increased school leavers. The second role is to provide skilled manpower that is needed at all levels of the economy. The skills so developed should be able to lead to self-reliance in the absence of salaried employment and enhance the industrialization process.(Mous tafa, 2017)

Currently , the trend and practices in most of the TVET institutions requires necessary support and attention since a lot of innovations and technologies are being generated in the institutions inspite of their workings towards realization of mandate for TVET. According to (Marope et al., 2015), Technical and vocational education and training (TVET) is steadily gaining popularity at the global debates and government priorities for education and national development agendas. Globally and locally, TVET is gaining momentum at the global, regional, and national level. A good framework is required to generate IPs and manage it in order to reap the potentials of TVET for contributing to socio, economic, and environmental sustainable development.

In developed countries, collaboration between universities and industries has been identified as critical for skill development (education and training), the generation, acquisition and adoption of knowledge (innovation and technology transfer), as well as the promotion and encouragement of entrepreneurship, i.e. start-ups, spin-offs and incubation centres. Among the benefits of academia-industry linkages include coordination of Research and Development are wide-reaching and can help in coordinating R&D agendas, avoiding research duplications, stimulating additional private sector investment, exploiting synergies and complementing science and technological capabilities. It is imperative to establish linkage with the productive sector to proffer solutions to industrial problems; to generating income to the HEI and promoting economic and social development. In order to establish linkage with relevant industries based on these goals, a number of considerations are to be critically examined These include; type of Partnership, management support, capacity, Infrastructure, IP, Industries and Industrial Visitation, Communication, Monitoring and Evaluation (Post Linkage). There are many partnerships that exist, and can be formed with relevant industries ,among which are partnership to transform teaching and learning , partnership for Research and Development (R&D).Depending on the TVET institution, research priorities is key in engaging relevant industries in the area of Science, Technology and Engineering (STE) with smart partnership for R&D. For an effective linkage with industries, full support of management is critical and highly needed. Such support shall include funding (financial incentive for R&D partnerships) which would strengthen the interest of researchers in solving industrial problems, and further show commitment to the established partnership with concerned industries. Also, regular specific training/education should be organized for researchers. All the necessary support needed to enhance the linkage is to be timely for high-quality results/outputs to be achieved. Obviously, industries has never been reliant on the solutions from Nigeria academia , and a teeming population of Nigerians is also yet to reap from the intellectual capability and outputs of Nigerian academics. Evidently, there are fewer products made by Nigerian academic, which are either in Nigerian or international market.This could be linked to the lack of a successful handshaken of academic –industrial linkage. A simple and workable framework is required for any institution whose primarily responsible is to solve societal and industrial problems. Method: This requires extensive review of literature on academic - industry linkage for intellectual property (IP) generation and management .An emphasis on the linkage involving TVET shall be explored and reviewed. A logical sequence of prodecures and processes , and various components and implementers would be presented using model diagrams. Responses would be gathered from a group of experienced academics and researchers at a business meeting and a number of critical questions bordering on the need for academia to be proffering solutions to the numerous societal and industrial problems would be discussed. An in-depth analysis of the responses of group of experienced academics and researchers and IP professionals would be presented and used as model inputs.The model would be developed with TVET perspectives for both mission and vision realization. Each component contains objects in the framework; the object can be linked to another object for information flow, decision making and necessary actions of the various components linkage between an institution and industry. Result A simple , robust and component-based framework for effective IP generation and management between the academia and industry is presented. Each component is broken into sub-components in a sequential order.Each sub-component is broken into element.This is to allow complete representation of processes and procedures in the framework. The framework is a two-way mode framework which could either be demand or supply –driven to and from the industry.The first framework presents IP generation mechanisms while the second framework presents IP management mechanisms. Significance While many institution especially research-based have different framework or approaches for IP generation , this novel framework can serve as model framework that can be adopted and used especially in a TVET research –based institutions in a developing country. The framework can be used to solve societal and industrial problems as it can be identified in the any industry and society.

ABSTRACT. While both Ecuador and Colombia maintain a top-down policy type in the design of their public policies, the differences between these two countries as regards STI are quite significant. In order to explain the different policies at hand we use policy networks, which can show us how actors define themselves in direct relation to other groups of actors, what their capacities are, as well as their preferences and the structure of their relations.

This contribution explores how the actors’ programmatic ideas are stabilized and how strategic learning is enhanced within the policy domain of science, technology and innovation. We compare the change in STI public policy of Colombia and Ecuador that began in 2006, the year during which a change of government took place in both countries, where innovation has been used as a go-to answer for policy change and a touchstone for governmental legitimacy.

Networks show the structure of the relations established by actors and institutions during the change of public policy, and this structure is fundamental to understand the programmatic ideas of the actors, their preferences and the negotiations of influence and domination that they exercise between each other. In both cases we took the change of government in 2006 as the policy window that allowed the STI policy change. In Ecuador this window contributed to the institutionalization of policy through the linear model, with new governmental entities and projects like Yachay City of Knowledge. In Colombia, this policy window helped to turn public policy towards productive innovation and the deepening of the triple helix through Law 1286.

ABSTRACT. 1 BACKGROUND AND RATIONALE Research over the years have revealed important findings related to the role of capabilities in innovation and policymaking, both in Global North and Global South. Over these years, concepts such as technological capabilities (Bell, 2009), dynamic capabilities (Teece et al., 1997), absorptive capabilities (Cohen and Levinthal, 1989; Cohen et al., 1996), have been developed and used to explain the role of capabilities in innovation. Although these concepts have helped to shape and reshape capability research in academia, and in the practice and application of capabilities in organisations (firms and governments alike); (Borras and Edquist, 2019 forthcoming) argue that these concepts need further conceptual clarification. Relatedly, various theories and frameworks have been put forward to either help explain capabilities for innovation, advance research in the field, or guide capacity building. Some of these frameworks include the National Systems of Innovation (NSI) (Freeman, 1987, Lundvall, 1992, Nelson, 1993), Regional and Sectoral Systems of Innovation (Malerba, 2002, Malerba, 2005), and Technological Innovation Systems (TIS) (Bergek et al., 2008).

However, despite the progress made so far in capabilities research, we find that many gaps remain.

First, the capabilities field remain narrow in terms of the discourse and interdisciplinary nature of the research that is being undertaken. The new challenges (e.g. social, economic, environmental, development), technological advancements (e.g. Artificial Intelligence, and 4th Industrial Revolution), and the global agenda towards sustainability; require new capabilities or updating of existing capabilities available to organisations, sectors, national and regions. Consequently, various scholars argue for the need to broaden the capabilities discourse and research in ways that more specifically address the needs of Low- and Middle-Income Countries (LMICs).

For instance, Dutrénit (2004) ‘argues that there is no simple linear progression from the early stage of accumulation of the minimum levels of innovative capability to the management of knowledge as a strategic asset and the deployment of core capabilities’ (p 1). This relates to the accumulation of stocks, which in this sense refers to physical stock (technological artefacts, infrastructure and human capital) (Cirera and Maloney, 2017). Relatedly, new theoretical framings in the field of innovation systems study, argue for broader understanding of the NSI, in ways that incorporate and emphasise development and social inclusion, which is not only about reduction of poverty but reduction of inequalities (see for example, Dutrénit and Sutz, 2014).

Second, the emergence of ideas such as those addressing climate change and environmental issues, the UN global framework of Sustainable Development Goals (SDGs), or the need for inclusion (i.e. the inclusion agenda); raise questions as to the stock of capabilities not readily available to researchers, firms and policymakers in LMICs. These global agendas require capabilities to fulfil the expectations of countries and their commitment to tackle greater challenges such as emissions reduction (e.g. National Determined Contributions, NDC).

Third, we find that these capabilities concepts (such as those outlined above), global agenda (e.g. SDGs), and theoretical or policy frameworks (e.g. different flavours of National System of Innovation and Systems Transformations) the majority of which have their origins in the Global North, continue to be applied across board in the Global South, oftentimes regardless of the context, educational status, capability levels, or development stages of the LMICs. For example, the operationalization of the SDGs, practices around sustainability transitions approaches, or the application of the NSI assume a uniform level of education, knowledge and capabilities across the Global North and Global South. In relation to the NSI, various authors have shown that the “systems” in LMICs are weaker, less capacitated, less organised and linked, and that the actors generally interact less, when compared to the Global North or High Income Countries (HICs). Despite these challenges, the tendency to diffuse, and apply global frameworks across LMICs persist.

Put together, we argue that these gaps call for a revisit to the research on capabilities in LMICs. We hope to show through this research that these gaps in capabilities have implications for practice, innovation and policymaking in organisations (be it farms, firms or government organisations). In order to examine these issues raised, we focus on three areas: (1) identifying and assessing the current and future stocks of capabilities, for development and eventually transformations. Here, we will look at education and skills; (2) capabilities for innovation in firms; and (3) capabilities for policymaking in government organisations such as the ministries of science and technology or similar agencies, charged with the formulation, implementation, evaluation or governance of science, technology or innovation policies.

While systems capabilities for firms and regions have been explored (e.g. Alvarez Tinoco, 2011); capabilities for policymaking take two perspectives: (a) individual capabilities, i.e. skills unique to an individual policymaker or (b) organisational capabilities such as process and routines (Daniels, 2015). This research delves deeper into capabilities for innovation in firms and capabilities for policymaking in governments.

2 METHODOLOGY This project started in 2018, and is still in its early stages. The research follows a mixed method approach that involves: (1) an in-depth review of research and literature on capabilities, going as far back as the early ideas on capabilities by Penrose (1959) and Sen (1985), through Bell (2009), to recent discussions on capabilities in relation to LMICs (Cirera and Maloney; 2017); (2) a full-day workshop to explore the issues and themes emerging from the literature review; (3) a survey questionnaire to key innovation actors involved with capabilities research/management in five LMICs in Africa and Latin America; and (4) expert interviews with selected individuals (researchers, innovation managers and practitioners in firms, and policymakers in governments) to follow-up on outstanding issues from steps 1-3, useful for triangulation of the empirical data captured.

3 ANTICIPATED RESULTS Anticipated results include: (a) empirical findings, at systems levels, based on the assessment of the current capability gaps in LIMCs, and what is needed to address the modern-day challenges facing LMICs; (b) lessons on some of the capability concepts, theories, approaches, and frameworks, that have proved to be most widely used in LMICs over the past decades, and insights on the outcomes; and, (c) an improved understanding of how capabilities research in LMICs could be (re)designed and operationalised in ways that lead to better capacity-building results, socio-economic outcomes and development impacts at national and regional levels; which in turn contributes to achieving global agenda such as the SDGs.

4 SIGNIFICANCE Capabilities accumulation has been at the centre of the economic growth of nations. This paper reviews the ideas/concepts, research, and literature on capabilities in order to provide fresh insight on the state of research on capabilities – including individual, organisational, and technological capabilities – at systems level in LMICs. By revisiting the research and literature on capabilities this research seeks to advance knowledge that contributes to LMICs ability to (1) better address their pressing socio-economic and development challenges, and (2) contribute to progress in achieving the SDGs. In our opinion, the SDGs will not be realised if the goals are not achieved in LMICs, particularly in African countries, China, India, and Latin America, where some of the poorest populations reside. Capabilities are needed in LMICs if the countries are to continue in, for example, their pathways towards sustainability transition, climate change mitigation and adaptation, or achieving the SDGs. This paper contributes in these regards

ABSTRACT. Introduction Recently, many policymakers and researchers have been made to consider public procurement as one of the instruments of innovation policy, especially in the advanced countries. Public procurement is utilized as innovation policy measures with the purchase of innovative products, services or technologies by the public sector. Public procurement has policy goals of not only the development of technology but also to the diffusion of innovative products, unlike traditional means of supply-based innovation such as R&D subsidies and tax benefits. Such public procurement for innovation (PPI) is recognized as a representative means of demand-side innovation policy. In the case of supply-side innovation policy measures such as R&D subsidies and tax benefits, the policy impact has been verified through various empirical studies. However, for demand-side innovation policy including PPI, the policy evaluation is largely limited to qualitative description based on case studies, and quantitative study based on enterprise and industry level data is insufficient. In this study, we try to quantitatively evaluate PPI policy by using the concept of additionality. PPI differs from policy objectives and effect paths according to the market and technology maturity level of the targeted goods or services. In other words, the evaluation of PPI should be made taking into consideration the characteristics of such purchasing objects. This study will evaluate the policy impact of pre-commercial PPI which induces the research and development of the company by concluding preemptive purchase contracts of products or services that have not yet been developed.

Research Background Edler and Georghiou (2007) defined PPI and recognized it as one of the main instruments of demand-side innovation policy. Demand has long been recognized as a major driving-force of innovation (von Hippel, 1986), and various studies introduced the case of several innovations which were introduced through demand from the public sector (Mowery and Rosenberg, 1979; Geroski, 1990). Although research on the policy impact of PPI has been steadily carried out, most studies are at the level of analyzing specific policy cases or making qualitative narratives through the theoretical background. However, policy evaluation of PPI needs to focus on various variables, as the impact is spread to private consumers, suppliers, other companies and industries through the diffusion of innovative products. In addition, since the purpose of PPI is different according to the characteristics of the purchasing object, analysis of the policy effect needs to be done differently according to the purchasing object. Innovation policy evaluation in quantitative manner has been concentrated on the analysis of supply-side innovation policy measures. Some studies including Karhunen and Huovari (2015) focused mainly on firms’ outcomes and proposed various methodologies to eliminate the effects of firm characteristics and macroeconomic environment and to examine the pure policy impact. The center of these methodologies is to capture additionality effect of the policy, which assumes a situation that could arise if the firm did not receive the policy when it actually did the policy program. Some studies including Czarnitzki et al. (2018) tried to capture additionality effect of PPI. However, these studies did not provide general implications for the performance of PPI, since the outcome variable was set without recognizing the various impact paths of PPI which are different from the types.

Research Objectives There are many studies on the taxonomy of PPI, but this study classifies them into four categories according to technology level and market maturity (Shin, 2017). This study focuses on the pre-commercial PPI that guarantees the purchase of products or services that have not been developed yet. Pre-commercial PPI lowers technology and demand uncertainty by tendering purchase contracts prior to the development of innovation, thereby increasing the company to invest on innovation activity. In particular, pre-commercial PPI is highly utilized in terms of linking R&D policy with innovation policy. It is necessary to assess the policy impact of PPI through not only technological input but also performance outcomes, such as sales performance and value added since PPI leads the diffusion of innovation. In addition, PPI affects individual agents in various ways within the innovation system, so changes in firms’ behavior by policy should also be considered. In this paper, we will also examine behavioural additionality (Busisseret et al., 1995) of PPI by considering dynamics of firms’ routine such as the change of asset structure and management and employment structure.

Data and Methodology This study utilizes National Science & Technology Information Service (NTIS) data, which is a collection of public research and development support information in Korea. The Korean government is linking R&D and PPI by the program which guarantees the purchase of project outcomes with R&D subsidies. This study regards this program as pre-commercial PPI and captures the firms which received this program. The purpose of this study is to analyze the policy impact of pre-commercial PPIs through the average treatment effect of treated sample (ATT) using propensity score matching (PSM) methodology widely used in public policy performance evaluation. The PSM methodology has been used to evaluate the impact of various public policies since Rosenbaum and Rubin (1983). Recently, the PSM methodology has been utilized in the policy evaluation of demand-side innovation policies including PPI. This study steps further by taking the difference for outcome variables between the matching year and the year after the policy when calculating the ATT. This methodology has recently been used in some innovation policy research under the name conditional difference-in-differences (CDID). This study will introduce various dependent variables to confirm not only input and output additionality but also the behavioral additionality of the firms. Variables such as R&D expenditure and R&D efficiency can be used as input variables, and value added can be added along with business performance indicators such as sales revenue and operation income as output variables. In addition, the behavioral additionality can be identified through variables such as changes in asset composition and management, and changes in labor composition.

References Buisseret, T. J., Cameron, H. M., & Georghiou, L. (1995). What difference does it make? Additionality in the public support of R&D in large firms. International Journal of Technology Management, 10(4-6), 587-600. Czarnitzki, D., Hünermund, P., & Moshgbar, N. (2018). Public procurement as policy instrument for innovation. ZEW-Centre for European Economic Research Discussion Paper, (18-001). Edler, J., & Georghiou, L. (2007). Public procurement and innovation—Resurrecting the demand side. Research Policy, 36(7), 949-963. Geroski, P. A. (1990). Innovation, technological opportunity, and market structure. Oxford economic papers, 42(3), 586-602. Karhunen, H., & Huovari, J. (2015). R&D subsidies and productivity in SMEs. Small Business Economics, 45(4), 805-823. Mowery, D., & Rosenberg, N. (1979). The influence of market demand upon innovation: a critical review of some recent empirical studies. Research Policy, 8(2), 102-153. Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70(1), 41-55. Shin, K. Y. (2017). Study on economy-wide impact assessment of public procurement for innovation for new emerging industry: A computable general equilibrium approach (Master Thesis). Seoul National University, Seoul, Korea. Von Hippel, E. (1986). Lead users: a source of novel product concepts. Management Science, 32(7), 791-805.

ABSTRACT. Grand challenge mitigation is framing many innovation policies and strategies worldwide (Mazzucato, 2018). In this context, Edquist and Zabala-Iturriagagoitia (2012) introduced the potential of public procurement for innovation as a relevant policy instrument aiding in grand societal challenge mitigation. According to the results of the ERAC consultation (see ERAC 1209/15), the latest estimations for public procurement expenditure on works, goods and services were close to €2.3 trillion per year, equalling 19.4% of European GDP. In particular, in Sweden public procurement amounts to SEK 683 billion per year (i.e. 17.5% of GDP). Public procurement thus represents a substantial portion of the EU economy and the economies of many countries around the world. However, the share of the whole procurement spending used to stimulate innovation remains insignificant, even if no comprehensive statistics exist as to date about this. Innovation-related procurement is acknowledged as a relevant policy instrument, particularly as a mission-oriented innovation policy instrument related to grand challenge mitigation, but in terms of its implementation and the mechanisms for its effective rolling out, it is still at its infancy. This is related to the discussions that are increasingly taking place in the academic realm as to the need to address the “implementation” of innovation policies (see Robinson and Mazzucato, 2018). The purpose of this paper is to analyse how public procurement can enable and be a driving force for innovations. The paper is about public procurement that requires or facilitates innovation. Therefore, an important task is to distinguish such procurement that leads to, or can lead to, innovation from such procurement that does not lead to innovation or prevents innovation. The paper argues that the most important way of achieving a higher degree of innovation in procurement is to conduct functional procurement (i.e. to formulate functional requirements in the procurement documentation). Therefore, the focus of this paper is on functional procurement, and thus it will not treat traditional/regular public procurement where one describes and purchases well-known products (product procurement) more than as a starting point (although this type of product procurement still constitutes the largest volume of all procurement spending). The important issue is how product procurement can be converted into functional procurement in order to enhance innovation. A large number of regular public procurements are perfunctorily conducted; the procuring agency or unit describes the same product as in previous procurements in a routine manner (i.e. path dependency and inertias) (Edquist 2014). These products must obviously be existing ones, since they can be described by the procuring organization. Often this description is quite - or even very - detailed. They may even be obsolete. If that is the case, qualitatively superior, products (i.e. innovations) may be excluded in the procurement process. A routine of simply describing the previously procured products makes it difficult or impossible for new products (innovations) to be accepted. This is a major obstacle to innovation in public procurement. Hence, product procurement generally does not lead to innovation. Of course, the procuring organization wants to buy products to use them for something. In fact, with the help of the product you usually want to solve a problem or get a need or function fulfilled. And this is (mostly) done in the interest of the citizens. An alternative to product procurement is that the contracting organization describes these problems, needs and functions in the procurement documents. When such a description exists, the term " functional procurement" is used in this paper. Functional procurement can be defined as the procurement of products by an authority/unit that describes a function to be performed or a problem to be solved (functional specification) instead of describing the product that is to perform the function (Edquist, 2017). In functional procurement, a public agency specifies what is to be achieved rather than how it is to be achieved. Functional regular procurement is pursued by means of functional specifications instead of product specifications. Hence, it is a matter of the manner in which a procurement call is set up and the tender documentation is formulated. Needs are translated into functions to which potential suppliers can respond. Needs are accurately identified and presented as requirements in terms that suppliers can respond to. It opens up for innovation but does not require it. Innovations are not excluded or disadvantaged. However, it should also be noted that a functional tender requires a process by which the need is identified, accurately specified and that potential suppliers are informed and engaged prior to the formal (functional) tender. Functional procurement is thus innovation-enhancing in the sense that it opens up for innovations, but it does not require innovations per se as it happens in other forms of public procurement for innovation, such as direct and catalytic public procurement (Edquist and Zabala-Iturriagagoitia, 2012). However, it does not necessarily have to lead to innovations - if the functional description includes existing products. What characterizes functional procurement is that the expected result is described as a function that must be fulfilled through the procurement. Functional descriptions may include old (existing) products or not. Including old products reduces the risk, but still opens for innovations. In addition, there are strong arguments for making the descriptions broad in order not to exclude unexpected innovations. The biggest difference when it comes to stimulating innovations is, however, the difference between product procurement and functional procurement as such. Product procurement can only exceptionally lead to the development of innovations. Functional procurement opens in principle all procurements for the development of new and better products. If you want to promote innovations through procurement, then functional specifications should be used as much as possible. Explaining and defining functional specifications rather than traditional descriptions of product/process characteristics is key to support innovation-enhancing procurement. Innovations may, of course, sometimes occur in regular product procurement, even if it was not a requirement of the procurement - if the product description is generic enough to include innovations (better products) that emerge anyway. One of the roles of innovation policy is, however, to create conditions and incentives for the systematic emergence and development of innovations that help address and respond to socioeconomic and environmental needs, both in the present times and in the future. From this perspective, innovations may be very much facilitated by functional specifications, as compared to product specifications. To achieve innovation through public procurement it is, somewhat paradoxically, more important to emphasize functional specification than to pursue innovation procurement. Functional specifications opens up for innovations in all types of public procurements, and not only those requiring innovations.

ABSTRACT. Background and Rationale Innovation does not occur automatically but thrives only under certain conditions. Some nations, including but by no means restricted to China, have adopted policies that can appropriately be labeled “innovation mercantilist.” These policies include forcing international companies to transfer technology to domestic firms or simply aiding and abetting its theft. Domestic firms may also be unfairly assisted by discriminatory tariffs, subsidies, public procurement, and national standards. While such policies may have short-term benefits for domestic firms and their customers at home and overseas, they are harmful for consumers and producers in the long-term. They reduce the returns from private and public investments in innovation industries globally, triggering reductions in such investments, stranding emerging technologies, locking-in the status quo, and slowing the pace of innovation as a whole.

Methods and Anticipated Results This paper will review the relevant literature on innovation mercantilism and provide a work-in-progress update on original case study research in two clean energy industries. The paper seeks to articulate a new framework for understanding the interactions of trade and innovation policies and provide preliminary empirical evidence at the industry level. The global transition to lower-carbon energy resources is creating opportunities to build new and rapidly growing export industries, estimated to be well into the hundreds of billions of dollars per year. These industries are therefore a major focus of trade and innovation policy-makers and thus appropriate cases for this project.

Relevant literatures focus on (1) interactions among nations in making trade and innovation policies, (2) strategic choices that firms must make when faced with tradeoffs between access to important markets and the risk of losing key innovation assets, and (3) the dynamics by which these tradeoffs are perceived by investors, including venture capitalists who seek to create disruptive start-up firms.

Innovation mercantilist policies can be distinguished from other policies through (1) their targeting of innovation industries with subsidies and trade barriers that benefit only domestic firms and (2) the use of coercion to extract innovation resources, such as intellectual property and know-how, from foreign firms. Sources of data about such policies include official documents of governments implementing such policies, other governments responding to them, and international organizations; testimonies and filings in legal cases; and interviews.

One case study on solar panel manufacturing will be retrospective. Production rapidly expanded in China in the 2010s, while innovative manufacturers in high-cost regions went bankrupt. The benefits of the shift for consumers who wanted to put solar panels on their roofs have been explored in depth, but the second-order, longer-term effects on the pace of innovation in solar energy products have been neglected. Low-cost crystalline-silicon solar panel technology has been frozen in place for a decade, while entrepreneurs who seek to develop more efficient panels based on new materials and methods have been left without resources.

A second case study will be prospective. The automotive industry, the world’s largest manufacturing sector, is entering a period of transformation, including the likely substitution of battery-based electric propulsion for internal combustion engines. China has stated in no uncertain terms that it seeks to be a global leader in the nascent electric vehicle (EVs) industry and has begun to apply its innovation mercantilist toolkit to the task. The prospect that EVs and batteries could replicate the path taken by solar panel manufacturing has alarmed many in the industry. New battery chemistries and alternative storage technologies may be stranded and current lithium-ion battery technology locked-in if the Chinese policy is carried through as intended.

Significance The project seeks to help focus research and policy discussions on important long-term issues for the world economy. While the specter of the Smoot-Hawley tariffs and the ensuing Great Depression hangs heavily over trade policy, not all forms of protectionism are equally significant. Trade and innovation policies, and the research communities that study them, have not been as tightly connected as they should be.

The project also seeks to direct attention toward multilateral action and institution-building. If each nation looks out only for its own interest, the global public goods that are generated by the global innovation system are put at risk. No single country is powerful enough or large enough to make the rules for the world economy, yet an anarchic system would be doomed to failure. The only pathway with the potential to create the conditions in which the global innovation system can thrive is through multilateral action and institution-building.

ABSTRACT. Purpose: To assess the challenge and opportunities of a new design of a monitoring and evaluation mechanism (MEM) of the SNCTA, one that would allow the System, considering the limitations of information, knowledge, institutional capacity, resources, and time, to verify its performance and efficiency in promoting technical change in the agricultural sector, and to improve its management.

Problem: For science technology and innovation activities, monitoring and evaluation (M & E) are fundamental activities to better know the process of knowledge generation, transfer, and adoption, to analyze the change generated on a population or target sector, and to support decision making and strategic direction.

In Colombia, the decision-making process that influences the development of the SNCTA is not supported by a systematic and precise information process derived from a monitoring and evaluation mechanism; one that assesses objectives and results of the science technology and innovation policy, and the activities in the agricultural sector.

The absence of an MEM allows for the existence of flaws in the planning of the policy and blocks the achievement of its objectives, harming the farmers and the general development of the sector. The above implies, in addition, high transaction and opportunity costs when making decisions and formulating strategies in an uninformed, incomplete, or wrong way.

Theoretical framework: This work is mainly based on literatura related with M&E mechanisms meant to guide a system through relevant information for decision making, aligning public policy or governance.

It is also based on the studies and results on New Public Management, Management for Results, Knowledge Utilization, Knowledge Management, National Innovation Systems, Program Theory, Theory of Constraints, and Participatory Monitoring and Evaluation, always from the idea of ​​having timely information about the detailed operation of the system in question for the illustrated decision making.

Methodology: For the development of this study we conducted interpretative analyses of documents for the understanding of the phenomenon and the characterization of its context, especially in aspects of regulations, public policy, and sectoral demands. We also developed non-standardized data collection exercises such as focus groups and interviews with experts with the purpose of validating hypotheses and a better understanding of the problem posed. The abductive method and the prospective public policy research, were used to create new ideas outside the standardized knowledge from the very conditions of the object studied. These were based on the researchers practical experience on existing M&E mechanisms operating in Colombia and academic work on the topic.

Results: The main result of the study includes the development of a new logical model of the SNCTA and its M & E process, supported on the approaches of the Innovation Systems and the Theory of the Program. These were based on the understanding of the role of the actors, relationships, processes, activities, and "bottlenecks", which practically constitute sub-systems with their own dynamics.

As a public policy proposal, we presented the characteristics that this MEM must have in order to orient the System towards technical change through "enlightened decision making", as well as some considerations and recommendations to carry out its implementation.

Conclusions: This new look at the SNCTA, from the evolution of knowledge management approaches, constitutes a transformative element that is worth exploring further in order to have national and territorial systems of innovation directed and coordinated towards the achievement of their goals. This view could be replicated in other subsystems of the National System of Competitiveness, Science, Technology, and Innovation (SNCCTI, by its initials in Spanish).

This scenario allows us to overcome traditional methods such as management by results, which vision falls short when identifying the causes of unexpected results and allowing them to be solved with the exercise of M & E.

This will imply an arduous work of the teams that today lead these processes of change of the policy and the management of the System; and it is hoped that it will become one of the strategies guiding the regulation of the SNIA. For this it is necessary that this type of actions stop being formulated in isolation, they must be integrating and co-responsible with the sectors, actors, and territories, especially in what refers to the agricultural sector which dynamics must be defined from the regions.

Bibliographic references:

ARNOLD, E. (2004). Evaluating research and innovation policy: a systems world needs systems evaluations. Published in Research Evaluation. Volume 13. Number 1. April 2004. Pp. 3-17. Retrieved June 2016 at: http://www.scienceofsciencepolicy.net/sites/default/files/attachments/2005-11.Arnold.Evaluation.Research-Innovation.Policy.pdf

EDQUIST, C. (2005). Systems of innovation. Perspectives and challenges. Published in The Oxford Handbook of innovation. Oxford University Press 2005. Editors Jan Fagerberg, David C. Mowery, Richard R. Nelson. Pp 181-208.

HEKKERT M. P. et al. (2006). Functions of innovation systems: A new approach for analyzing technological change. Published in Technological Forecasting & Social Change. 74 (2007). Pp. 413-432. Retrieved July 2016 at: http://www.transitiepraktijk.nl/files/Hekkert_et_al_2007_%20functions_of_%20innovation_systems.pdf

LIGERO, J. A. (2011). "Two methods of evaluation: criteria and program theory". Work documents. CECOD series. Number 15. 2011. CEU Editions. Retrieved August 2016 at: http://www.cecod.org/LinkClick.aspx?fileticket=VSYEl2XAhNY%3d&tabid=862&language=en-GB

ORDÓÑEZ-MATAMOROS, G. (2013). Manual of analysis and design of public policies. Bogotá: Externado University of Colombia.

URIBE, C. et al. (2011). Sowing Innovation for the competitiveness of the Colombian agricultural sector. Bogotá: Ministry of Agriculture and Rural Development and National University of Colombia.

ABSTRACT. Context and purpose:

The Colombian agriculture sector is one of the main engines for competitiveness, productivity and economic development. In 2017, the participation of the sector in the global gross domestic product indicator (GDP)- of the country was 6.31% from 309.2 billion USD. Science and technology activities in the sector achieved participation levels of 0.2% of the national GDP and 0.79% of sectorial GDP (19.51 billion USD). This evidence the need to strengthen the management structure, for a more efficient resources allocation to promote technological change.

The dynamics of research, technological development and innovation (R&D&I) activities as a main active in the National Innovation Systems, is supported on public policy instruments and the organizational structures that integrate its stakeholders. These activities focus on generation, transfer, adoption, and evaluation of technological offer (knowledge, technologies and technological services).

The aim of this study was the design of an integrative management approach for the National Agriculture Innovation System (NAIS) in Colombia based on three key elements: The Agriculture Science and Innovation Policy (ASTI), previous work on the area, and the main management instruments of Science, Technology and Innovation (STI) in Colombia. Previous work comprises the interaction analysis of the main ASTI policy instruments, the conceptual approach to national, regional and sectoral innovation systems, and the first relational structure for STI policy instruments.

Methods and material:

Based on the results of previous studies that establish a circular approach for the NAIS, a sequential methodological scheme was constructed based on its main instruments, which include: the sectoral strategic plan (PECTIA), the survey of sectorial evaluation, the Agenda for R&D&I, and the sectorial knowledge management system (SIEMBRA).

This scheme embraces four stages, each one with specific methods and activities. First, the 5w+h questions approach is used to characterize the main instruments in the NAIS policy. Second, an analogy of dynamic systems of upstream and downstream information is used to represent the interaction between main instruments. Third, the general theory of systems is used to represent the instruments interaction across the organic components of the NAIS (sub-systems). Fourth, the Colombian Technical Norm for innovation management -NTC-5801 is used to standardize a structural integrative approach for policy instruments, decisions instances, and key stakeholders.

Results:

Five main policy instruments were characterized including the national agriculture innovation policy, as well its information feedback relations. This characterization approach could be used for any other support instrument (current or future). Based on this approach a systemic diagram is designed, to represent the main components of the NAIS: the research and technological development subsystem, the innovation capacities for education and training subsystem, and the rural extension and technical assistance subsystem.

This diagram includes not only the interaction between subsystems but also in the national innovation system and among the main instruments. The systemic representation model is used then as a reference point to standardize the NAIS. The standardize approach comprise an integrative structure of subsystems, instruments, instances, and actors based on the NTC-5801 innovation management processes. These processes for the NAIS are: Knowledge management based on information systems, governance and strategic decisions, resources and capabilities, R&D&I activities and monitoring and assessment strategies.

Finally, an equivalence analysis is used to describe the management processes for agriculture innovation, a matrixial scheme for sub-systems characterization is proposed and a matrixial scheme for sub-systems, instruments, instances, and stakeholders is designed. These products converge as the first approach of a management model for the NAIS.

Significance

This research is an initiative to support the regulation, operation, and appropriation of the NAIS Law. It also contributes to promoting technological change in the agriculture sector in Colombia.

ABSTRACT. Conspicuous change is seen in less than a decade since 2008 in global higher education environment driven greatly by developments in the Asian universities, particularly India, China and Japan. With steep rise in number of universities and increasing awareness on research initiatives, India is poised to contribute significantly to academia spawned innovations. This is aligned with technology-based economic development initiatives which emphasize on stimulating technological entrepreneurship in universities via start-up creation, incubation units, patenting, licensing, and university–industry partnerships. Many studies suggest such developments as ‘academic entrepreneurship’ or university’s third mission in contributing to the national system of innovation.

Linking to the Indian Context

In India, innovation management has been characterized by creation and adaption to a range of organizational and institutional developments particularly in the last decade. Though, the outcomes have not been very promising as compared to academic institutes of good global standing; knowledge production and its effective transfer across 1050 odd universities in India has the potential to capitalize significantly. The innovation capability is still limited to very few universities and dominated by institutes of national importance. This paper is an attempt to understand and document the progress in research output of Indian universities, highlight the stellar performances, and suggest policy measures. An emerging global economy, India does need an environment for research to critically focus on creating an ecosystem that would contribute significantly to the system of innovation. Universities across the globe play a critical role in building an innovation ecosystem that meaningfully contributes to creation of new knowledge and its dissemination. This is in accordance with Nelson (1993), Lundvall (2004), Mowery and Sampat (2005), and others who observe that universities play a crucial role in technical advance, and that a growing number of both industrial and developing economy governments, seek to use universities as instruments for knowledge based economic development and change. In India, similar views have been echoed by Basant and Chandra (2006), Chandra and Krishna (2010) among others.

Data and Methodology

The paper attempts to examine the determinants of different means of knowledge creation and transfer in academia, focusing on their institutional mechanisms, and core parameters such as research publications, intellectual property such as patents and new firm formation. Data and information relevant to building innovation capability is searched through scientometric indicators of universities in two international databases: Clarivate Analysis’ Web of Science and Elsevier’s Scopus as also through official patent websites, annual reports and other official authentic sources.

Research and Preliminary Findings

A comprehensive exercise was conducted that involved listing of the total universe of 1047 degree awarding institutes in India up to December 2018. Around 35 per cent of the total universities have been set up in the last five years. The initial analysis shows that top 10 per cent or 100 universities have a total of around 61700 publications per year during 2016-18 in Web of Science index. This accounts for around 71 per cent of total annual publication count of 86800 in the same period. It also shows a significant increase from 64000 articles by 600 Indian universities with at least one publication reported in 2015. Similarly, in Scopus database, the top 100 universities (roughly ten percent) accounted for an average of 85100 publications annually during 2016-2018 or nearly 68 per cent of total publication adding to 124250. The top five institutes produced an average of 2300 publications as listed in Web of Science in the year 2018. Other indicators of quality research such as average citation, h-index and number of articles per author for top 100 universities during 2016-18 were in the range of (5.8 to 6.1), (25 to 27) and (2.6 to 3.5) respectively. However, on the other side, the study also shows 50 percent universities in 2018 had less than 50 papers that got reflected in either Web of Science Expanded or Scopus Index. There are nearly 22 percent universities which do not have a single publication listed against their name.

The findings also highlight the performance of universities in terms of their intellectual property, industrial consultancy, sponsored research, incubation units and start-up firms. The research performance in these functionalities, however are limited to very few institutes.

The Rationale and Relevance As there is no systematic study on universities in India and other developing countries that document the state of affairs in research activities; this paper attempts to do the same. While the database has an elaborate classification of universities, their geography, year of establishment and such details; the key focus is to link the universities as contributors to the innovation system globally, emphasizing on policy initiatives and the need to create institutional mechanisms that foster innovations in an academic setting.

Objective and Future Direction

A majority of the universities in the US and Europe have experienced the nuances of academia-industry interface in a wider perspective, so as to understand the ways and means of fostering a productive relationship between the two entities. The primary focus of this paper is to study the interface modes linked to academia spawned innovations in select European and American universities and juxtapose the related issues with a few other universities in India and possibly in the developing countries. A key future novelty will be the concurrent focus on different modes for knowledge transfer driven by innovation policy framework practiced globally. In a rapidly changing external environment, for many universities, restructuring of the academic organization is a key challenge. It is equally important to engage in a systemic framework that promotes grassroots innovations and start-ups which is inclusive and involves people from all strata of society. This is a promising area of extending the findings of this study.

The structure of the paper would comprise an overview of the university system globally in relation to innovation eco-system; a section on the significance of innovation management including policy and relevant literature review that draws largely from the theoretical framework of ‘National Systems of Innovation’. The findings discuss the empirical results of select universities.

References:

Basant, R. and Chandra. P. 2007, "Role of Educational and R&D Institutions in City Clusters, "An Exploratory Study of Bangalore and Pune Regions, "World Development, V 35, No 6: 1037-55. Chandra, N. and Krishna, V.V. 2010. "Academia-industry links: modes of knowledge transfer at the Indian Institutes of Technology". International Journal of Technology Transfer and Commercialisation, 9 (1-2): 53-76. Lundvall, B.-Å., 2002. The University in the Learning Economy. DRUID Working Papers, No. 6. Mowery, D. C. and Sampat, B. N. 2004. Universities in National Innovation Systems in: Fagerberg, J., Mowery, D. C. and Nelson, R. R. (Eds.), The Oxford Handbook of Innovation. OUP, Oxford, 209-239. Nelson, R. R., (Ed.), 1993. National Innovation Systems: A Comparative Study. OUP, Oxford. Siegel D. S., Waldman D. A., Link A. N. (2003), ‘Assessing the impact of organizational practices in the productivity of university technology transfer offices: an exploratory study’, Research Policy 32 (1), 27–48.

ABSTRACT. Reasonable compensation of indirect cost(IDC) plays an important role in sustainable development of research universities and institutions. Nowadays, block grant and competitive project funding are two main funding sources of research universities and institutions. IDC is the compensation for public service expenditure from institution to projects. Under-compensation results in project funding crowding out block grant funding, which affect the sustainable development of research universities and intuitions. Over-compensation reduces the efficiency of project funding. In the long term will influence the sustainable development of competitive project funding mode. In practice Completely Differentiated Compensation Model (CDCM) and Non-Differentiated Compensation Model(NDCM) are two typical IDC compensation model. However, both of these models have their own weaknesses. CDCM has the problems of high management expenditure and low system efficiency. NDCM cannot effectively exerts compensation and incentive role due to under-compensation or over-compensation. This study aims to propose a novel IDC compensation model to solve the above problems. Firstly, based on the case study and principle agent analysis of government scientific research funding system, an optimized IDC compensation path is proposed. Secondly, based on the review of the concept and the compensation process of IDC, 9 potential institutional characteristics affecting the level of IDC are proposed. Following that, Through the empirical analysis of 91 universities in the United States, 5 significant institutional characteristics are verified. Finally, A novel IDC compensation model with classification through institutional characteristics as core to keep institutional sustainability is proposed.

ABSTRACT. The Success of Start-Ups that Participate in Business Incubators and Accelerators

This study investigates how participating in a business incubator or accelerator influences the success of nascent high technology firms. The emergence of incubators and accelerators as drivers of entrepreneurship has garnered the attention of entrepreneurs and policymakers alike since they provide a much-needed source of infrastructure to support new firm development. Nascent firms are important for economic health and innovation, which influence the competitiveness of countries, regions, and even cities (Audretsch and Keilbach 2007). Private and university-led incubators and accelerators attempt to improve the success of nascent firms (Mian et al., 2016) by “buffering” firms from environmental threats and “bridging” via network relationships to build legitimacy and resources (Amezcua et al., 2013; Eveleens et al., 2017). Few studies have looked at the long-term outcomes associated with participation (for exceptions see Schwartz, 2009; Mian et al., 2016). This study develops a deeper understanding by building an empirical link between different programs and participants’ outcomes. We ask if the participation in an incubator or accelerator improves a firm’s success. We compare three outcomes important for nascent firms: closure due to bankruptcy, asset sale, or performance, raising venture capital (VC), and obtaining government grants.

BACKGROUND Private business incubators provide young companies with office space and basic business services including shared administrative support (Bollingtoft & Ulhio, 2005; Grimaldi & Grandi, 2005) and some provide lab space (Feldman & Francis, 2003). These tend to be unstructured with few admission requirements. About 800 private incubators in the U.S. in 2012 (NBIA, 2013). Private accelerators have a competitive selection process and usually make seed stage investments in exchange for firm equity. They offer services and resources beyond incubators such as structured programming (Pauwels et al., 2016), business assistance, mentoring, access to capital, and business networks (Mian, 1996; Cohen, 2013). Accelerators “make seed-stage investments in promising companies in exchange for equity. Over 175 private accelerator programs existed in the U.S. in 2016 (Gust, 2017). University incubators are sponsored by a college or university and are like private accelerators (Mian, 1996; Grimaldi & Grandi, 2005) as participation is often competitive and may have a structured program with access to courses, advising and/or mentoring. Some university programs require a university affiliation. University incubators may also provide access to faculty advisors, student employees, libraries, technology transfer services, and R&D labs (Quintas et al., 1992; Colombo & Delmastro, 2002). Participating firms also benefit by gaining legitimacy through their affiliation with a university (Mian 1996; Rothaermel & Thursby, 2005). Recently, some university incubators have evolved into an accelerator model by including funding and formal training. In 2012, about 400 university incubators existed in the U.S. (Powell, 2017). These programs increase the creation of new firms by lowering barriers to entry and reducing the hazard of exit (Schwartz, 2009; Mian et al., 2016; Cohen et al., 2018). Further, these programs offer proximity to other entrepreneurs, which increases the likelihood that participants will exchange useful information. Thus, the tangible and intangible resources provided by these programs support participants’ success (McAdam & McAdam, 2008; Mian et al., 2017; Patton et al., 2009). Hypothesis 1a: Private business incubator participants are more likely to succeed than other firms. Hypothesis 1b: University incubator participants are more likely to succeed than other firms. Hypothesis 1c: Private accelerator participants are more likely to succeed than other firms. The knowledge-based view suggests that when firms access a broader resource pool the entrepreneurs increase their network and learning opportunities, gaining relevant and novel information to facilitate innovation and growth. For example, firms originating from universities may benefit more from participating in a private incubator than in a university-based incubator where they likely encounter familiar resources. Likewise, firms that participate in more than one venture development program may be able to access a greater variety of resources than other firms. Thus, we propose: Hypothesis 2: Participation in a start-up accelerator or university incubator improves a firm’s likelihood of success more than participation in a private incubator. Hypothesis 3: Firms that participate in an incubator or accelerator with a dissimilar foundation than the founders’ occupational background will have a higher likelihood of success than those that participate in incubators with similar foundations.

METHODS The study uses a database on U.S. nanotechnology firms started between 1997 and 2012 to compare the participation of venture development programs. To identify start-ups with nanotechnology capabilities, we examined over 10,000 pages of industry lists, directories, press releases, university websites, scientific publications, and websites. We focused on firms within 60 miles of an incubator or accelerator program and on industries most prevalent in accelerators and incubators to concentrate on those industries most generalizable to incubator participation. The final sample was of 421 firms. Demographic and outcome data were collected for each firm including location, industry, and accelerator and incubator participation. The data were triangulated across multiple sources. Background data regarding each founder’s education and employment were collected from databases. The models control for firm-level and environmental-level variables that can influence firm outcomes such as year of founding, if founded by a team, and industry. We used event history analysis via STATA with maximum likelihood estimation and robust standard errors, clustered by firm with a Weibull distribution (Blossfeld & Rohwer, 2002).

FINDINGS & DISCUSSION Over one-third of nanotechnology firms founded between 1997 and 2012 participated in an incubator or accelerator program. Participation was split evenly between private incubators (16 percent) and university incubators (18 percent). Approximately three percent of firms participated in accelerator programs. We find that venture development programs do help firms succeed, but in contrasting ways. In summary, firms that participated in private incubators were more likely to obtain VC funding and government grants than firms that did not participate in a program, which support Hypothesis 1a. These results are interesting since accelerators offer as many if not more resources and programs than private incubators. Although this may be due to the experiences of the founders while in the incubators, it may also be an artifact of the selection criteria for participation. Specifically, incubators may target firms that require funding or are attractive to investors. Firms that participated in university incubators were more likely to obtain a government SBIR or STTR grant and were less likely to close due to unfavorable performance than other firms, supporting Hypothesis 1b. Again, these findings may be due to the selection criteria of the university program that may focus on firms that fit with the government grant program needs. There was little support for Hypothesis 2 that diversifying resources by participating in a program dissimilar to the founders’ backgrounds increased a firm’s likelihood of success. This study provides unique insight into the value of incubator and accelerator programs. The success of nascent firms is a goal not only for their founders, but also for policy makers who strive to improve the economic and technological competitiveness of their location. We show that these programs indeed influence the success of participating firms in differing ways. This study is a first step in understanding the heterogeneity in these programs and the necessary unpacking of how the underlying mechanisms play a role in economic development.

ABSTRACT. Business leaders and policy makers stress the importance of stimulating entrepreneurial activity for the continued vitality of the US economy, but many policies to promote entrepreneurship have had mixed effectiveness. One reason is the lack of exposure to the processes of innovation and starting a new firm. This research uses the idea that television is often credited for impacting socio-economic outcomes and attitudes, to assess if a televised business plan competition (in this case, the ABC show Shark Tank) can shift opinions about the willingness and desirability of starting a business.

We measure exposure to the show by means of Nielsen ratings. Our outcomes come from a range of sources including the US Small Business Administration (i.e. counseling and training counts), National Establishment Time Series (NETS -- i.e. counts of new businesses), and the US Patent and Trademark Office (i.e. patent application counts). Within-market variation in the show's popularity over time suggests that exposure to Shark Tank may indeed impact entrepreneurial activity.

While our findings are preliminary, they appear to be consistent with prior (economics) literature on media influences (e.g. DellaVigna and La Ferrara 2016). As far as we know, we are one of the first to explore the impact of the media on entrepreneurship and innovation.

This work is supported by a SciSIP grant from the National Science Foundation, Award number 1664383.

ABSTRACT. Introduction & Background Surprisingly little research examines how financing that supports entrepreneurial businesses promotes regional growth and influences industry emergence. Studies tend to examine one program in isolation. Yet the combined impact on a regional economy—the sum of effects on individual firms—remains unexplored. There is general consensus that the social rate of return to public R&D investment is high (Toole 2012), which is used as a justification for government funding. A related stream of literature considers the degree to which public funding affects private funding. Empirical evidence suggests government R&D subsidies lead to an additionality effect and induce private, firm-level investment. The literature is relatively silent on the relationship of government financing to other sources of financing at the aggregate regional level.

In complex systems theory there is a property called emergence, attributed to George Lewes (1875), which means the action of the whole is more than the sum of the actions of the parts. Emergence is a socially defined, agent-based result of myriad small efforts. Accordingly, this paper argues an emerging entrepreneurial ecosystem can be considered a result of myriad small efforts of ecosystem agents. Being one such agent, the steady hand of government funding creates opportunities for entrepreneurs. As entrepreneurs create new firms, the private sector has a vehicle to make investments. Through this type of public-private interaction an industry emerges.

This paper extends the literatures on entrepreneurial ecosystem building and industry emergence, as well as public-private funding interactions and R&D funding policy. We analyze the development of one industry in one region over a long-time horizon, focusing on the interacting roles of state and federal public funding and private funding of new firms in order to explain emergence. Specifically, we ask how the interplay of these three sources of funding influenced the emergence of the Research Triangle region’s life sciences industry.

Empirical Context North Carolina's Research Triangle region's life sciences cluster is one of the largest in the country, anchored by the region's three research universities, a long history of pharmaceutical branch plant location, and a large number of entrepreneurial firms. The origins of the cluster can be traced to Research Triangle Park’s 1958 establishment—the result of a collaborative effort involving politicians, academics, and financiers to alter the industrial and competitive basis of the region (Leyden & Link 2011). Yet, unlike Cambridge MA and the San Francisco Bay area—the country's leading life sciences regions—the Research Triangle region was not an obvious candidate to develop a life sciences industry, making it interesting to study emergence.

We define the Research Triangle’s life sciences industry broadly and focus only on entrepreneurial firms. We categorize firms into the following sectors: dedicated biotechnology, human therapeutics, diagnostics, medical devices, biomaterials, health IT, and services. While the 1980s saw in increase in public support for the industry, it was not until the 1990s that the number of start-ups began to noticeably increase. Over time, the region has slowly nurtured an entrepreneurial ecosystem, thanks to mergers and layoffs from high-profile multinationals, a more aggressive technology transfer stance from the universities, and the development of a plethora of support institutions.

Data, Sample, & Methods Firm data is obtained from the PLACE: Research Triangle Database, which contains information on the universe of life sciences firms founded in the Research Triangle region (Feldman & Lowe 2015). State-level funding data was obtained directly from its sources—the NC Biotechnology Center, a quasi-public entity dedicated to promoting the life sciences in North Carolina, and NC IDEA, a nonprofit that uses State funds to provide seed grants to new ventures. Data on private funding was gathered from CB Insights. Small Business Administration awards (SBIR and STTR), combined with NIH awards measure federal funding. The sample is limited to the universe of 670 firms founded between 1983 and 2012. Of these firms, 147 received national funding, 107 received state, and 117 received private.

The paper employs mixed methods to examine industry emergence, beginning with an historical analysis of the cluster’s development. An important question when considering the interplay between public and private funding of start-ups is whether public funds act as a signal for private investors or vice-versa. In order to investigate this question, we apply Granger causality tests for each sector (Dumitrescu & Hurlin 2012). Granger-causality tests demonstrate predictive, though not truly causal, relationships between funding sources. We next use discrete event history analysis, testing multiple hazard and frailty specifications, to investigate how the variety of multi-level public and private funding influences ecosystem emergence through firm survival. This improves upon prior literature that often invokes simplifications to implement continuous survival methods.

Preliminary Results Granger causality tests indicate relationships between funding sources are highly differentiable across sectors. The human therapeutics and medical devices sectors exhibit the most predictive relationships. For human therapeutics, federal and state funding evolve together, while federal funding predicts private. This sector deals with a high degree of uncertainty due to the nature of drug development. For medical devices, which are easier to get to market, we see state funding predicts federal funding, while the state-to-private relationship is mutually predictive and federal funding predicts private.

Event history results show that firms that received federal funding experience less chance of failure. The hazard ratio is 0.971, meaning federal funding decreases the probability of failure by about 2.9%. While statistically significant, this effect size is economically small. Still, it supports the argument that federal funding plays a role in industry emergence. We also find that firms that received private funding experience less chance of failure (3.7%), and the effect is non-linear. This result confirms prior literature on the importance of VC and other private funding to emerging ecosystems. We examine variations on the functional forms of the state, federal, and private funding variables and these results confirm the previous ones. Furthermore, firms that receive funding from a variety of sources experience the largest decrease in probability of firm failure.

Discussion & Significance These results shed light on the process by which new industries and ecosystems emerge in a regional context. Descriptive findings uncover wide heterogeneity in the way firms are funded by national, state, and private funding sources based on sector. Granger causality tests confirm sector level differences in the relationship between the multi-level public and private funding. Survival analysis provides evidence that the mix of multi-level public and private funding of these firm matters for firm survival, and therefore ecosystem emergence. Notably our results show the importance of having a strong base of federal funding and suggest the State may be funding more high-risk firms. Though federal funding provides fewer dollars to firms, it funds the greatest number of firms in the region. Private funding, which provides substantially more funds in terms of dollar amount than both public sources combined, also decreases the chances of firm failure.

This paper contributes to the literatures on entrepreneurial ecosystems, public and private funding of innovation, and regional development. Furthermore, findings offer practical insights for policy makers and business strategists. Empirical demonstration of the interrelations existing between multi-level public and private funding sources that support economic growth will provide a better understanding of the innovation systems policy makers often aim to strengthen.

ABSTRACT. Background and Rationale

The supply and demand mismatch within the academic job market is augmented by a growing complaint that trainees are not prepared for a range of careers beyond the academic. A growing literature underscores that career opportunities and interests of doctoral trainees are broad, with doctoral graduates in STEL having increased  interest in non-academic careers, both research and non-research career pathways.  In fact, 39 percent of PhD recipients in the biological and other life sciences are employed in four-year academic institutions, down from 42 percent in 2013 (U.S. National Science Board, 2016).

What career resources do doctoral candidates have to support them in pursuing broad career interests? The reality is that university career development resources are typically geared to the undergraduate career placement needs, and doctoral training is mostly intended to prepare specifically for tenured academic careers. Further, an individual’s doctoral advisor or postdoctoral supervisor may be the most direct support for trainees. Yet, faculty do not necessarily have the exposure, knowledge, or connections to facilitate the training of doctoral students for broader career pathways, tending to provide advising and mentoring that align with academic careers. This can ultimately leave trainees who have broader career interests on their own to locate the resources they need to identify the career that they want, and also have the skills to pursue and succeed in that career.  Yet, the career development strategies that they pursue and the preferences that they have for different types of career development resources is not well understood.

Our interest is in how the support of key relationships in the doctoral process, including both peers and doctoral advisors (whether or not they can assist with non-academic career preparation) mattes in the strategies that trainees use to pursue career development resources. We draw from existing higher education and social capital theory literatures to characterize how trainee preferences for career development resources are shaped by the career support received from their Principal Investigator (PI) and peers, as well as their own career search self-efficacy.

Methods

Our study is based on two data collection efforts that were part of the assessment of NIH BEST programs at four U.S. academic institutions. To address the strategies that doctoral trainees have used to seek out career development resources, we examined survey data that was designed to examine trainee career interests and access to career development resources prior to their engagement with the BEST program. Using a survey of both BEST and non-BEST trainees (those not formally in a BEST program), we conducted descriptive and logistic regression analyses of survey data to assess the factors affecting trainee preferences for three different types of career development models: an intensive cohort career development experience (BEST “cohort”), ad-hoc resources (“cafeteria”), or choosing not to seek any career development resources at all. A second aspect of this study examined the specific resource preferences and interventions that trainees found valuable. This aspect of the project involved a series of semi-structured interviews with trainees at three institutions. Qualitative interview data were transcribed and thematically coded using NVivo.

Results

For our first question on factors that influence trainee preferences for career development resources, we find that social capital in the form of a supportive environment and peer support was critical for shaping career development preferences. Cohort programs were particularly attractive to trainees interested in careers outside of academia and who had low career self-efficacy. Trainees who reported high levels of PI support were less likely to pursue other career development resources, while students reporting low levels of PI support were more likely to choose to participate in a career development focused BEST cohort community. Trainees who reported low levels of PI, department, and peer support were less likely to participate in formal career development events or resources offered by academic institutions.

For our second question regarding specific resources utility and value, we find a variety of resources that were helpful to trainees in their career development. Here, trainees point to traditional resources such as career development workshops, especially those that were informing trainees of different career options, which were reported as the most beneficial resource to the trainees’ development. Most notably however is that we find that psychosocial support from mentors and peers matters for trainee career development. We were able to tease out the role that peer and mentor support plays, finding, as with the first analysis, that peer and mentor support play different roles in career development. From the interviews, we found that the helpfulness of peer and mentor support were associated with different types of career development barriers and had different effects on the extent of the trainees’ confidence changes over their time in BEST.

Significance

This study contributes to the doctoral trainee research base by: 1) taking a quantitative approach to cohort based interventions for career development, concepts historically largely examined by qualitative methods; 2) distinguishing among the types and sources of support to better tease out the different types of relationships trainees may have;  3) identifying these issues for both the experiences of the doctoral student and the less-studied postdoctoral fellow; and 4) moving beyond a single institution study context by examining data from three different university programs, which allows us to control for institutional and demographic characteristics which importantly is recognized as a significant need in cohort model research.

ABSTRACT. Background and Rationale: How long foreign nationals who obtain their PhDs from U.S. educational institutions remain in the U.S. has implications for the S&E workforce and the state of the U.S. S&E enterprise. The National Science Foundation has funded research for two decades on the proportion of these individuals staying in the U.S. for 5, 10 and 16 years after graduation. “Stay rates” vary by country of citizenship at time of graduation and degree field. Additionally, for those graduating in 2002 and 2005, estimated stay rates based on the rank of the academic departments from which they obtained their PhDs have been developed. This analysis was done to gain insight into how institutional rankings impact the stay rates of their doctoral recipients. This presentation will focus on findings related to differences in stay rates by the “top-ranked vs. all other departments” for the 2002 and 2005 cohorts.

Method: The stay rate estimates were derived by assembling groups of Social Security Numbers of foreign doctorate recipients from a census survey, the Survey of Earned Doctorates* and obtaining a special tabulation of data from the Social Security Administration. If a foreign doctorate recipient earned $5,500 or more, paying taxes on these earnings for the year(s) specified, he or she was defined as a stayer. For the department rankings, we used rankings published by U.S. News and World Report and also by the National Research Council, each of which uses a survey to measure the research reputations of the faculty within a department. For each of nine different discipline groupings, a doctorate recipient was designated as having graduated from a top-rated program if either one of these two separate ranking sources listed their program among the top 20.

Results: Looking across both cohorts, without regard to country of citizenship or degree field, there is clearly a lower “stay rate” for those graduating from top departments. However, we show that the top-rated programs as a group have a more diverse mix of doctoral recipients by country of citizenship. While 38 percent of their doctoral recipients come from high stay rate countries (stay rates above 75%, mostly China and India), compared with only 30 percent from the low stay rate countries (stay rates below 45%), the programs that are not top-rated rely even more heavily on students from the high stay rate countries. Taking country of citizenship at time of degree into account explains a significant portion of the differences in stay rates seen overall for top-rated versus all other departments. Additionally, we look a stay rates by discipline in top-rated versus all other departments, finding that controlling for country of citizenship within disciplines again mitigates the lower stay rates seen from top-ranked departments compared to all other departments.

Significance: Understanding the influence of foreign nationals who are U.S.-educated S&E PhDs on the U.S. S&E workforce is critical to guide investment and policy decisions related to the health of science and innovation in the U.S. and around the world. Our new findings from the 2005 cohort confirm earlier findings for the 2002 cohort. Our research using administrative data has provided a consistent method for estimation of stay rates by key demographics to allow for a time series of data to improve our understanding of what the future may hold for this important population.

*The Survey of Earned Doctorates is sponsored by the National Center for Science and Engineering Statistics (NCSES) within the National Science Foundation (NSF) and by five other federal agencies: the National Institutes of Health, Department of Education, Department of Agriculture, National Endowment for the Humanities, and National Aeronautics and Space Administration.

ABSTRACT. Background and rationale

Numerous studies analyse and discuss how the conditions of doctoral education and training have been altered by reform processes introduced to improve its quality and efficiency. Most of these studies address the introduction of structured doctoral programmes and suggest a convergence of doctoral training across countries with regard to structure and organisation. In contrast, considerably less attention has been devoted to analyse the role and purpose of the doctorate as an academic degree which qualifies for careers in and outside the academic sector. This in particular applies to Germany which is characterised by a high share of graduates transitioning into doctoral training – especially in fields like chemistry, biology, and medicine – and a high share of doctorate holders employed outside academia as compared to international standards. While in the United States, United Kingdom, and Australia doctoral degrees e.g. in terms of research, professional, and industrial doctorates are formally codified – reflecting that doctoral education does not exclusively serve the reproduction of scientific workforce – no such distinction does currently exist in the German higher education and science system. Numerous studies from English-speaking countries argue that the emergence and growth of professional doctorates are the outcome of a complex interplay of factors such as the employability of doctorate holders in non-academic labour markets, the critique of the research doctorate, the rise of the knowledge economy and changing role of higher education. However, attempts to delineate professional from research doctorates respectively to develop definitions of professional doctorates based on the analysis of programmes have produced ambiguous results. Hence, it is a highly debatable point whether the ideal-typical dichotomous distinction between the research doctorate that serves as an entry ticket for an academic career and the professional doctorate that qualifies professionals for doing research or reflecting their professional practice, provides a useful analytical background for describing and explaining trends in the functional differentiation of the doctorate on an empirical basis.

Research questions

In this paper I explore differentiation in the role of the doctorate by using the example of Germany. Thereby I especially focus on the following questions: - What are the aims und purposes of doctoral education in Germany and how do these become manifest in actual practices of doing a doctorate (e.g. in terms of motivation and career goals, nature and source of the topic of the doctorate, cooperation practices and context of research)? - How is differentiation in doctoral degrees discussed in German science and higher education policy? - To which extent do current practices of doing a doctorate in Germany reflect or challenge the common dichotomous conception of a rather professional-oriented vs. a research-oriented doctorate? - Which dimensions are appropriate to describe functional differentiation of the doctorate in the German higher education system?

Methodological approach

In order to empirically explore functional differentiation in the role of the doctorate in the German Higher Education System, I combine two different methodological approaches: Firstly, I use thematic content analysis to study the body of position papers and recommendations from stakeholders in the German science and higher education system (e.g. German Rectors’ Conference, German Council of Science and Humanities, National Academy of Science and Engineering) in order to analyse whether and based on which arguments those demand for a differentiation of doctoral degrees and identify potential science-internal drivers of differentiation. Secondly, I use survey data from doctorate candidates in order to analyse how current practices of doing a doctorate suggest that a functional differentiation is already in place. Therefore I draw on data of a large-scale pilot study of the German National Academics Panel Study (Nacaps, www.nacaps.de) conducted in May 2018. 10,458 persons at 26 German universities where contacted by e-mail yielding a final sample of 1,766 persons who completed the survey. The cluster sample aimed to achieve a maximum of heterogeneity concerning disciplines and institutional settings covering classical full universities with a broad range of fields of studies as well as specialised technical universities. Based on this survey data, I explore different types of the doctorate by analysing motivations and career intentions of doctoral candidates, formation and nature of their topic of thesis, as well as their involvement in co-operational and translational practices during the PhD. Further analysis will involve latent class analysis.

Preliminary results and next steps

First preliminary analysis of position papers and recommendations shows that differentiation is discussed with regard to three main dimensions, which are rather weakly linked: Firstly, debates set out from the ideal-type of a research doctorate which is characterised by a substantive and independent contribution to research, while potential alternative models are to be demarcated by a deviation from this ideal-type. Secondly, an independent discussion thread which deals with the design and delineation of professional doctorates can at most be identified for single subject fields. Hence content analysis will be extended to specific expert associations. Thirdly, recent developments about cooperative doctorates are connected to differentiation in the higher education landscape, while it remains an open question whether their emergence will foster functional differentiation of doctoral degrees. In sum, the preliminary analysis of position papers and recommendations suggests that the doctorate is discussed intensively in science and higher education policy. The analysis also reveals that the discourse hardly refers to empirical data which provides insight to which extent the doctorate in Germany is focussed on research, rather than being an extended period of learning or professional training. Findings of descriptive analysis of survey data can be condensed as following: In accordance to previous studies it shows that the population of doctoral candidates splits up in group of roughly one third of the doctoral candidates who strive for employment in the higher education respectively publically funded research sector and a second large group of more than two fifth seeking employment in the private respectively industrial sector. Moreover, far more than half of the respondents are motivated by the fact, that the doctorate will improve their career chances outside academia. In contrast, there is only a very small fraction of persons who are motivated to do a doctorate in order to reflect about their professional practice. However, for a surprisingly large share of more than half of the doctoral candidates, the doctorate is not only used to acquire research skills and symbolic capital but explicitly serves as a phase of professional orientation. This especially concerns career perspectives outside academia, while commitment to an academic career is taken already in an early career stage. Consequently, I conclude that the multi-functionality of the doctorate and the lack of differentiation in formal degrees makes the German doctorate attractive as an intermediate career stage even if an academic career track has never been intended. The findings suggest that in Germany, the doctorate as an institution is shaped by conflicting demands: while science policy debates revolve around sharpening the still predominant normative model of the research doctorate and specific fields demand for the introduction of professional degrees, its latent function as an orientation phase to navigate in questions of professional development outside academia is hardly recognised.

ABSTRACT. Background and rationale: In spite of being a large portion of the biomedical workforce, postdoctoral researchers are often the forgotten population in academia, with very little data existing on their salary, benefits, and career outcomes. Future of Research works for greater transparency on these issues, holding institutions accountable, and enabling early career researchers to advocate for evidence-based change in academia.

Methods: By contacting HR departments and postdoctoral offices, we monitored the compliance of institutions with a federal labor law mandating postdoctoral pay raises nationally. Subsequently, we submitted FOIA requests to U.S. public institutions to gather information on actual postdoctoral salary amounts and associated titles. We also helped to create an online platform (postdocsalaries.com) for self-reporting of individual postdoctoral salaries and benefits. In addition to these factors, in collaboration with Rescuing Biomedical Research, we examined biological sciences postdoc numbers within U.S. institutions on a yearly basis from 1980 to 2015 to shed light on how postdocs are counted nationally.

Results: Our institutional compliance monitoring affected decisions to raise postdoctoral salaries by comparing institutions side-by-side. The FOIA studies took this work further and showed large national discrepancies in pay levels based on gender and postdoc title, and a complex relationship between salaries and institutional federal funding levels. Finally, examining postdoc numbers showed institutional variability in reporting, resulting in fluctuations that mask larger trends in postdoc employment.

Significance: These studies demonstrate the lack of data availability and transparency around postdoctoral salaries and numbers, acting as a barrier to institutional change. We hope these studies shed light on the experiences of postdocs within institutions, and the need for reforms to improve institutional policy and tracking of postdocs nationwide.

ABSTRACT. Introduction

The UK is often discussed as a case of a country where government has increasingly been pressuring university scientists towards producing research that is not only excellent, but also has significant societal impact. While the UK scientists mostly aim to influence policymaking with their research (Kenny, 2015). it has been repeatedly stressed by the policy users that not all evidence is equally useful, and for policymaking, academic research seems to be one of the less useful types of evidence (Arinder, 2016; Vilkins and Grant, 2017). We investigate why this might be the case by shifting the attention from scientists’ efforts to generate and evidence impact of research to whether, how and under what circumstances scientific research is mobilised by policy organisations. In the case study presented in this paper, we probe into the practices of scientific research use to address pressing transport policy issues in a medium-sized city in the north of England. Our research contributes to further understanding between the institutional conditions in policy organisations and types of impact of scientific research. We also develop a nuanced enquiry of studying impact of research as a process that unfolds over time, and how diverse sources of knowledge from various disciplines contribute to addressing a complex societal problem.

Background

Previous studies that scrutinised the use of scientific research in policymaking have tended to stress the different nature of scientific research from what is required by evidence-handling procedures in policy organisations. For example, Arend (2014) stresses fundamental nature of questions asked by science and the long timelines of scientific process as opposed to changing agendas and applied issues that concern policy organisations. These barriers relate more to the different nature of academic research outputs and the knowledge that could be useful for policymakers (Newman et al 2016). Non-scientific research actors, such as NGOs and think tanks, may better align with policy priorities and may be better positioned to address them than academic scientists. The scholarship on linkages and interactions between science and policy usually recognises that scientists cannot guarantee impact because much of the pathway is beyond their control. From the political science side, scholars who study the use of scientific research in policymaking tend to analyse scientific impact as one of the variables in the political process, the outcome of which is determined by multiple factors ranging from configuration of advocacy coalitions to how the general public perceives the issue (Smith 2013). The analysis of organisational conditions of policymaking organisations is limited to formal and legal channels within which policymakers collect and examine codified scientific evidence (Beland 2005; Palmer 2010; Wu et al. 2017). As a result, it has been suggested that the best way to ensure impact takes place is to create incentives and opportunities for the two communities to interact (Ramos-Vielba et al. 2015). As a way forward, we propose to further conceptualise how institutional conditions in policymaking organisations may influence the uptake of scientific research. The Science – Conditions – Impact (SCI) framework (Edler et al, 2017) zooms into the ‘black box’ of institutional conditions of the user side of scientific knowledge. Our theoretical stance follows a political science reflexive institutionalist approach. We admit that the policy process is interest and power driven, yet policy problems, normative and material interests are being constantly redefined by the actors capable of reflecting on, and changing, their strategies (Hall, 1993, Edler, 2003). As a result, ideas can be the source of institutional change (Schmidt, 2010). Scientific knowledge and ideas embedded in it can therefore play an important role in the policy process. The SCI framework consists of three key components: the scientific knowledge and ideas embedded in it; institutional conditions, subdivided into institutional conditions influencing intra-organisational processes and inter-organisational interactions; and the various forms of impact of science on policy. Institutional conditions affect the impact of scientific knowledge, but the scientific knowledge itself is shaped by the expectation of impact it might have. Following Scott (2014), we distinguish between three types of institutional conditions that help us to understand what shapes the identities, motivations, practices and strategies of “users”: - Cognition: frames of interpretations that give meaning, shape the understanding of cause and effect relationships; - Normative world views that determine what kind of knowledge is asked for. They act as filters for the absorption of scientific knowledge (Rein and Schon, 1993); - Incentive structures and regulative conditions, i.e. the system of formal and informal rules, rewards and recognition within bureaucracies.

The Case Study

The approach towards analysing not only the pathway of concrete instances of the use of scientific research in policymaking, but also of all those instances where the impact took place but cannot be measured, requires longitudinal qualitative methodology. To analyse the conditions that affect the extent to which scientific research affects policymaking, we are developing a longitudinal case study within the process methodology (Gulbrandsen and Woolley, 2018). We forward and backward trace various interactions between policymakers and scientists and employ document analysis, interviews and observation. We study a policy organisation in a medium-sized city in the north of England. The city has grown substantially in the past few years and has enjoyed significant investment from the national government towards industrial development. As a result, the city’s transport network has not been able to cope with its growth, and issues such as air quality and congestion emerged as the key issues hampering further growth. The authorities have proposed multiple measures in the past decade to address the problem of congestion, ranging from a congestion charge to the most recent efforts to promote cycling and walking as alternatives to car use. Our preliminary work (to be fully reported in the presentation) suggests that we will have findings in the three areas of our SCI framework. Empirically, we conducted a discourse analysis of the key transport policy documents and their underlying evidence base reports. The results reveal that only about 20% of citations come from scientific sources. The city authorities rely on commissioned research from local consultancies. Academic citations are unevenly distributed across thematic sections of the documents, which suggests differences in internal research processing routines. The qualitative enquiry reveals an extensive network of informal personal, as well as formalised project and partnership links that the organisation has with multiple British universities. However, the knowledge obtained through interactions in these networks is not readily useful for policy development. Attitudes to science and scientists differ in different organisational units.

ABSTRACT. Individual paper to be presented on the following proposed panel session: Session organizer: Magnus Gulbrandsen Panel session: “The user side of impact: Understanding how research generates societal effects by investigating the process from the perspective of the non-researchers”

Reframing policy issues through research: how obesity became a global warming concern

Current science and innovation policies are largely framed under headings such as grand challenges, societal challenges and missions (Kuhlmann & Rip 2018; Mazzucato 2018; Schot & Steinmueller 2019). Although these terms represent somewhat different perspectives and policy options, they all indicate a desire to tie research and innovation activities closer to the perceived most important current and future problems in society. Through targeted policies and new governance mechanisms, research-based knowledge networks are reconfigured and given certain directions that can increase the chances that even wicked problems (Rittel & Webber 1973) may find some solutions or see some measure of progress.

This paper deals with a case where research is invoked not necessarily to find direct solutions to a challenge, but to reframe it with the possible hope of getting out of a locked-in situation with few policy mechanisms that seem to work. As such, the paper complements other investigations of the relationship between research and grand challenges – we look in particular at how research helps to define the challenge itself and through this change the decision-making space for policies. The paper also contributes to the literature on research impact, by investigating a case of different types of impacts – conceptual, instrumental, political – of research on policymaking (cf. Joly et al. 2015; Donovan 2011).

Our case is obesity, certainly a global and serious challenge, and our starting point was an observation that some actors interested in obesity recently started referring to global warming and the need to handle that particular challenge as well, forging a conceptual link between the two societal problems. In the paper, based on an ongoing case study, we explore this change process and ask what kind of research that played a role, how and why it was used and what it may have led to. Our first impression is that research has been used to reframe obesity from an individual to a societal issue, opening up new opportunities for intervention. The rest of this abstract give an introduction to the case based on a preliminary analysis, to be supplemented by expert interviews and additional document analysis spring/summer 2019.

Obesity is on the rise worldwide. Since 1975, obesity has nearly tripled. In 2016, 39% of adults over 18 years were considered overweight, and 13% were obese, according to the World Health Organization (WHO). In the same period, research on overweight and obesity has bourgeoned, ranging from research on behavioral versus biological risk factors, nutritional strategies and preventive measures and different health consequences, to mention a few. One outcome is the acknowledgement of obesity as a multifactorial disorder, caused by a combination of several genes and behavioral factors. Still, the main formula guiding weight management remains the so-called energy balance hypothesis, stating that energy in should equal to energy out. Strategies to activate and manage weight loss have therefore largely targeted the individuals who carry the extra kilos, and therefore have to alter their eating habits (energy in) and/or physical activity (energy out). Within this paradigm, obesity is largely framed as a health problem that is managed by the individual patient and health professionals.

While much research has been concerned with individually oriented causes of obesity, there has also been a surge in research that focuses on possible external or environmental driving forces, and how society’s development and organization can shed light on the increase of obesity. Such research is often based on population surveys that for instance couple nutrition trends such as the increase in sugar intake to weight gain at the population level. By linking societal trends with the growth of obesity, the assumption that obesity is a social phenomenon and not just due to the characteristics of the individual has received increased attention. As a result, attention is moved to societal characteristics such as food consumption patterns, access to healthy food and the organization of food production, simultaneously turning overweight and obesity into a political matter. However, judging from the rise in obesity, possible actions have failed, and many have attributed the failure to obesity as a ‘wicked problem’. It has no singular cause or solution, and is deeply embedded in dynamics reinforcing the problem. This also means that it cannot be researched through simple ‘trial and error’ strategies suited for ‘tame’ problems (Rittel & Webber 1973).

In Norway, the link between health (including body weight), nutrition and food production/-markets has been a salient political issue for decades which has been attended to by the state Nutrition Council, providing advice on nutrition and health to the public and to the government and health services. For a long time, the council was a success in the sense that it provided the latest research on nutrition to the people through awareness-raising campaigns, but also politically, by providing guidance for state subsidies and taxation of various foods. The politicization of the council also became its curse, and after it came into conflict with key business interests, the council was temporarily closed down. At the same time, food consumption is increasingly decoupled from state regulation as food production and distribution are ever more left to market mechanisms, illustrating the challenges of handling obesity politically.

Obesity is now recognized as a global epidemic. Yet while declaring it an epidemic has increased attention to the issue, actions to tackle the issue remains inefficient. Explanations for this failure have been, in addition to the problem’s ‘wicked’ nature, that obesity is a low-status problem associated with shame and largely individualized as a problem. Can framing the issue differently spur increased political attention and efforts to create effective solutions to the obesity problem?

In January 2019, a Lancet commission established to provide advice on obesity based on the latest research, declared obesity as one of three pandemics – together with undernutrition and climate change – underlying the new global syndemic threatening the planet. The commission adopted a systems perspective on obesity, and identified common systemic drivers and solutions across the three pandemics, claiming they are symptoms of “deeper underlying systemic problem that require systemic actions”. For example, the commission identified common societal feedback loops, including that of transportation and land use systems, such as red meat production being both a driver of global warming and weight gain.

One of the innovations of the Lancet report, is that they couple research on obesity and on global warming, and they calculate the effect of action on one area (obesity) in another area (global warming), thus demonstrating a win-win effect of measures to reduce obesity. As such, they frame new dietary requirements as a solution to a more high-status or mature (political) issue, possibly prompting more attention and urgency to the matter. In the full paper, we will explore in more detail the research behind this change and how it has been used, as well as to explore some national responses to the commission.

ABSTRACT. Paper proposal to the Atlanta 2019 conference - to proposed panel on “the user side of impact”

Background and approach

How do policymakers access, distribute and use research information in the policy process? Which conditions influence the use of research and are these conditions equally important in different policy contexts? To address these questions, we have made a systematic review of the literature on research uptake and utilization in public policy organizations, and have used these insights to develop and test a survey-based method to investigate such questions empirically. In this paper, we will present findings from both studies and use them to discuss future developments and applications of methods to capture perception, application and use of scientific knowledge in policy contexts.

Literature review

The first part of the paper present findings from a systematic literature review on determinants of research utilization in public policy organizations. We performed multiple literature searches in Scopus and included into the corpus only empirical studies that targeted policy makers and their perceived use of research in policy processes. We identified 31 studies that met our strict inclusion criteria. These included empirical evidence from several policy domains as well as continents. Within this field of study, health policy and practice is overrepresented. We analyse the literature with the goal of identifying key factors that influence research utilization at multiple levels; individual, organizational and policy field level determinants.

The studies have different objectives, but together they tell a comprehensive story of interrelated factors that affect use of research among policymakers. For research to be utilized it needs to be timely and credible, it must be accessible and the policymakers need to have sufficient time and absorptive capacity to access, assess and understand the research results. Absorptive capacity is linked to educational level and prior work experience. To facilitate processes of research utilization the organisational culture and practices must value academic research, and a strong relationship and collaboration between policymakers and researchers is beneficial.

In terms of differences between policy domains, within policy domains that directly influence peoples’ lives and welfare, policy makers are more likely to use research. Examples of such issues are climate, food, health, welfare and education. Evidence from the literature review also indicate that in policy fields that are characterized by a high degree of controversy and public attention, policymakers are more likely to make decisions based on research evidence.

The literature has also looked at different uses of research; the classic formulation is the division between instrumental, conceptual and symbolic use of research, which are seen as complimentary. The relationship between preconditions and uses of research across multiple policy domains has however received limited treatment in the extant literature, and our ambition is to develop further knowledge on this issue in particular.

Pilot study of research utilization among Norwegian policymakers

In the second part of the paper, we present some results from a pilot study on the use of research knowledge in public policy organizations. The survey was developed based on the literature review and an additional review of the methodological tools utilized in prior empirical studies. Scoping interviews with ten policymakers were also carried out to understand the specific communication patterns and determinants of research uptake and use in different policy domains.

The survey was structured to shed light on multiple determinants of research utilization, as well as patterns of research access, distribution and different kinds of use. The main dependent variable is research utilization measured both in terms of utilization intensity (frequency of use acts reported) and the different kinds of uses of research. Independent variables include individual characteristics, characteristics of policy fields and organizational level characteristics. In addition, respondents also answered questions about the characteristics of research that is seen as most useable from a policy perspective and also how research knowledge is accessed by policy makers and how such knowledge is diffused within public policy organisations.

In the pilot 277 individuals employed by seven public policy organizations in Norway responded to the survey (a response rate of 44 %), and the organizations that piloted range from having responsibilities for welfare and health policy, environmental and climate policy, energy policy, educational policy and regional policy/internal affairs. The majority of the informants work as policy advisors, and the remaining 1/4 held managerial positions, and almost 100 percent of the sample has education at the postgraduate level.

Preliminary results

The results indicate that policymakers in Norway that often use research, match the characteristics found in earlier investigations: they are senior, they have a high education level and they have work experience from other sectors, particularly research performing sectors (1 in 4).

Most respondents state that they gain access to scientific knowledge in an informal manner, i.e. they search for research information on the internet or they ask colleagues inside or outside their organisation. This informal system of sharing research is also the main mechanism for diffusion of research knowledge inside policy organisations.

Employees with background from natural sciences and technology are more inclined to access scientific resources directly from scientific sources, and rely less on “mediated content” (summaries of research, information from colleagues etc.). We also find that policymakers that draw on external sources of knowledge of one kind (e.g. research journal articles) are more likely to draw on multiple sources of scientific knowledge.

When asked about the use of research in policy, we asked whether the informants had directly used research in their own work, including having cited research and/or having summarized and drawn conclusions from research in policy documents. About half of the sample have cited and drawn conclusions from research publications in policy, but with considerable variety across organisations. Employees that access research the most are also most likely to report use of research in this way. In correspondence with prior studies, we have attempted to discern different kinds of use of research in policy – conceptual, instrumental and symbolic uses of knowledge. The respondents highlighted that conceptual use (i.e. to provide new understanding of a policy issue) or instrumental use (to develop policies or programs) are the most frequent uses of research in policy contexts. Symbolic uses of knowledge (i.e. to justify decisions already made) is the least common use of research according to the informants. There are differences between organisations and policy field, as well as individual level differences that explain different uses of research in policy contexts.

Conclusion

The paper sheds light on the preconditions and uses of research in policy, seen from the perspective of policymakers as potential users of research. We have developed a conceptual framework to investigate research utilization empirically. The main contributions of the paper is to discuss preconditions and uses of research across multiple policy domains through an updated systematic summary of existing research, as well as some empirical findings in a Norwegian policy context. In the full version of the paper, we will perform a full analysis and discuss the results in terms of prior research and outline further research avenues.

ABSTRACT. Presentation for Panel “The user side of impact: Understanding how research generates societal effects by investigating the process from the perspective of the non-researchers” Organizer: M.Gulbrandsen

One decade ago the SIAMPI project was launched. Funded by the European Commission, SIAMPI developed a conceptual framework and tools to assess research impact through the “Productive Interactions” established between researchers and users. Since then the concept of “Productive Interactions” has been used in many research impact assessments, has been extended and adapted, and also subjected to some criticism. The objective of this paper is to assess how and to which extent “Productive Interactions” can be used to analyse research impact when the assessment is carried out from a user perspective and “tracing backwards” from a problem area. The paper will offer a definition of “productive interactions”, discuss the assumptions underlying an impact assessment approach relying on the identification of “productive interactions”, and the implications of these assumptions when attempting to analyse impact from a user perspective.

“Productive Interactions” can be defined as “exchanges between researchers and stakeholders in which knowledge is produced and valued.” (Spaapen&van Drooge 2011). When this contact leads to an effort by the stakeholder to engage with the research produced we refer to it as a ‘productive interaction’; when ‘productive interactions’ result in stakeholders doing new things or doing things differently we say that the research has had an impact.” (Molas-Gallart&Tang 2011).

Therefore, “SIAMPI assumes that for social impact to take place, a contact between researchers and non-academic stakeholders must have taken place”. Note, however, that for the interaction to be considered “productive”, it does not need to lead to a change in practices, or to new products or services. It suffices for the stakeholders to “make an effort to engage with the research”. Such engagement may extend from reading and considering a report, to an intense collaboration with a research team. Therefore, we assume that research can be of practical value, even if it does not lead to changes in the practices that would have taken place had the research not been carried out. The justification for this approach lies in the existence of forms of use of research results that do not lead to action: research results can be used to confirm that a policy intervention was adequate, may be used to recommend against a change in policies, or may not find application because such application would run against the interests of current incumbent groups, even if it would benefit the majority. Research results can also be helpful when they can help explain tragic occurrences after the fact, even when they have not been used to provide solutions to the repetition of such occurrences.

There is an important consequence of this set of assumptions: not only does SIAMPI focus on processes rather than the impacts, but assigns value to the processes themselves. This focus on processes has relevant methodological implications.

“Productive Interactions” were initially applied to situations in which the point of entry into the study was a specific research project or research investment, “tracing forward” from the specific investment through the interactions the investment generated between researchers and research users. The approach was therefore well suited to the evaluation of specific research investments, and could draw on a well-bounded set of informants. SIAMPI developed a set of interview protocols that were first applied to researchers working on the projects and programmes under study. Stakeholder informants were nominated by successive rounds of interviewees (“snowballing”). This approach focuses on those interactions based on direct contacts and where there is scope for mutual influence (i.e.: stakeholders can also influence the research agendas and instances of collaboration and co-creation can be identified). Yet, when the exchange between research and stakeholder is limited to a traditional one-directional flow from researcher to user or beneficiary, as when the “impact pathway” relies basically on stakeholders accessing written materials produced by the researchers, such snowballing techniques are of limited use. Researchers will often be unaware of whether and how their publications may have reached distant user communities, and how they have used them.

Taking as a starting point a problem area and a set of user communities allows, in principle, for such distant use of research results to be identified. Yet, it makes it difficult to provide the detailed studies of specific research-user “productive interactions” that SIAMPI projects have typically delivered (Spaapen, et al. 2011). Tracing forward from the research activity allows for a well-bounded subject; although the impact pathway will become complex overtime and involve an increasing set of actors, the community under investigation remains well bounded. In comparison starting with a problem area and a set of research users to “trace backwards” to influential research is likely to generate a much broader set of potential lines of inquiry: users are likely to have engaged with a broad variety of research results and researchers throughout their professional life. Research-users may also seek to engage with a researcher or a research group not in relation to, or through, a specific research activity, finding or outcome, but due to recognition of their general competence and authority in relation to knowledge that is (potentially) relevant to a problem area. This further complicates attempts to trace backwards from a problem area, as the relationship between productive interactions and specific knowledge objects becomes less precise.

From the perspective of research-users, such interactions will also be influenced by the socio-technical imaginary (Jasanoff&Kim 2015) shaping the societal benefits desired and the populations that are expected to be the ultimate beneficiaries of research. Relevant interaction networks are thus likely to appear more diverse and distributed when starting from the ‘problem area’ perspective, and the distinction between researcher-users and beneficiaries may become blurred in certain problem area contexts. It will often be the case that a broad set of “productive interactions” can be identified but a detailed analysis of the processes by which they have emerged may be made impossible by the sheer variety of engagements with research that the users have entered into. When this happens the analysis of impact processes from a user perspective will tend to focus on how the user deals with researchers and research information, rather than a detailed “pathway” analysis of how an interaction has developed over time. But it will also potentially open up a perspective on productive interactions as inclusive of beneficiaries; an impact process that may also transform the identity of beneficiaries into research users or hybrid user-beneficiaries. We will, therefore, conclude that a “productive interaction” approach from a user perspective will necessarily be very different from the interaction process studies that have characterised much of the SIAMPI work.

References Jasanoff, S., and Kim, S.-H. (2015). Dreamscapes of modernity: Sociotechnical imaginaries and the fabrication of power. Chicago: University of Chicago Press. Molas-Gallart, J. and P. Tang (2011). "Tracing "Productive Interactions" to identify social impacts: an example for the Social Sciences." Research Evaluation 20(3): 219-226. Spaapen, J. and L. van Drooge (2011). "Productive interactions in the assessment of social impact of research." Research Evaluation 20(3): 211-218. Spaapen, J., et al. (2011). SIAMPI Final Report. Amsterdam.

ABSTRACT. The National Institutes of Health (NIH) plays a critical role in funding scientific endeavors in biomedicine that would be difficult to finance via private sources. One important mandate of the NIH is to fund innovative science that tries out new ideas, but many have questioned the NIH’s ability to fulfill this aim. We examine whether the NIH succeeds in funding work that tries out novel ideas. We find that novel science is more often NIH funded than is less innovative science but this positive result comes with several caveats. First, despite the implementation of initiatives to support edge science, the preference for funding novel science is mostly limited to work that builds on novel basic science ideas; projects that build on novel clinical ideas are not favored by the NIH over projects that build on well-established clinical knowledge. Second, NIH’s general preference for funding work that builds on basic science ideas, regardless of its novelty or application area, is a large contributor to the overall positive link between novelty and NIH funding. If funding rates for work that builds on basic science ideas and work that builds on clinical ideas had been equal, NIH’s funding rates for novel and traditional science would have been the same. Third, NIH’s propensity to fund projects that build on the most recent advances has declined over the last several decades. Thus, in this regard NIH funding has become more conservative despite initiatives to increase funding for innovative projects.

ABSTRACT. Introduction

Scientific production in Brazil is intrinsically connected to graduate programs as most of the country’s research is carried out in universities. At present, the Coordination for the Improvement of Higher Education Personnel (CAPES), a federal institution responsible for evaluating those programs, is reviewing their evaluation system. At present, one of its core components is a journal ranking mechanism named QUALIS. As CAPES' evaluation system acts as a policy tool to allocate public funding, this process needs to be carefully planned to safeguard the quality of research in the country.

Created in 1998, QUALIS is composed of different journal ranking lists, one for each knowledge field, in which faculty and students publish their work. The journals are classified into strata indicative of quality: A1, A2, B1, B2, B3, B4, B5, and C, in descending order of importance according to their relevance to each field. The system is based on specific criteria that combine bibliometric and non-bibliometric indicators. Nowadays, 49 different journals lists are published annually by CAPES, one per knowledge field, which generates misunderstandings among the academic community as the same journal can have a different evaluation in different knowledge fields. Besides, the majority of the fields use the Impact Factor from the Web of Science as the main indicator to determine the stratum of a particular journal. This practice goes against international trends - the San Francisco DORA [Declaration on Research Assessment], the Leiden Manifesto for Research Metrics and the Metric Tide - to stop using the Impact Factor as the main gauge to evaluate the quality of academic research (DORA, 2012; Hicks et al., 2015; Wilsdon, J., et al., 2015).

As an alternative solution to the problem, CAPES is building a single ranking list of journals, replacing the current practice to keep one list for each knowledge field. As Scopus is one of the most extensive and widespread databases on different knowledge fields (Guz & Rushchitsky, 2009), the new system intends to use metrics based on that database. Relating to this proposal, the Social Science and Humanities fields (SSH) are reluctant to adopt it. Their discourse is aligned with the widely known concern related to differences in publication practices according to knowledge fields. A strong argument is that research in the Social Science and the Humanities often concentrates in national or regional problems that do not interest the international community, and because of that a large part of the publications in the area is made locally, in Portuguese, in journals not indexed by Scopus or other major international databases.

To circumvent these limitations, CAPES has proposed to use Google Scholar h-index to compute the relevance of publications in Social Sciences and the Humanities for journals not covered by Scopus. The use Google Scholar (GS) metrics is more attractive because it is a broader database of citations that includes citations in books and conference proceedings, and it also covers a wider set of journals than the Web of Science. In this direction, this work aims to investigate whether Google Scholar h-index would be an alternative indicator for the relevance of publications in the Social Sciences and Humanities, considering the former Brazilian journal ranking. To achieve that, we examined whether there is a correspondence between SJR indicator from Scopus and GS-based h-index. Then we verified if there is any agreement between h-index and the categories of the former QUALIS raking in the Social Sciences and the Humanities.

Methodology

For this study, data from two periodic evaluations carried out in Brazil (from 2010 to 2012 and 2013 to 2016) in the field of Education, Business and Management, and Literature and Linguistic were analyzed. On average eighty journals were selected from the QUALIS list by each periodic evaluation and field. In the selection, a random process was used, but only those journals indexed by Scopus were considered. For this selection, the indexed journals were ordered by the SCImago Journal Rank indicator (SJR). This indicator is based on the transfer of prestige from a journal to another; citations are weighted to reflect whether they come from a journal with a high or low SJR. The h-index for each selected journal was retrieved from Google Scholar using the software Publish or Perish. H-index is defined as the greatest number of publications h for which the count of lifetime citations is greater than or equal to h. The period used in the search was from 2010 to 2012 and 2013 to 2015, a three-year citation window. Correlation analyses were made between SJR and Google Scholar h-index as well as between the former indicator and the QUALIS categories. Spearman Rank Order Correlation Coefficients were calculated using STATA version 15.0.

Results and Discussion

The h-index and SJR were significantly correlated in all fields in both periodic evaluations (p-value<0.001). The coefficients ranged between the periodic evaluations, from 0.548 to 0.838 for Education, 0.558 to 0.750 for Business and Management, and 0.672 to 0.688 for Literature and Linguistics. Our results showed that the degree of relationship between Google Scholar and SJR rankings become stronger over time for Education, and Business and Management. The agreement between these two impact metrics suggests that, for the studied fields, Google Scholar h-index could provide an alternative for the journals not covered by Scopus in the Brazilian journal ranking system. On the other hand, a non-significant Spearman correlation result was found between h-index and the QUALIS categories for Education, and Literature and Linguistics in both periods. Moreover, a significant negative correlation was found for Business and Management in both periods, which ranged from -0.67 to -0.44. These results indicated a weak agreement when comparing Google Scholar and the former QUALIS rankings. Regarding the Brazilian research evaluation, it relies mainly on peer review and indicators are used as an auxiliary tool to analyze merit. These are among the fundamental principles established in the aforementioned international manifestos. However, there are some fields such as Education, Literature and Linguistics that don’t use any bibliometric indicator in their analysis. It is possible that the qualitative criteria used in these fields are not being efficient to detect the ‘impact’ or ‘visibility’ of those journals, even considering data from a larger database such as Google Scholar. In this sense, a question arises: how can bibliometric data be used in scientific production assessment without overpowering other information that may exist and may contribute to the evaluation process? This seems to be one of the main challenges in the area at present.

Bibliography

Dora. San Francisco Declaration on Research Assessment. Dora. Available in: . Access in: May, 07. 2018.

Hicks, D. et al. (2015). Bibliometrics: The Leiden Manifesto for research metrics. Nature, 520, 429–431.

Wilsdon, J. et al. (2015). The metric tide: Report of the independent review of the role of metrics in research assessment and management. Sage. DOI: 10.13140/RG.2.1.4929.1363

Guz, A. N., & Rushchitsky, J. J. (2009). Scopus: A system for the evaluation of scientific journals. International Applied Mechanics, 45(4), 351.

ABSTRACT. Background: To date, the most comprehensive study of research proposals is by Ginther, who analyzed 83,000 R01 proposals submitted to the NIH between 2000-2006. That study focused on a very sensitive topic—the possibility that the peer review process at NIH has gender/racial biases. Smaller scale studies in Europe have also shown possible gender bias in proposal evaluation. Overall, the literature on proposal success/failure focuses, almost exclusively, on the issue of bias.

Proposal analytics has significant potential beyond the analysis of bias. The purpose of this presentation is to illustrate the types of contributions that proposal analytics can make to our theoretical understanding of the science of science and to practical issues dealing with research evaluation and planning. We present preliminary results from our analysis of 3459 research proposals submitted by the University of Michigan Medical School (UMMS) to NIH between 2010 and 2016, and address both theoretical and practical issues.

Theory: One of the central questions in the science of science is topic choice—which topics do researchers investigate over time and why? This is sometimes described as the hedgehog/fox phenomenon: some researchers seem to dig into a single topic while others seem to bound from topic to topic over their careers.

The existing literature on topic choice focuses on ex post analysis using papers and citations. We do not currently look at the documents that could provide us with ex ante predictions of what researchers are planning. Research proposals represent the intentions of researchers and are an ideal data source to understand topic choice.

For example, we have identified 71 researchers at UMMS that submitted at least 10 research proposals to NIH over the seven-year period. Roughly 84% of these research proposals were not funded. Do the successful 16% simply represent “better versions” of the remainder or is the truth more nuanced? Is it possible that many research proposals were rejected because the proposal didn’t conform to the norms of the reviewers? To what extent does the researcher have a vision of what they want to do (and persist in achieving this vision)? Do researchers compromise because of funding biases? Do researchers primarily chase funding or are they guided purely by curiosity for exploration of the unknown?

We don’t have an answer to these questions … yet. However, we believe that these are the right questions to ask and we are finding that the information embedded in the research proposals submitted by these people may provide answers. Almost every research proposal has an extensive bibliography—the number of references in a research proposal to NIH is more than the number of references in a typical medical article. The description of the research strategy is usually limited to 10 pages—roughly the size of a research article. All the tools that have been developed to analyze publication patterns (i.e. citation and text analysis) can be applied to this robust and as-yet largely unexplored corpus. One can tell which proposals build upon each other. Proposals can be accurately assigned to topics. There is a huge opportunity to gain insights into topic choice by looking at research proposals.

Practice: From a practitioner’s perspective proposal analytics is all about helping people. There are two groups of people in a research university that can significantly benefit from proposal analytics: research administrative leadership (such as Provosts, Deans, and Department Chairs) and faculty support staff.

Research proposals are unique in that they can help administrative leadership identify rising stars among young faculty. At present, junior faculty are evaluated based on ex post events and artifacts (as mentioned above). While this may be appropriate for evaluating associate and full professors—those who have had time to build up academic wealth in the form of these outputs—we have found that such reliance on ex post events tends to reinforce racial/gender biases for early career researchers to the detriment of their careers and the diversity of scientific thought. For a variety of reasons, some demographic groups are unduly privileged when evaluation prioritizes ex post productivity, such as through papers (our results mirror those of Ginther on racial/gender bias). Further, an under-emphasis on proposal analytics ensures that opportunities for effective institutional intervention remain unexamined. For instance, if a number of researchers repeatedly try and fail to secure grants within a certain scientific topic it could be a signal for capital investment in that area. Perhaps they simply lack the necessary equipment or core services to be competitive.

UMMS has an entire unit that is dedicated to helping researchers write better proposals. The people in the department have years of experience in proposal writing. But would you be surprised to learn that the unit does not track intervention effects? There is no feedback on whether their advice helps. Or stated in the language of biomedicine: treatments are prescribed without evidence of efficacy. For example, one of the most oft-repeated pieces of advice offered to proposal writers is to focus on clarity. Consequently, one of the first studies we did with proposals was to test the assumption that clarity resulted in higher proposal success rates. We reported those results at this conference two years ago. It was a very small study—judgments about clarity were made by Professor John Swales about 20 matched pairs of proposals. He judged that 12 of the pairs were of equal quality. He correctly identified which proposal was accepted in 7 of the remaining 8 pairs.

We’ve now expanded this study (n=3459) using a standard indicator of writing clarity (the Fog Index). Would you be surprised to find out that writing clarity had no impact on proposal success? Yes, we could dismiss this result as a matter of poor measurement technique. But if it’s true that clarity doesn’t improve proposal success, we may be doing more harm than good by emphasizing the need to rewrite and rewrite with this as a primary aim. Instead of spending extra time on refining a single proposal for clarity, it may be more effective to focus on refining the overarching narrative structure of the proposal—the story arc. Or it may be better to simply write another proposal. These are common recommendations for fiction writers. A successful writer focuses on the story, not the individual lines, and builds up an inventory of works in progress. Analogously, the sign of a good researcher (as mentioned previously) may be someone who writes compelling stories and builds up an inventory of (failed) proposals that explore different topics of interest to the researcher.

Implications: Our purpose at this conference is to encourage others to analyze research proposals. Convince the Deans of your school that the codification of research proposals can be of significant benefit to them and can be accomplished at only marginal cost. Please talk to us about how to overcome the barriers that will inevitably be placed in your path. Over the past three years we have encountered many and can offer practical advice for overcoming them. Proposal analytics can make a significant contribution to both our theoretical understanding of how research evolves while also helping us to make evidence-based decisions about research administration going forward.

ABSTRACT. China’s participation in global science has risen at a spectacular rate over the past 30 years (Zhou & Leydesdorff, 2007). China’s scientific output has risen in most fields. Citation strength of Chinese output has not risen at the same rate as publications, but it is also rising significantly (Xie et al., 2014). China has usurped the United Kingdom to become the number one collaborator for US researchers (Wagner et al., 2015). We know it is growing in size but is Chinese research also growing in novelty? China’s rise in global science raises the question of whether the work is also increasing in novelty, creativity, and innovation. Innovation requires expertise; clearly China is developing the expertise needed to transition to the process of innovating. Fields where Chinese institutions are investing heavily in R&D capacity would be ones to look at for quality and innovation. Secondly, a question is whether Chinese capacity has reached a level to create quality research without international support or collaboration. We hypothesize that international collaborations are more likely to be working on frontier research. In addition, if this is the case, a question arises as to whether Chinese researchers are producing quality work on national teams, on an individual basis, or only as a result of international collaborations. A third testable question is whether openness to exchange is critical to innovation. China has become increasingly open to international exchange, and as it has done so, scientific production has grown. It is not a “highly open” scientific nation, however (Wagner & Jonkers, 2017). This raises the question as to whether electronic censorship may reduce the ability of Chinese scholars to innovate at the frontier. This presentation will discuss work in progress testing whether Chinese researchers are producing more novel work over time, and if so, whether it is being done primarily with international partners. Are Chinese scholars producing increasingly innovative or novel research over time? We expect to find that more novel research is emerging in the internationally collaborative projects, but at a decreasing rate as China develops national capabilities. We are testing this hypothesis now and will have results by the time of the conference. Further, we expect to find that Chinese-only authors are growing more novel over time. This will be tested by isolating the China-only work, and even within that, to see if researchers were trained in China or trained overseas. We expect to find that China is becoming more novel on its own, but at a very slow rate. Further, we expect that more conventional work is Chinese-only, based upon the hypothesis that many parts of the Chinese system are still learning the basics of ‘normal’ science, and that these conventional articles are less likely to be internationally coauthored. This will be tested as well. Creativity and novelty are universally valued in scholarship, but this value is difficult to measure. The team of Brian Uzzi, Satyam Mukherjee, Mark Stringer, and Ben Jones, hypothesized that truly novel advances are accompanied by strength in conventional know-how, joined to a novel idea, often resulting from unique or atypical combinations of prior knowledge (Uzzi et al., 2013). Drawing upon the tradition where the reference to preceding work serves as an elementary building block of the scholarly attribution and reward system, they used this artifact as a proxy for novelty. Rheingold (1985) suggested that more familiarity produces more novelty because it provides a depth of expectation against which to detect departures from the familiar. By acquiring expertise in a given domain, one becomes deeply familiar with its intricacies, which in turn creates a cognitive scaffolding from which to notice minor novelties, or abnormalities that escape the untrained eye. We will follow the hypothesis and measurement tools developed by this team to examine the combination of novelty and conventionality in Chinese science and technology publications over 10 years with data from the Web of Science. We study the relationship between atypical combinations of conventionality and novelty and internationally coauthored work, with the expectation that international collaboration would also be more likely to produce atypical combinations. We will analyze probability distribution of 10th percentile z-scores of all articles. The referenced journal pairs with 10th percentile z-score less than 0 will be collected into a set signifying “novel combinations”. We expect that less than 10 percent of records will fall into this category. For z-score values more than 0, these records will be classified as “conventional combinations” (Mukherjee et al., 2016). Any continuous variables will be binarized and then combined into a nominal variable. With the HN/HC category identified, we will conduct additional analysis. First, we will create sets of articles that are (1) internationally collaborative; (2) nationally collaborative; (3) sole-authored (Chinese); and for LN/HC, a set of (4) nationally collaborative or sole-authored Chinese only.

References Mukherjee, S. Uzzi., B., Jones, B., Stringer, M. (2016). A new method for identifying recombinations of existing knowledge associated with high-impact innovation. Journal of Product Innovation Management, 33(2) 224-236. Rheingold, H. (1985). Development as the acquisition of familiarity. Annual Review of Psychology, 36(1), 1-18. Uzzi, B., Mukherjee, S., Stringer, M., & Jones, B. (2013). Atypical combinations and scientific impact. Science, 342(6157), 468-472. Wagner, C. S., Bornmann, L., & Leydesdorff, L. (2015). Recent developments in China–US cooperation in science. Minerva, 53(3), 199-214. Wagner, C. S., & Jonkers, K. (2017). Open countries have strong science. Nature News, 550(7674), 32. Xie, Y., Zhang, C., & Lai, Q. (2014). China’s rise as a major contributor to science and technology. Proceedings of the National Academy of Sciences, 111(26), 9437-9442. Zhou, P., & Leydesdorff, L. (2006). The emergence of China as a leading nation in science. Research policy, 35(1), 83-104.

ABSTRACT. ** This paper is part of the panel session title: "From Invention to innovation – tech transfer offices, patents, and technology adoption"

Background and context Technology adoption has multiple benefits: productivity increase, higher quality of products, improving flexibility and reducing production costs (Beaumont and Schroder, 1997; Rischel and Burns, 1997; Small, 1998). For instance, to increase productivity, a firm must produce the same outputs with fewer inputs which will lead to a reduction in production costs. The adoption of advanced technologies can also improve product flexibility (Young et al., 1993; Spina et al., 1996). In a study conducted by Baldwin and Lin (2002), Canadian manufacturing firms reported that the most two important benefits of the adoption of advanced technologies are an improvement in productivity and an increase in product quality. These benefits lead to increase innovation and economic performance. Moreover, the adoption of advanced and ICT technologies can have an impact on collaboration and open innovation practices. Particularly, digital technology can play an important role in facilitating collaboration between different organisations internally and externally. In manufacturing, technology is defined as the set of tools (automation + integration) used in the different stages of design, manufacturing, planning and control of the product (Ettlie and Reifeis, 1987). These technologies are complemented with business intelligence technologies that are often ICT driven. As described by Bantau and Rayburn (2016), the first and second wave of ICT were used to automate individual activities and increase connectivity respectively. Today, advanced technologies that are often ICT-driven have become part of complex products and play a key role in productivity improvement (Breur, 2015, Porter and Heppelmann, 2014). Taking all that into consideration,

Methodology - Econometric Models Using the Survey of Advanced Technologies (SAT) 2014 conducted by Statistics Canada, we estimate the factors that influence the propensity to innovate (Innov), in particular the impact of advanced technology adoption. For this purpose, a simple logit or probit will first be used. Because we are interested in the obstacles that may slow down or prevent the adoption of advanced and ICT technologies, we have to turn to an instrumental variable model where the adoption of these technologies will be endogenous. Our general model will be as follows: Innov_1i= Y_1i^*=X_1i β_1+Y_2i β_2+X_3i β_3+ϵ_1 (3)

Adopt_2i=Y_2i=X_2i γ_2+X_3i γ_3+μ_1 (4)

where we observe Y_1i={█(0 if Y_1i^*<0@1 if Y_1i^*≥0 )┤ (5)

and where Y1 measures whether a firm innovates or not, while Y2 represents the adoption of advanced and ICT technologies, measured by a continuous variable (the natural logarithm of the number of advanced technologies adopted) ; and, where X1 is a vector of exogenous variables such as those related to open innovation practices and collaboration, X3 is a vector of control variables such as the size or the sector of the firm, and X2 is a vector of instrumental variables that has an effect on the adoption of advanced and ICT technologies (Y2 being endogenous), such as the recruitment of new employees pertaining to the adoption, capital expenditures and the measures to mitigate the obstacles encountered when adopting.

Results and contribution Findings suggest that the more technologies a firm adopts, the higher its propensity to innovate. This is true across the main four families of technologies we looked at. In addition to that, we find that open innovation practices and quality management strategies increase the propensity to innovate. In fact, findings are in line with literature and suggest that collaboration and strategic alliances with other firms, universities and government research organizations all have a positive impact on innovation. We also found that the expected result that innovation type is impacted by industry. For example, high-tech firms are more prone to product innovations while consulting firms may be better equipped to do organizational innovations. Furthermore, we looked at what impacted the number of adopted technologies. The first instrument we looked at is the number of measures taken to mitigate adoption obstacles and we found a positive correlation between the number of measures and the number of adopted technologies. This is independent of the family of technology adopted and the type of innovation. Other instruments such as the recruitment of employees (dummy variable) and capital expenditures pertaining to the adoption of new technologies also increase the adoption. Our results contribute to the technology adoption literature where we find an endogenous effect between technology adoption and the propensity to innovate. In fact, firms need new technology to innovate but they also must mitigate obstacles and find the right resources to implement these new technologies. This has a lot of implications in practical state because C-suite executives will need plan a clear strategy pertaining to recruitment and capital expenditures. In era of ERPs, the outcome of implemented a new software was a simple pass or fail and it was known very quickly. In today’s world however, companies are adopting more than technology at the same time and it can take year to realize the positive or negative consequences of it. As a future research, SAT 2014 gives us the ability to look specifically at which bundle of technologies companies are adopting which can allow to provide better policy implications. Finally, although this paper provides insight into the early days of the industry 4.0 revolution, it is imperative to repeat this study with a more recent data set to understand how technology adoption has evolved.

ABSTRACT. A central theme of science and innovation policy research is how the risky and difficult path from invention to innovation can be de-risked and facilitated. This session explores three common sources of innovation in companies – university collaborations, patents owned by companies themselves, and advanced technology adoption. The first paper highlights the different roles played by different sizes of firms in terms of R&D contracting: collaborating with startups is more likely to generate income than with large firms. The role of the university within the innovation ecosystem, needs to be clearly understood as pressures to generate income from discovery increases and has the potential to change our universities into a more “utilitarian role”. The second paper focuses on a largely science-based sector: life sciences. The research shows that signalling patent ownership on a company website is associated with substantially greater amounts of external funding, hence suggesting the importance of the visibility of invention and innovation potential. It may also be the opposite as VC require increased positive buzz around the technologies in which they have invested to favour better exit strategies… The third paper examines the impact of technology adoption on various types of innovation, from the classic product, process, marketing and commercialisation innovation, to combination variations (tech vs non tech, product vs the other three as in the new Oslo manual). The paper provides insights of technology adoption at the very start of the Industry 4.0, or the so-called “4th industrial revolution”.

The insights from these studies have implications for policy and program development. For example, technology transfer program design can directly benefit from the insights on which companies are most lucrative to work with discussed in the first paper, and a better understanding of how companies publicize their technologies (paper 2) and the role of the technologies provided in the innovation process (paper 3) might provide opportunities to negotiate better agreements with companies. In addition, improved understanding of the role of patents and advanced technologies in the innovation process can inform science policy, including with regards to the role of support for research and technology development. These conclusions highlight the importance of systemic approaches to study innovation, encompassing several points of view and research angles, to try to make sense of it all. This group of papers also stresses the importance of multiple methodological approaches to perform such systemic analyses of the innovation phenomenon that is put forward as necessary to economic development and wealth creation.

Paper list:

How does exclusivity affect company behaviour toward university research commercialisation? Aksoy, A., & Beaudry, C. (Polytechnique Montreal, Canada).

Revealing Patent Ownership: A Study of Life Sciences Companies. Schillo, R.S., & Sanni, A.M. (Telfer School of Management, University of Ottawa, Canada).

Advanced technology adoption and its impact on innovation performance. Hage, G., Beaudry, C. (Polytechnique Montreal, Canada), & Therrien, P. (Innovation, Science and Economic Development Canada)

ABSTRACT. Session Title: From Invention to innovation – technology transfer offices, patents, and technology adoption

Paper Title: Revealing Patent Ownership: A Study of Life Sciences Companies

Please note: We are following the conference’s submission guidelines in not including references. We have completed a substantial literature review in planning and conducting this work. Background and Rationale Especially in the Life Sciences, patents are often useful tools to protect the intellectual property rights of companies during the generally lengthy periods required to bring inventions to market. There is an extensive body of literature on the role and value of patents in company’s strategies. For example, patents have been shown to enhance companies’ ability to attract Venture Capital (VC) and they are often used to enhance a company’s technological positioning, e.g. by protecting against imitation of competitors, controlling or preventing market entry of rivals, or blocking competitor R&D efforts. In many cases, such benefits require that external parties, such as VC investors or potential rivals become aware of the existence of patents, to the extent that this signalling effect has been the subject of much academic literature, and various contingencies have been documented. However, we have not found references to scenarios in which companies obtain patents, but then do not publicize their ownership of the patents. Theoretically, such a scenario should be relatively rare. While companies may have important motivations to not disclose their technologies and thus not to patent, once a patent has been obtained, the technological disclosure is made, and the information becomes accessible to any serious competitor following patent data. Thus, it seems not disclosing patent information on a web site would not offer much protection, ostensibly only from those competitors who do not follow patent data regularly. In this study we show that, in practice, the phenomenon of companies obtaining patents, but not revealing them on their web sites, is quite common in the life sciences. We present analyses to better understand company strategies in revealing patent ownership.

Data and Methods This study is based on a combined data set from three sources: CapitalIQ, a commercial financial database, the US Patent Office (USPTO), and the web sites of the focal companies. The sample consists of companies in the pharmaceutical, biotechnological and life science industry, derived from the financial database Capital IQ. To ensure relative homogeneity in operating environments, we focus on companies currently operating in Canada or the United States. We focus on small and medium-sized companies with 5-500 employees, and in particular those that were able to raise external funds. Taking into consideration that it takes approximately 18 months for a patent to be approved, we consider data up to 2017. Our data set is comprised of all companies in the pharmaceutical, biotechnological and life science industry in Capital IQ that have data on the amounts raised and web site information. It consists of 283 companies. For each company, we checked the USPTO database to determine if it had patents, and if so, how many. In addition, we checked each web site for the occurrence of the term ‘patent’, and more specifically if the company indicated that it owns patents. The first phase of analyses consists of deriving descriptive statistics to determine to what extent companies reveal patent ownership on their web sites. In addition, it consists of multinomial regression analyses to determine whether our existing variables can predict the combination of companies owning patents and revealing them on their web site, or not. This set of analyses has been completed and we outline the results below. In a second phase, we are expanding the data set to identify more detailed information about the companies with the goal to better understand the revealing strategies. We anticipate a substantial part of this work to be complete by the conference date and look forward to presenting the results.

Results and anticipated results Firstly, our descriptive analyses confirm that companies do not necessarily reveal patent ownership on their web sites. In fact, in our sample, 65% of the companies have patents according to USPTO records, and over 40% of these companies with patents do not reveal that they own patents on their web sites. This suggests that the phenomenon we investigate exists and occurs frequently. Secondly, our logit regressions show that companies that both have patents and reveal them on their web sites have significantly higher amounts of external funding than those that do not have patents or only mention patents on their web site. The only significant difference between companies that have patents and reveal them on their web sites, and those that do not reveal them on their web site is that companies that are subsidiaries are less likely to reveal patents publicly. There also is a significant difference between the two countries included in this study (United States and Canada), in that companies that have no patents in either the USPTO or their website are less likely to be located in the United States than those having patents and revealing them. Thus, from our first phase of the study, we show that a) there are different patterns of revealing of patent ownership, and that b) they are associated with the amount of VC funding, company status, and location in terms of country. These insights are important in and of themselves, as outlined in the last section of this abstract. They also suggest that the analyses planned for the second phase of this study are likely to deliver additional useful insights on company’s strategies with regards to revealing patents.

Significance To the best of our knowledge, the patent revealing behaviour of companies has not been studied yet, even though the potential signalling effect of patents is widely acknowledged in the literature and in entrepreneurial practice. Our study shows that it cannot be assumed that companies will necessarily publish the fact that they own patents, in fact our preliminary analyses suggest that over 40% do not reveal their patents on their web sites. Although the academic literature has not discussed this topic, informal discussions with experts (not reported in detail) suggest that entrepreneurs are not necessarily surprised by this finding, as Venture Capital funders may prefer to maintain confidentiality with regards to the prospects of their portfolio companies, especially those in the early stages. However, we also find evidence that both ownership of and revealing of patents are associated with higher amounts of funding obtained. Based on our preliminary analyses we thus expect that the signalling effect outweighs the concern over confidentiality. These early findings are substantial with regards to their implications for both academia and practice. We hope that additional analyses using a larger data set will allow us to describe patent revealing strategies in more detail.

ABSTRACT. Catherine Beaudry: From Invention to innovation – tech transfer offices, patents, and technology adoption.

Background and rationale

Universities and governments around the world invest a lot of money into research and development. The mechanisms through which these investments generate return on investments are still a source of debate and so are the indicators to measure them. In this paper we concentrate on returns through direct payments to technology transfer offices and industrial research funds. Our aim is to study the effect of exclusivity and the size of the licensee company on strategies, investments, and incomes. Previous research shows that universities use three main strategies to generate income through their services and products: R\&D contracting, licensing or startup creation.

The amount of R&D funding, licensing income and startups launched has been shown to be dependent of the university size and willingness to invest into commercialisation. Others further demonstrated the importance of the field and the local ecosystem. Unsurprisingly, large technical universities geared toward commercialisation in wealthy industrialised regions are more likely to generate funding and income.

The literature indicates that the probability for companies to enter into R&D partnership with a university is determined by its size and its R&D propensity. Yet it’s unclear how the size of the company affects the outcome of licensing negotiations. One would think that since larger companies have more resources to spare they would be better providers for R&D projects and would generate more licensing income. However, logic dictates that large companies didn’t become large by handing out money and we can expect them to be more experienced and well funded in negotiations. Small companies and startups might not have such luxury. Not only are their options of R&D partners more limited, but they also lack the resources and experience to navigate through these types of negotiations. Our aim is to better understand which type of company is more lucrative to work with for universities.

Methods

We use the association of university technology manager’s (AUTM) statistics access for technology transfer database (STATT) to collect information about north American universities. We combine this data with regional GDP values retrieved from Statistics Canada and the Bureau of Economic Analysis of the U.S Department of Commerce (BEA) which we converted into Canadian dollars using a purchasing power parity index obtained from the International Monetary Fund (IMF)

We fit a linear model with panel regressions using maximum likelihood estimations. We further verify results through clustered ordinary least square regressions (OLS) with clustered sandwich variance estimation.

Our dependent variable set is composed of the amount of research funds raised by the university from industry, the amount of research funds tied to licensing deals, as well as the number of licenses generating income and the amount of gross licensing income and their sub components : the number of licenses generating royalty payments, the amount of royalties, the amount of other licensing related income excluding equity sales, and the number of licenses generating more than one million dollars.

Our independent variables include the proportion of: exclusive licenses granted to large companies, non-exclusive licenses granted to large companies, exclusive licenses granted to small companies, non-exclusive licenses granted to small companies, exclusive licenses granted to startups, and non-exclusive licenses granted to startups.

Our control variables involve the size of the university measured in federal research expenditure, the presence of a medical school, the number of disclosures recorded, the number of licenses granted, the number of technology transfer staff, the amount of legal fee expenditure, the local GDP and year dummies.

Results

Our results indicate that exclusivity is used as a bargaining chip by universities and firms alike. The most lucrative way to generate income through R&D contracting is to collaborate with startups the proportion of exclusive licenses to these companies has a positive effect on both the amount of industrial research expenditure and the amount of research funds tied to licensing deals. The amount of research fund tied to licenses is also positively affected by the proportion of exclusive licenses to large companies. Our other independent variables don’t indicate any effect on neither. This indicates the more prudent stance of large companies when investing in university research as they only invest in products they are sure to benefit from and with universities they already partnered.

The number of licenses generating income and the amount on licensing income generated are tied to non-exclusivity. In both cases the results are positively affected by the proportion of non-exclusive licenses to incumbent companies be it large or small.

We also observe a preference for royalty schemes by small companies while large companies favour other types of payments. The results further suggest that licenses are more likely to reach the million-dollar mark when granted to small companies or startups as the proportion of non-exclusive licenses to these companies has a positive effect on the number of licenses generating more than a million dollar in the case of our OLS regressions.

Significance

These results suggest that universities can benefit from using exclusivity as a bargaining chip, exclusivity is only being granted in case of early R&D investment or in exchange of equity when launching a startup. However, like patents, the temporary monopoly granted by the exclusivity can be a considerable advantage for the licensee on the marketplace. No wonder then that companies will not invest in R&D and most startups will not launch without an exclusivity clause.

Universities should reconsider their R&D subcontracting strategies. While large companies have the resource and expertise to enter into university-industry partnership they also have larger networks and can have their own laboratories. Thus, startups are more likely to pay generous amounts because of their lack of other options when it comes to R&D.

Most startups are not launched to generate immediate income but rather equity value growth over a long period of time. Hence, the bulk of licensing income is paid by incumbent companies. It’s however interesting to see that companies don’t seem to pay a premium for exclusivity as one would have imagined. More importantly, just as in the case of R&D, large companies once again seem to have the upper hand compared to their smaller counterparts. While both small and large companies generate licensing income, the proportion of licenses to small companies indicate larger coefficients. Furthermore, startups and small companies produce more licenses generating more than a million dollar. This shows that in the case of licensing for money smaller might be better. These results are probably due to a difference in experience and resources when negotiating contracts.

ABSTRACT. Research Question: Do R&D tax credits increase the levels of innovation in a state’s economy?

Background and Rationale

Since Schumpeter first proposed the notion of creative destruction, the process by which companies founded on new ideas replace companies that are unable to adapt, the rallying cry of economic development is that innovation drives growth. According to endogenous growth theory, research and development (R&D) creates new knowledge that does not solely benefit the firm — instead the knowledge created by R&D spills over into the rest of the economy. States have largely turned to R&D tax credits as the leading policy to incentivize R&D within a state’s economy. While scholars have repeatedly established the relationship between R&D tax credits and R&D as well as the relationship between R&D spending and patenting, limited work has been done to extend the effect of the tax credit past R&D spending to innovative activity itself. Studies focusing on the effectiveness of R&D tax credits tend to consider only these credits on R&D spending by firms. This is problematic because R&D spending represents a budgetary allocation of resources to generate new ideas, and processes, (i.e. innovation) but does not measure outcomes associated with increased R&D spending. To extend our understanding of these tax credits, it is necessary to account for the direct and indirect effect these tax credits have on innovation within a state’s economy. Additionally, most of the studies connecting R&D spending to innovation have relied on patent counts as the sole measure of innovation. Patenting data has some inherent measurement error because not all patents have the same economic impact. In order to address measurement concerns with patents and to measure the direct and indirect effects of R&D credits on innovation, I propose using a structural equation model. Finally, this study presents a novel way of measuring the use of R&D credits in states.

Data & Methods Data for this project comes from a variety of governmental and non-governmental sources and is organized into a longitudinal repeated measure design. The data includes measures from all 50 states and DC from 2001-2015. Data from the states are placed into a panel with each observation representing the i-th state and the t-th state. The key variables of interest are the amount of R&D credits claimed by companies in a state, the amount of R&D spending performed by firms in a state, a state’s GDP, and a factor of innovative activity in a state. Innovative activity is a latent variable measured by the number of patents filed in a state, amount of Small Business Industrial Research (SBIR) grants, amount of Small Business Technical Transfer grants (STTR), and the Kaufman startup index. The R&D tax credit variable was collected using a novel approach. Most studies focusing on the effectiveness of these tax credits measure the tax credit in terms of either a binary variable or the rate of the credit (Becker, 2015). This approach is problematic because it ignores considerable heterogeneity that exists in the way states structure their credits. For example, New Mexico offers a credit only to new firms, while Delaware only offers a credit to firms whose R&D spending reaches a certain level. However, this study uses the amount of R&D tax credits claimed by companies as the measure of the size of the tax credit. This approach has a few advantages over other measures of the credit. Namely it measures the extent to which companies took advantage of the credit and introduces greater variation in the R&D credit variable itself. Data for the amount of R&D tax credits claimed comes from annual state tax expenditure reports. This study leverages a structural equation model in order to understand the relationship between R&D tax credits and innovation. Structural equation models have the benefit of being able to measure the direct effect and indirect effect between R&D tax credits and innovation. Additionally, they allow one to account for measurement error through the use of latent variables. The hypothesized structural model argues that R&D tax credits have a direct effect on innovation, and an indirect effect on innovation mediated by R&D spending. R&D spending also has a direct effect on innovation, and an indirect effect on innovation moderated by overall economic activity (GDP). Autocorrelation is always a concern for longitudinal studies, but particularly problematic for structural equation models. Most preferred solutions for this problem involve the use of wide data, in which each unit has a single observation, and each measure has a different variable for each year. This approach requires a large sample size in order to retain statistical power. Unfortunately, generating a larger sample is impossible in this case as there are only 50 U.S. states. As a result, I intend to use a partially lagged model, where measures in year t are a function of variables from year t-1. This should partially mitigate the autocorrelation problem, though caution will still be urged in interpreting the results.

Anticipated Results I hypothesize that there is both a positive direct and positive indirect effect of R&D tax credits on innovation. Initial econometric analysis of the data is already yielding promising results regarding the relationship of R&D tax credits and patenting. Other initial analysis has confirmed the hypothesized factor of innovation made up of patenting, SBIR awards, STTR awards, and the Kaufman startup index. Next steps include fitting the full structural model and applying rigorous robustness tests and sensitivity analysis.

Significance This study has three important contributions to the literature. First, it leverages a novel way of measuring R&D tax credits. Second, it offers a more robust measure of innovation than previous attempts to use patents as a proxy of innovation. Finally, it measures the direct and indirect relationship between R&D tax credits and innovative activity. The study also offers an important contribution to the wider policy debate surrounding economic development. For years, there has been a large disconnect between economic development in theory and practice. Considerable consensus exists in the literature that few firm-based subsidies benefit the states that offered them. Though, policies that aimed at incentivizing R&D and high-technology firms tend to perform better. However, the spectacle of Amazon’s H2 bidding process and the resulting backlash from communities has invited new scrutiny into these policies, which motivates the salience of testing the efficacy of policies like R&D tax credits.

ABSTRACT. Background and rationale 2018 marked the 10th anniversary of the 2008 Global Financial Crisis (GFC) that started in the United States and was followed by a global recession. The GFC created in the Eurozone an asymmetric shock, affecting particularly six heavily indebted Southern European countries. Not only in these countries but in all OECD countries debt-to-GDP ratio increased substantially.

To reduce budget deficits, government spending was cut. This affected severely the public sector in the heavily indebted countries. With their small open economies the Europe’s Nordic countries (except Iceland) were also confronted with the impact of the GFC, but the Nordic model turns out to be less vulnerable and more resilient.

These cuts in government spending had also an impact on public R&D expenditures and in particular on the public funding of the higher education sector and public research institutes. 10 years after the start of the GFC little work has been done on the impact of the reduction in public funding on the research performance of the higher education sector and the public research institutes.

In the present study, for a panel of OECD countries we investigate trends in the publication output of these institutes and the potential existence of causality between funding and output.

Data In most OECD countries nearly all basic research is conducted at universities and public research institutes and to a large extent funded by public authorities. Although they do not cover the total research output on country level, papers published in scholarly journals processed for the Web of Science (WoS) are often used as a proxy for this output.

The publications are aggregated at country level, applying two counting schemes: the ‘whole’ or full and the fractional counting scheme. The former gives equal weight to the countries mentioned in the by-line of a publication. The latter allocates to each country a fractional weight based on the number of countries in the by-line.

As a proxy for the expenditure on research of a country’s higher education institutes the OECD’s Higher Education Expenditure on R&D (HERD) is used. Similarly a proxy for the expenditure on research carried out by public research institutes and other governmental organizations is the OECD’s Government Expenditure on R&D (GOVERD). The funding data are corrected for differences in inflation rate and purchasing power among countries.

Methods In this study time series analysis is applied on national research expenditure and publication data of a panel of 14 OECD countries. Eleven EU countries were selected based on the degree of severity of the impact of the GFC: Group 1: Countries strongly affected by the GFC: Ireland, Greece, Spain, Portugal; Group 2: Countries with severe problems in their banking sector and large debt to GDP ratio’s: Belgium and Italy; Group 3: The Nordic EU countries, more mildly affected by the GFC: Denmark, Finland and Sweden. Of this group only Finland is a Eurozone member; Group 4: Two leading scientific EU countries, moderately affected by the GFC: France and the UK. Group 5 consists of three OECD non-EU countries, powerhouses in research: Canada, Switzerland and the USA. In a time series an unexpected shift can occur. To test the assumption of a structural break the Chow test is often used. The Chow statistic tests whether a single regression line or two separate regression lines fit the data best.

When studying time series one often looks at correlations to test whether and how strongly two series are related. However, correlation does not imply causation. Causation indicates that one event is the (partial) result of the other. The Granger Causality test (GC) uses a bivariate linear autoregressive model of two variables x_t and y_t:

y_t=μ_0+∑_(i=1)^k▒〖α_1i y_(t-i) 〗+∑_(i=1)^k▒〖β_1i x_(t-i) 〗+ε_1t x_t=φ_0+∑_(i=1)^k▒〖γ_1i x_(t-i) 〗+∑_(i=1)^k▒〖δ_1i y_(t-i) 〗+ε_2t

where k is the maximum number of lagged observations included in the model, μ_0, α_1i,β_1i,φ_0,γ_1i and δ_1i are parameters of the model, and ε_1t and ε_2tare residuals (prediction errors) for each time series. If the variance of ε_1t is reduced by the inclusion of the x_t terms in the first equation, then it is said that x_t Granger causes y_t, i.e. if the coefficients β_1i jointly significantly different from zero.

Preliminary results and discussion For all countries in the panel except Canada, Sweden, Switzerland, the UK and the US, the growth rate of the full-counted publication output was lower in the last 4-5 years of the period 1999-2017 compared to the years before. These trends cannot be explained by changes in the coverage of the WoS. One of the possible causes may be the reduction in public research funding. However there are exceptions. Although the public funding for higher education and public research institutes increased steadily between 1999 and 2017, for Belgium the growth rate of the full-counted publications was lower in the last quarter of the period compared to the earlier years.

For the fractional-counted publication output the reduction in the growth rate was more pronounced compared to the full-counted number of publications and for some countries this growth rate was flat or even negative in the last quarter of the period. The difference in growth rates between the two counting schemes can be explained by an acceleration in the international collaboration. However it is remarkable that this acceleration happened 3 to 5 years after funding reductions started to be implemented in 9 out of 12 countries in the panel. Moreover around the same year it is also observed in three countries were research expenditures grew at a nearly steady rate over the entire period.

Part of the time lag between the changes in the funding and the publication output could be explained by the research cycle: obtaining a grant, carrying out the research work, submitting a manuscript and lastly its publication in a peer reviewed journal. Although there are differences between disciplines, in general we can say that the publication process takes about a year and in most countries the majority of research grants cover 2 to 4 years.

To explore the causal relationship between funding and publication output the Toda-Yamamoto procedure for the GC-test is applied. For the period 1988-2017 at the 10% confidence level for 7 out of the 14 countries a unidirectional causality is found between funding and publication output. At the same confidence level this relationship is even bidirectional for Denmark, Ireland and the US. This results suggest that at a country level changes in public research funding affects with a certain delay the nation’s publication output. But also more counter-intuitively that changes in publication output have an effect on public research funding.

The interpretation of the results of the analysis must however be done carefully. Besides the reduction of a complex world to a model with 2 proxy indicators and the problems related to the use of OECD funding data, the GC-test has its own methodological limitations.

Notwithstanding the necessary caution, for time series analysis powerful analytical tools have been developed in econometrics. To the best of our knowledge they have not yet been used in bibliometric studies.

ABSTRACT. 1. Introduction The Nobel prizes in Physiology or Medicine, Physics, and Chemistry have attracted continued attention from the history of science and quantitative science and innovation perspectives. This paper contributes to a debated yet understudied topic: North America’s global leadership in science. First, there is some disagreement about when North America, and especially the United States, became the global center of science during the 20th century. Some argue that research-intensive North American universities became globally competitive in the 1920s. Others locate the shift after the Second World War, however, pointing to the considerable emigration of eminent European scientists after 1933 and the unprecedented growth of government-sponsored “Big Science”. Second, the enormous growth of scientific research in Asia since the 1990s, especially in China, has sparked discussions about the impending decline of North American hegemony in science. This paper studies the changing global distribution of Nobel science laureates (n=599) over 117 years, distinguishing three world regions: Asia-Pacific, Europe, and North America. We identify when global leadership in science shifted from Europe to North America and when the latter started to lose some of its global share, an indication of growing global competition at the beginning of the 21st century. In this context, we examine institutional factors associated with North America’s scientific strength and more generally with dynamic science systems at the scale of world regions.

2. Material and Methods We compiled a dataset for all Nobel laureates who received the Nobel Prize in Physiology or Medicine, Physics, or Chemistry from 1901 to 2017 (n=599). We collected data on: (a) the year, organization, and country of both the first and the highest academic degrees (HDs); (b) the year, organization, and country of where each laureate’s prize-winning work (PWR) was performed; (c) the year, organization, and country where the laureate worked when awarded the Nobel Prize (designated from here as NP); (d) whether the laureate had a Nobel mentor while a graduate student, postdoctoral researcher, or junior collaborator or during extended research sabbaticals; and (e) year of birth. We determined the absolute and relative frequencies of Nobel laureates for the three career events (HD, PWR, NP). Hence, the distributions in all three career events, including HD and PWR, are mapped onto time periods defined by NP. The three regions are defined by the university or research organization in which HD, PWR, or NP occurred. In addition, we measured the number of master–apprentice relations for Europe and North America. These relations are relevant because they indicate effective transfer of the ability to conduct ground-breaking research from one generation of scientists to another. We also measured the number of universities and research organizations hosting laureates for the first time (“newcomers”) across the three career events (HD, PWR, NP). Finally, we measured inter-organizational mobility: these are the sums of laureate moves from HD to PWR and/or from PWR to NP, respectively.

3. Empirical Results North America’s rise as global power in science started in the 1920s and its leadership position was consolidated in the 1970s. North America’s hegemony in science lasted about 30 years (1970s to 1990s), preceded by a transition period in which it superseded Europe (1940s to 1960s). During North American hegemony, its universities and other research organizations educated up to 60 percent (HD) of all laureates and provided the work environment where roughly 75 percent of all laureates conducted their PWR and received the NP. Yet, beginning in the 2000s, we identified a period of increasing global competition. First, North America’s global share of laureates dropped significantly, across both the three major career events (HD, PWR, NP) and the three disciplines (Physics, Chemistry, and Physiology or Medicine). North America’s shares dropped to 48 percent (HD), 47 percent (PWR), and 58 percent (NP) in the 2010s. At the same time, Europe’s shares consolidated at 37 percent (HD), 36 percent (PWR), and 26 percent (NP), and Asia-Pacific’s shares grew considerably, reaching 15 percent (HD), 17 percent (PWR), and 15 percent (NP). Contradicting speculations about China as a rising global science center, this growth resulted almost entirely from Japan. Second and related, although North America was effective in transferring capacities for conducting ground-breaking research from one generation of scientists to another for a long time (via master–apprentice relations), Europe and the Asia-Pacific region seem to be catching up. During its hegemony (1970s to 1990s), North America accumulated laureate apprentices, reaching its highest level in the 1990s (68%). Then, its share of apprentice–master relations plummeted, reaching 42 percent in the 2010s. At the same time, Europe consolidated (37%) and the Asia-Pacific region increased (16%) their respective shares. Regarding inter-regional migration of scientists, we found that during its hegemonic period (1970s to 1990s), North America attracted on average about 11 laureates every 10 years from Europe and the Asia-Pacific region. The net migration effect for North America remained constant over the hegemonic and the post-hegemonic periods, when North America attracted around 12 laureates per decade. Finally, we examined the share of “newcomers” entering the inter-organizational competition for scientific talent: research organizations hosting laureates for the first time in either of the three career events (HD, PWR, NP). Considering all three career stages combined, there is a general trend of decreasing shares of newcomer organizations across the three regions over the entire observation period (1901–2017). Two findings are noteworthy in this regard. First, the share of newcomer organizations across all career events combined is significantly higher in North America compared to Europe not only during the catch-up period (1920s to 1940s) but also during the hegemonic period (1970s to 1990s). Second, we found that both Europe and North America are more dynamic in the organizational context in which prize-winning research is conducted. In contrast, newcomers are less important among the set of universities and research organizations where future laureates either receive their Ph.D. and/or their NP.

4. Discussion Our results on the Nobel population are broadly corroborated by recent bibliometric comparisons that show a declining citation impact gap between North America and other world regions since the 2000s. In particular, several cross-country bibliometric comparisons show that while North America was in the lead until the early 2000s, its share among both the 1 percent and the 10 percent most highly cited publications has declined since the mid-2000s. In the meantime, countries from Europe and the Asia-Pacific region have increased their shares, respectively. Thus, our analysis complements existing bibliometric cross-country and cross-regional comparisons. Future research could focus on systematic cross-regional comparisons based on archival data on R&D expenditure and the scientific work force at the national and/or regional level. These data should include measures such as number of inhabitants, number of employed scientists, gross domestic product, or patent applications and citations. It would be interesting to know whether and to what extent the overall picture, as shown here (and in other bibliometric comparisons), would change if such measures were to be taken into account.

ABSTRACT. Background It is well-known that US universities are better placed than their European counterparts in international rankings. This happens despite the fact that, on the aggregate, the European university system produces more scientific papers and even receives more citations. This ‘transatlantic gap’ in international research visibility has been explained by differences in the institutional environment. While the US system would promote differentiation of university profiles and allow for concentration of excellence in a few top research universities, in the European system excellence is scattered with individual universities having only few top-research departments. This would be explained by a funding system in which research resources are largely provided as a supplement to student funding and, therefore, excellence is not financially primed in terms of resources. The goal of this paper is to provide more a more in-depth analysis of the differences between the US and the European university system. We move from a recent study showing that most of the differences in international research visibility is accounted by the presence in the US of a small number of highly funded research universities that consistently show at the top of the international rankings. This analysis also identifies the concentration of research funds from private donors as a potential mechanism for the emergence in the US of this group of universities, which has almost no counterpart in Europe. These results suggests, first, that the observed differences between the two systems might be related to the presence (or absence) of groups of universities with a specific profile, while the two systems might be more similar for what concerns other groups like universities with a strong educational component (public universities in Europe vs. State universities in the US). Second, they hint that different mechanisms account for the emergence of groups of universities and that, therefore, adopting policies tailored to the (few) top-research universities at the system level might not be optimal, but differentiated policies are required taking into account the different mixes of activities and goals of each type of institutions. Approach First, we develop a classification of US and European universities into distinct types. The underlying assumption is that only a limited number of configurations of activities emerge, associated with interdependences between the types of resources and of outputs. Second, we compare the classes in terms of their activity profile (education vs. research vs. third-mission) and of resources. Based on these findings, we speculate about the mechanisms accounting for the (differential) presence of classes in each region. Third, we discuss the implications for the design of research and higher education policies. Data and sample Our dataset is derived from the Integrated Postsecondary Education Data System for the US (IPEDS; http://nces.ed.gov/ipeds/) and the European Tertiary Education Register dataset (ETER; www.eter-project.com), which provide a highly representative sample of universities in both system, as well as data on finances, staff, students and graduates at the institutional level. The database has been enhanced with bibliometric data derived from the Web of Science version at CWTS in Leiden and with patent data derived from the PATSTAT version at IFRIS in Paris. This is therefore the first combined dataset for US and European universities, which includes data both on the input and on the output side. We construct our sample by adopting the criterion of the Carnegie classification for ‘research universities’, i.e. those universities that awarded in the reference year at the least 20 PhD degrees. According to 2014 data, this sample is composed by about 1,000 universities, including nearly all top-500 institutions in the Shanghai ranking and is therefore very representative of international excellence. Unlike the Carnegie classification, the sample also includes about 80 specialized institutions, mostly medical schools and some specialized universities in technology, business and theology. Variables The selection of variables is rooted in the higher education comparative literature. First, the activity profile in terms of three main activities and outputs, i.e. education, research and third-mission. While traditionally it has been considered that education and research are associated in universities, the differentiation in the HEI landscape implies that the relationships between the two activities might vary by HEI types. Moreover, third-mission has become an important dimension of HEI activities, up to the definition of Entrepreneurial HEIs as a distinct class of institutions. Accordingly, we observe these dimensions through distinct indicators: the number of students for education; publications, citations and PhD degrees for research; patents for third-mission. Second, the subject scope, i.e. the diversity of the subject domains covered by HEI activities. Subject specialization is relevant for market positioning: HEIs that are active in many subjects cover a broader range of educational demands and tend to have larger enrolment, while specialized HEIs might leverage their distinctive identity to attract students. The literature distinguishes between the generalist HEI covering most subject domains, and the specialist HEI, whose identity is defined by the subject. We therefore introduce in the model an indicator of subject concentration of students, as well as measures of the HEI orientation towards different subject domains. Further, we include organizational size as measured by the number of staff given that it strongly impacts on activities and is a proxy for the ability of HEIs to acquire resources. Finally, we take into account two regulatory (exogenous) characteristics that are likely to influence the HEI profile and classification, i.e. the legal status of the HEI (public vs. private) and the research mandate, as represented by the legal right to award a PhD. We further introduce financial variables, i.e. the volume of revenues and their composition (core vs. third-party, public vs. private), as descriptive information to characterize the classes identified by the activity profiles in terms of their resourcing behavior. Methods To attribute HEIs to classes, we use latent class clustering. This class of models fits the distribution of a set of observed variables conditional to the observations belonging to non-observed (latent) classes; compared with conventional clustering methods, latent-class clustering can incorporate prior assumptions on classes and statistical distributions. The model computes the distribution functions and the posterior probability for each HEI to belong to a class and searches iteratively for the solution maximizing the model fit. It must be run with a pre-specified number of classes, but the optimal number can be selected ex-post by comparing fit statistics. To interpret results, we assign each case to the class with the highest probability and we compute descriptive statistics by class and by region (Europe vs. US). Expected results At the empirical level, this work will develop a comparative typology of research universities, providing evidence of the presence (or absence) of groups in the US vs. Europe. We specifically test the hypothesis that much of the difference in international visibility is related to the presence in the US of a class of research universities with a volume of resource well-above the average university (in absolute terms and relative to students), which have no counterpart in Europe. At the policy level, differentiating groups (and their revenue structure) is expected to provide information for the design of more tailored policies. Such policies might be more effective in addressing the multiple goals of higher education and balancing excellence and access, concentration and regional outreach.

ABSTRACT. The Knowledge in the Making (KNOWMAK; www.knowmak.eu) project has developed an interactive tool, which allows users to visualise and analyse the production of knowledge in the European Research Area (ERA), with a focus on knowledge related to Societal Grand Challenges (SGC) and Key Enabling Technologies (KET). This presentation is intended to introduce the tool, discuss its characteristics as well as the indicators that can be visualized and extracted for further analysis. Further, based on preliminary analyses, we will provide examples of how the tool can be exploited for deeper explorations of knowledge production.

Backgound Understanding knowledge co-creation in key emerging areas of European research is a critical issue for policy makers. However, current methods for characterising the field have limitations concerning the nature of research and the differences in language and topic structure between policies and scientific topics. A central issue is how to map diverse kinds of knowledge outputs to topics in the science policy debate. Traditional classification systems are simple, stable, and have widespread coverage. However, combining such schemes in order to depict an overall view of scientific knowledge production is challenging. Furthermore, mapping these classifications to policy-oriented topics presents a further issue due to terminological and conceptual divergence. On the other hand, textual methods such as keyword extraction and overlay maps, provide fine-grained views of scientific fields, but are not easily scalable to broader topics and remain bound to the specific language of each source. We address these problems through the use of ontologies. Ontologies share with classifications the fact that they are constructed upon some intellectual understanding of reality. The ontology that has been developed in the KNOWMAK project is organized around two central topic areas to European policy makers: Key Enabling Technologies (KET; https://ec.europa.eu/programmes/horizon2020/en/area/key-enabling-technologies) and Societal Grand Challenges (SGC; https://ec.europa.eu/programmes/horizon2020/en/h2020-section/societal-challenges). From a policy perspective, KETs are considered as essential generic technologies for the future competitiveness of the EU, and SGCs as the knowledge domains crucial for the major societal challenges of the future. These 13 topics have been further disaggregated in around 150 subtopics, such as ‘Genomics’, ‘Energy efficiency’ and ‘Social inequality’. Keywords have been generated semi-automatically from various sources. Finally, the source documents (publications, patents, research projects, social innovation projects) have been assigned to the ontology classes based on the combined weighted frequency of keywords.

Methodology and empirical base KNOWMAK integrates data sources on knowledge production, including data on scientific publications derived from the Web of Science version at the University of Leiden, on European collaborative R&D projects from the EUPRO database developed at the AIT Austrian Institute of Technology (Roediger-Schluga and Barber 2008), and on patents from the PATSTAT version at IFRIS in Paris. Additionally, the project is integrating two additional data sources, one concerned with social innovation projects, the other with user attention as observed through social media. All source data have been geolocalised based on the authors’ (publications), participants’ (projects) and inventors’ addresses (patents). This allows for a flexible attribution to regions. A new regional classification has been developed to address some issues of the conventional EUROSTAT NUTS regions (https://ec.europa.eu/eurostat/web/nuts/background) and to develop a regional classification that recognizes the central role of metropolitan areas in knowledge production. The geographical coverage includes EU-28 countries, EA-EFTA countries (Iceland, Liechtenstein, Norway and Switzerland) and EU candidate countries (Albania, Former Yugoslav Republic of Macedonia, Montenegro, Serbia and Turkey) for a total of 553 regions. By combining this regional classification to the assignment of data to classes in the ontology, it becomes possible to generate indicators of knowledge production for various spatial entities by topics and subtopics. This includes for example the number of publications, patents and projects, but also quality-related indicators (such as publications in the top-10% cited) or indicators on the network centrality of regions. Further, composite indicators can be constructed, such as the overall share in knowledge production (combining volumes of publications, patents and projects), as well as the intensity normalized by the population.

Results a) The geographical distribution of knowledge production by region. The KNOWMAK tool allows evaluating how knowledge production is distributed in Europe and what are hot spots in terms of production volume (as a share of the European total) and intensity (normalized by population). This analysis shows that regions with higher knowledge production volumes are mostly concentrated in large metropolitan regions, with Paris, London and Munich ranking in the first three positions. On the other hand, medium-size metropolitan areas like Eindhoven (in the Netherlands), Vlaams-Brabant (Leuven – Belgium) and Uppsala (in Sweden) rank in the first three positions in terms of production intensity, while still having rather large volumes of knowledge production. This emphasizes the important role of such medium-size regions, which is likely to emerge even more clearly when analysing specific domains. b) Analyzing knowledge production in a specific KET or SGC in terms of different types of outputs (publications, patents, projects) and of the relative importance of subtopics. In the case of Nanotechnology, the KNOWMAK data display systematic differences between different knowledge production by type of data, with nanoscale devices being the most important subtopics for patents, nanoscale technology for publications and nanoscale materials for projects. This highlights differences in the science vs. technology orientation of domains within a KET, but also potential misalignments between EU funding policies and the European S&T basis. c) The relative importance of scientific vs. technological production in European regions. In the case of Genomics, we display systematic differences in the geographical distribution of knowledge production, with patents being clustered in a few regions in Western and Nordic Europe, while publications are more distributed across space, including some regions in Eastern countries. This is likely to reflect different geographies of public research vs. industrial innovation, as the top-technological regions in genomics are the seat of some of the largest pharmaceutical companies in Europe. In a future release of KNOWMAK, it will also be possible to identify the main R&D actors in each region in order to better understand the observed patterns.

Conclusions The examples presented clearly highlight some innovative characteristics of the tool: • The development of an ontology allows for a fine-grained analysis of knowledge production at the topic level (13 KET/SGC, about 150 subtopics) that is common across data sources and connected to political priorities at the European level. This moves beyond current topical analyses based on a single data source or on fixed classification schemes. • The regional classification developed by KNOWMAK is better suited to the geographical analysis of knowledge production, while it remains compatible with EUROSTAT-NUTS and, therefore, allows integrating regional statistics in the analysis. The combination of regional and topical breakdowns is powerful to investigate issues such as regional specialization, specifically interesting in context of the regional ‘smart’ specialization debate. • KNOWMAK has undertaken a systematic process of harmonization of data sources in terms of geography, topics and actors, which allows for fine-grained analysis combining different data sources on the same units of analysis. Eventually, this will also enable the development of smarter composite indicators. • Finally, KNOWMAK has developed a user-friendly on-line tool that allows visualizing the data, displaying different types of maps and charts and, finally, downloading the data for further analysis. This will contribute to widespread usage beyond the scholarly community.

ABSTRACT. Background and rationale

Thomas Kuhn proposed that science is "a series of peaceful interludes punctuated by intellectually violent revolutions" (Kuhn, 1963). He argued that revolutions are essential for the most impactful advances in science. Part of this idea has percolated into the goals of public and private funding agencies, which tend to favor innovative proposals over incremental ones. NSF and NIH make this goal explicit in their grant submission instructions (National Institutes of Health, 2018; National Science Board, 2007; National Science Foundation, 2019). The European Research Council (ERC) also favors innovation, stating in their 2016 annual report that investment in frontier research has enabled breakthrough discoveries (European Research Council, 2017), and the National Natural Science Foundation of China (NNSFC) follows a similar pattern (National Natural Science Foundation of China, 2019). However, we know little about whether innovative grants lead to impactful new scientific discoveries when compared to more incremental ones.

While previous research has found that it does exist a definite link between the novelty of publications and number of citations, it is unclear how to apply such analysis to grants. One difficulty is that at the time of submission, a grant only has preliminary analyses completed. Also, grants' citations are usually hidden once funded, making it difficult to analyze using traditional network science. We hypothesize however that similar trends exist between the innovation of a grant and the future impact of the publications and that one possible approach to measure it is to use text analysis and novelty detection techniques from machine learning. In this work, we examine whether Kuhn proposition for violent revolutions also applies to research proposed in grants.

Measuring innovation in science has been historically done by measuring how unusual ideas are combined in a piece of work. Uzzi, Mukherjee, Stringer, and Jones (2013) used the concept of unusual citation combinations to estimate "novelty." A similar approach was used by Wang, Veugelers, and Stephan (2017) to find the relationship between novelty and impact in scientific publications. These analyses can be logically extended to grants. For example, we could take the text in grants' abstracts to estimate how unusual the topics in such text are from grants previously funded. An unusual combination of ideas can, therefore, be measured using this text.

Materials and methods

We now describe the grant and publication data we use and how we measure novelty in the text of the grants.

Materials: We collect 99,053 NSF grants and 604,021 NIH grants (e.g., Project Number, Abstract, Year of the award, PIs, Total Cost) from Federal RePORTER in the period 2008 to 2015. We also collect grants 449,636 unique NSF publication records from NSF and 633,333 unique publication records from NIH. We only analyze R01 research grants from NIH—R01 research grants are the most common type of grants of NIH—leaving us with 224,108 NIH grants.

Before we analyze the data, we find that some grants appear as continuing grants in NSF and NIH. Continuing grants area not by definition new grants and therefore were eliminated from the analysis after their first funding cycle. We were left with 91,905 NSF grants and 61,233 NIH grants.

We retrieve citation data by using the Scopus API. We found 112,016 articles associated with NSF grants and 140,481 articles with NIH grants. We additionally obtained information about journals and knowledge areas. We match the journals from Scopus to their journal ranking (SJR ranking indicator) and knowledge areas in Scimag (De Domenico, Omodei, & Arenas, 2016; Sinha et al., 2015). This external dataset provides excellent coverage of the data reported to NSF and NIH agencies.

Methods: To measure the novelty in the text of grants, we use a technique from machine learning meant to estimate the novelty of data points. The basic idea behind this technique is to learn the distribution of a training dataset (e.g., past grants) and estimate the degree to which a new data point (e.g., a recently awarded grant) is novel. We represent grants by using topic models. We perform a non-negative matrix factorization of the term frequency-inverse document frequency (tf-idf) matrix and use a technique called one-class support vector machines ((Schölkopf, Platt, Shawe-Taylor, Smola, & Williamson, 2001) to learn the distribution of grants. After rescaling, the method produces values between 0 and 1, where 0 is an entirely incremental grant, and 1 is an entirely original grant.

Results

Innovative grants produce more impactful research

We use a linear regression model to understand the effect of grant novelty on future citations. We control for the journal's knowledge area, award amount, PI experience as measured by whether they have been funded in the past, the prestige of the journal (SJR index), and the interaction between agency and topic novelty. We run individual regression models on NIH and NSF grants for this purpose. For NIH, the R-squared of the model is 0.313. More specifically, we found that novelty has a significant positive effect on future citations (t(105,230) = 7.447, p < 0.0001). For NSF, the R-squared of the model is 0.2831, and we found that novelty has a significant positive effect on future citation (t(71,014) = 5.318,p<0.0001). Across these two funding agencies, our results suggest that a completely original grant would obtain 50% more citations than a completely incremental one. These results suggest that there is a strong relationship between novelty and innovation in grants.

Qualitative analysis of some highly novel grants of the past

We also performed a qualitative analysis of the results. For example, our method gave a very high novelty score to the NSF grant #1307429 "Toward Gravitational Wave Discovery With Advanced LIGO" which led to a Nobel prize for the discovery of such Gravitational Waves. Also, we were able to identify RAPID, CAREER, and EAGER as having a higher number of citations (although not significantly) compared to other grants. These observations suggest that topic novelty can be used as a valid technique as a proxy for the novelty of a grant.

Significance and conclusions

Our results show that we can effectively measure the innovation of grant using the text summary of it. We use machine learning techniques to measure the novelty of the natural language used in the summary of the grant and correlated it with the future citations of publications associated with the grant. Our results suggest that innovation in grants is an essential factor driving the impact of publications after controlling for several factors related to the field, researcher, and award amount. Future work will examine the mechanisms behind such a relationship.

Overall, our results strongly favor the policies of many funding agencies in that more innovative grants do have a higher impact. Future work will look at the interplay between incremental research and innovative research. Perhaps, scientific revolutions would not exist without the "peaceful interludes"—incremental research—alluded to by Kuhn.

ABSTRACT. Knowledge diffusion is a key metric in measuring the impact of investments in science and technology. Increased attention has been paid to the evaluation of the outcomes of government-sponsored research. Observable milestones in the life of a technological idea include (1) R&D funding, (2) patents, and (3) licensing agreements, each a measurable step in the maturation of knowledge derived from R&D. Recovering connections between these milestones is difficult, because there is no persistent identifier that is assigned when funding is awarded that then carries through to licensing; however, we can make connections between the steps using unique identifiers such as award numbers and patent numbers. This project links the knowledge flows through the R&D system, producing a linked set between R&D funding, patents, and indicators of use such a patent licensing agreements.

In this project we have collected and linked three types of public data: federal government grants and contracts, U.S. patents, and patents mentioned in SEC filings, creating a unique, novel data set. The government interest section of a U.S. patent is used to link patents to federal government investments. Data on federal government investment is found in grant databases and the federal procurement data system. Patent licensing is opaque and not generally available; however, reports of licensing agreements (and other patent events) are filed with the SEC. This project uses machine learning to extract patent mentions and classify them (granted, licensed, litigated, etc.), building a database of patents that are valuable enough to be mentioned in the SEC database.

We have processed all quarterly filings made available by the SEC extending from the present back to 1995, when public electronic records were first made electronically available. We have extracted patent references from this database of more than 12 million records. We link those patents to federal government investments through the government interest section in the patent record. We will present in-depth analysis of work extracting patent mentions from the SEC database and linking them to federal government investments as recorded in the following database: • National Science Foundation Awards Database • National Institutes of Health exPorter • USAspending.gov • Department of Energy’s Portfolio Analysis and Management System • Department of Health and Human Services’ Tracking Accountability in Government Grants System • Federal Procurement Data System • Small Business Innovation Research and Small Business Technology Transfer awards database The presentation will detail the knowledge flows from federal government investments into the public market and discuss the relevance of this data set.

**SciSIP support** Received email invitation to submit

ABSTRACT. In December 2018, the National Institute of Standards and Technology (NIST) published a green paper outlining a Return on Investment (ROI) Initiative for Unleashing American Innovation, whose goal is to “move the Nation to a new level of innovation performance that will increase the taxpayers (sic) return on their investment in federally funded R&D.”

This goal is similar to that for the Bayh–Dole Act (BDA) or Patent and Trademark Law Amendments Act, which was passed almost forty years ago (1980): the utilization and commercialization of inventions made with government support.. . . a greater return on the billions of dollars spent each year by the Government on its research and development programs.

Both the BDA and the ROI Initiative recognize that inventions increase welfare when they are diffused, rather than when they are invented. Accordingly in the BDA, the principal instrument was the provision allowing universities, small businesses and non-profit institutions to patent federally funded research. It was felt that such patent protection was necessary for firms to invest in further R&D to commercialize university inventions. This assumption regarding the need for patents is carried forward in the ROI initiative: “The discoveries that result from American R&D efforts must be transferred from the laboratory to the marketplace through innovations that bring products and services to consumers more quickly. Protection of intellectual property rights is often necessary to achieve this transfer by establishing partnerships with industry for commercial adoption”.

Because the details of the ROI initiative are still being fleshed out, and because its goals and assumptions are similar to the BDA, it is an opportune time to examine the extent to which the BDA accomplished its goals. Previous research examining the BDA’s impact, has focused primarily on university activity (patenting and licensing). Because the BDA allowed both patenting and licensing, it is not surprising that both increased following its implementation (Mowery et al. 2001). Note however that while licensing revenue has grown, it is dwarfed by research revenue ($1.4 Billion in licensing revenue in 2000 versus $28 Billion in grant revenue). If universities captured the full value of commercializing federally funded research, this would indicate a negative return of -95% on that research. Fortunately, the patent increase has not come at the expense of publications which have maintained their historic growth trajectory.

If we were to look strictly at the negative return on research grants at universities, then the BDA appears to have failed in its goal to stimulate a greater return on the billions of dollars spent each year by the Government on its research and development programs. This would be a myopic interpretation of the returns to federal research however, because universities cannot commercialize inventions. Moreover, firms will not license technology unless they can earn profits in excess of the licensing fees, so licensing fees understate commercialization. Accordingly, we assess the impact of the BDA on the commercialization of federally funded research by looking downstream at companies—the entities who can commercialize innovation. We look both at patterns in firm utilization of university research as captured by their patent citations to that research, and at the extent to which university research affects commercialization of firm R&D.

Using university research as a proxy for federally funded research, we find that its use by firms increased following implementation of the BDA . However the rate of increase matched that prior to the BDA. Not only did firm citations to university patents increase (which would be expected because that would not have been possible prior to BDA), so too did citations to university publications. Thus, one benefit of the BDA may have been the creation of university Technology Transfer Offices (TTOs) with the consequent increased visibility of university research within the industrial community.

Having said that, the larger goal of the BDA was to increase the commercialization of university research. We tested the impact of BDA on commercialization through a quasi difference-in-differences specification where we treated firms who utilized university research prior to BDA as a treatment group and compared their commercialization of university patents relative to their commercialization of university publications. We then compared their commercialization of patents and publications to a control group (firms who did not utilize university research until after BDA).

We found no evidence that the BDA increased commercialization of university research using our measure of firms’ commercialization of innovation. While BDA had no impact on the control group (no prior experience with university research), it appears to have decreased commercialization of university patents in the treatment group (those with prior experience with university research). Moreover the negative impact of university research on commercialization within the treatment group was restricted to university patents (the policy instrument within the BDA intended to increase commercialization). The impact of university publications on firm commercialization was insignificant.

Taken together, these results suggest that while the BDA may have increased the visibility and use of university research, we cannot reject that null that it had no impact on commercialization. Instead, it appears that providing university research with intellectual property rights may have led to lower commercialization than allowing the research to remain in the public domain. The question is why. One possible explanation is that companies substituted university research for their own R&D. If so, then R&D investment that formerly went to in-house research, where it enhances the firm’s capability to do research, is now going to universities in the form of licensing fees. This may compromise firms’ R&D productivity.

An alternative explanation is that absorbing the knowledge underlying university patents is more difficult than absorbing knowledge in the public domain (publications). Perhaps university research is embodied knowledge that requires co-development rather than arms-length acquisition. The fact that patents were detrimental, while patents were not, seems to argue against this. However, we explored this conjecture in greater detail in a single industry (pharmaceuticals). We utilized a websearch to determine which firms funded university research (as an indicator of knowledge co-development). While we were only able to code one-third of the firms, our preliminary results went in the wrong direction—the impact of university patents for firms who funded university research was more negative than for those who did not.

A final explanation is that creating a fence around the commons of university research makes its knowledge less accessible. Indeed, informal interviews with two Chief Technology Officers indicated that working with TTOs is so cumbersome, they only do so when they lack other alternatives. This explanation is most consistent with the difference in results between university patents and university publications, since patents require use of the TTOs, while publications do not.

Note this discussion pertains to the aggregate results. There is substantial heterogeneity in results across sectors. Exploration of those differences may provide greater insight into why firms’ success with university patents differs from that for university publications. Alternatively, it may identify strategies for restricting the domains which the BDA (and the new ROI Initiative) governs.

ABSTRACT. The prestige of an academic institution is closely related to most measures of the quantity and quality of its faculty's scholarly outputs. Faculty at more prestigious institutions produce more of the scientific literature, receive more citations and scientific awards, and train more of the faculty hired by other prestigious institutions.

Background.

The origin of these imbalances is often attributed to the competitive nature of the academic job market, which serves to sort individuals into positions at different institutions according to their reputation and record of past achievements, including their publication and citation counts. Reputations and achievements may be influenced by meritocratic characteristics, such as an individual's skill, effort, or potential, by non-meritocratic characteristics like age or gender, or by external factors such as work environment, social connections, or even chance events. Untangling the influence of these factors has proved difficult because of endogenous cumulative advantage, in which past achievement broadly correlates with future achievement. Identifying the social mechanisms that explain individual and institutional differences in scholarly output would clarify the degree to which academia operates according to meritocratic principles, inform efforts to address persistent social inequalities in academia, and ultimately expand scientific discovery.

The fact that more productive individuals tend to have been trained at prestigious institutions and also currently work at other prestigious institutions presents a causal puzzle: which is more important in explaining their greater productivity (number of publications) and prominence (number of citations), where they trained or where they work? To answer this question, we infer the causal effect of each environment on scholarly output, treating as a quasi-natural experiment the discontinuity in an individual's circumstances that is caused by moving from their doctoral institution to their faculty institution.

Data and Methods.

We construct matched-pair experiments from a unique and comprehensive data set that documents the doctorate-to-faculty transitions of 2453 tenure-track faculty at all 205 Ph.D.-granting computer science departments in the United States and Canada, spanning 1970-2011, along with complete records of their scholarly output through 2017, encompassing more than 200,000 publications and 7.4 million citations.

The annual faculty job market generates two kinds of quasi-natural experiments: it co-locates at the same institution individuals who trained at more or less prestigious institutions than each other, and it separates individuals with similar training into faculty appointments at more or less prestigious institutions than each other. To isolate the effect of prestige differences on post-hire productivity and prominence in each case, we combine exact and caliper matching techniques to mitigate the confounding effects of differences in the age, gender, subfield productivity norms, and postdoctoral training. If where an individual trained determines their early-career scholarly output, individuals with more prestigious training should be, on average, more productive and more prominent than co-located peers with less prestigious training. On the other hand, if where an individual works determines their early-career scholarly output, individuals with appointments at more prestigious institutions should be more productive and more prominent than similarly trained peers with appointments at less prestigious institutions.

Results.

For matched pairs of faculty with appointments at similarly prestigious institutions, the individual with the more prestigious training was not more productive in the first five years post-hire but received, on average, 301 more citations during this period. Among the pairs, the individual with more prestigious training was more productive in 52.1% of trials but more highly cited in 63.9%.

In contrast, for matched pairs of faculty with similarly prestigious training and with similar pre-hire productivity and prominence, the individual with the more prestigious appointment produced, on average, 5.1 more papers in the first five years post-hire, with 57.4% of trials exhibiting an advantage of any magnitude and significant differences in years {1,2,4,5}. Similarly, individuals with the more prestigious appointment received, on average, 344 more citations in this period.

Hence, conditioned on an individual holding a faculty position somewhere, we find no evidence that training at a prestigious institution confers any advantage to an individual's subsequent productivity while it does lead to marginally significant increase in prominence relative to peers with similarly prestigious faculty appointments. Furthermore, we find strong evidence that the prestige of an individual's faculty appointment, or correlates thereof, drives both their early-career productivity and prominence. That is, where an individual works---not where they were trained---explains the quantity of their scholarly output, and both environments contribute to their prominence.

The precise manner by which institutional prestige controls post-hire productivity remains unknown. Prestigious institutions could create environments that lead to higher faculty productivity through four different mechanisms, based on selection, expectation, retention, or facilitation. Institutions could (i) select inherently more productive faculty via hiring; (ii) require that all faculty meet high expectations for productivity; (iii) selectively retain more productive faculty at tenure or other formal evaluations; or (iv) facilitate productivity by providing a conducive working environment. We investigate the degree to which each of these four mechanisms can explain the observed prestige-productivity effect.

We find that mechanisms based on selection, expectation, or retention each provide at best weak evidence that higher levels of productivity at prestigious environments simply reflect more stringent requirements for faculty. These results leave the majority of prestige's effect on productivity and prominence to be explained by the fourth mechanism, departmental facilitation and its variation with prestige. Under facilitation, the characteristics of a department, such as its location, resources, and organization, enable or constrain the productivity and prominence of individual faculty, through more specific mechanisms. Applying explanatory modeling, our results suggest an important role for several aspects of a department's composition (including size, student-faculty ratios, and support staff) in driving scholarly outputs, each representing a testable hypothesis for further investigation.

Significance.

The emerging field of the science of science aims to develop a causal understanding of the social drivers of scientific discovery, which will improve the evaluation of and investment in good science. A common assumption is that faculty's scholarly productivity mainly reflects their scientific skill, which is often assumed to correlate with the prestige of their doctoral institution. Here, we show that this assumption is false: for early-career faculty, the characteristics of their working environment, and not the prestige of their doctoral training, drive their productivity, and the greater productivity of faculty in more prestigious departments cannot be explained by the preferential selection or retention of more productive scholars. Separately, faculty prominence is influenced by both training and work environments, allowing individuals to benefit from the prestige of either location.

More broadly, our findings have direct implications for research on the science of science, which often assumes, implicitly if not explicitly, that meritocratic principles or mechanisms govern the production of knowledge. Theories and models that fail to account for the environmental mechanism identified here, and the more general causal effects of prestige on productivity and prominence, will thus be incomplete. The causal importance of working environment indicates that past findings in the science of science should likely be reevaluated in light of this effect, and future studies should more explicitly account for it.

ABSTRACT. Background and rationale: Extreme competition for jobs, publications, and grant funding has been a defining feature of biomedical research for well over a decade, and this hypercompetition has outsized negative effects on the careers of young researchers. Focusing on academic faculty careers, competition is fierce: nearly 80 percent of newly minted biomedical Ph.D.s enter a postdoctoral research position and more than half of these new postdocs intend to pursue a faculty position, despite this being far more postdocs than there are faculty slots. Securing an F32 postdoctoral fellowship or a K-series mentored career development award from the National Institutes of Health can increase a postdoc’s chances of securing a faculty position and future funding, while having high-profile publications can be a predictor of success on the faculty track.

However, our understanding of what makes a successful biomedical faculty candidate is still incomplete. For example, the NIH has nearly 30 different awards that trainees can apply for, and while receiving specific training awards confers an advantage in advancing along the faculty track, it is not known if all training awards are equally beneficial in this regard. Graduate students, medical students, residents, and postdocs are eligible for NIH fellowships, designated as F-series awards. The K-series NIH career development awards are directed to medical residents, postdocs, and some faculty. As not all F and K-series awards have the same mission in supporting trainees, it is important to know which mechanisms predominantly support those who go onto successful faculty careers.

Methods: I analyzed publicly available grant information from the National Institutes of Health to determine the grants first-time NIH R01 (FTR01) awardees held during their training period. The R01 is an NIH grant of sufficient size and duration to sustain an independent research program, are highly sought by young faculty, and support more young faculty than any other NIH grant mechanism. The data available through public databases allowed me to connect individual R01-investigators to their training awards, their training and R01 institutions, and the years in which their awards were received. I found there were five training awards, one F-series award and four K-series awards, that were highly represented among FTR01 awardees. To better understand how institutions valued faculty candidates with these specific training awards, institutions were divided into quartiles based on how many FTR01s they received. I then analyzed the distribution of early-career researchers with specific training awards with regard to their training institution and their R01 institution.

Results: The proportion of the FTR01 population that held a training award increased steadily from about 23 percent in 2000 to 35 percent in 2017. However, because of NIH regulations, foreign-born FTR01 awardees were not eligible for most NIH training awards. Taking this into account roughly 50 percent of FTR01 awardees in 2017 who were eligible for an NIH training award had one.

The changes in the FTR01 awardee pool that held training awards could be ascribed to changes in five specific grant mechanisms: the F32 postdoctoral fellowship, the K01, K08, and K23 mentored career development awards, and the K99 pathway to independence award. Between 2000 and 2017, the percentage of FTR01 awardees with these training awards remained constant, but the total number of K awards increased while F32 awards fell. These data indicate the changes in FTR01 awardees with a training award is mostly due to changes in the availability of the training awards.

Next, I analyzed the distribution of FTR01 awardees with specific training awards across institutions. Top institutions preferred hiring trainees with K01, K08, and K23 awards, while FTR01 awardees with an F32 or K99 were evenly distributed across the research enterprise. Importantly, a majority of K01, K08, or K23 awardees received their training award and first R01 at the same institution. When examining only those who changed institutions, the distribution of K01 awardees more closely resembled that of F32 and K99 awardees. Finally, postdocs with training awards at top institutions were overrepresented in the faculty ranks of the research enterprise, regardless of mechanism.

Significance: Understanding the contours of hypercompetition and the realities of the environment in which today’s young scientists compete for funding, publications, and jobs is critical to develop policies that will improve the biomedical research enterprise. The data presented here explore a persistent perception among young researchers that NIH funding is a prerequisite for attaining a faculty position. While there has clearly been an increase in the number of first-time R01 awardees who had NIH funding as a trainee, the majority of FTR01 awardees attained their faculty position without having had prior NIH funding.

An exploration of the distribution among institutions of FTR01 awardees with the five most common training awards showed previously unappreciated complexities with regard to success rates and physical transitions to faculty positions. For example, K01, K08, and K23 awards are available to postdocs and to early-career faculty. One explanation why so many FTR01 awardees with a K01, K08, or K23 award stay at the same institution between their K and R01 is that they are already faculty members when they receive their K award. Additionally, F32 and K01 awardees are awarded to Ph.D.s while K08 and K23 awards are made to M.D.s. These eligibility differences partly explain the differential distribution of training awardees across the enterprise.

These data suggest a self-reinforcing model where trainees seek training awards to distinguish themselves from competitors for faculty positions, institutions favor faculty candidates with training awards, and this increases the pressure on trainees to attain a training award. Future work should overlay non-public information, like the numbers and types of grant applications submitted by trainees and faculty, with publicly available information, like that presented here. For example, do F32 awardees have to submit more or fewer applications before receiving an R01 than K awardees? And have the application and funding dynamics for those without training awards changed substantially over time?

This work is an important step in understanding the multifactorial interactions among a faculty member’s training award track record, their publication history, and their training location and how they play into where they land a faculty position and how this affects R01 applications and awards. Furthermore, understanding how these forces shape hypercompetition is needed to craft policies that effectively relieve these pressures.

ABSTRACT. Objective:

Historically Black Colleges and Universities (HBCUs) play a fundamental role in the American Higher Education system. Perna et al (2009), for example, show that though HBCUs comprise 3% of all institutions of higher education, they are responsible for 30% of all African American bachelor’s degrees in STEM fields. Substantially less attention is paid to recruitment and retention issues related to faculty at HBCUs. HBCU faculty are more diverse in terms of race, ethnicity and national origin than other institutional types (Betsey 2017). In fact, the National Study of Postsecondary Faculty has found that HBCUs have a substantially higher proportion of foreign born faculty and of naturalized citizens (Betsey 2017). There is extensive research on the recruitment and retention of foreign born faculty in the United States (Mamiseishvili 2011). Across all institutional types, foreign born faculty display stronger preference for research than US born faculty, and job satisfaction of foreign born faculty vary with regards to national origin and institutional type, but is generally lower with regards to research rewards. But very little attention has been paid to HBCUs, especially as to what might explain their ability to recruit and retain higher numbers of foreign born faculty. As such, this study is driven by two main research questions: What are the key factors that shape the recruitment of foreign born faculty at HBCUs? Do those differences lead to differences in current job satisfaction?

Methods:

The data for this project comes from the NETWISE II survey, a National Science Foundation funded project. The population for the survey includes tenured and tenure-track academic scientists in four general areas: biology, biochemistry, engineering, and mathematics. The population includes faculty at the ranks of assistant, associate and full professor from Research Extensive and Research Intensive institutions, Masters I/II institutions, and Liberal Arts colleges, and includes a number of Historically Black Colleges and Universities. Using data from the 4195 survey responses we are able to compare data  regarding job satisfaction, resources available to faculty, academic productivity, and the relative importance of different factors in their initial job search.

Specifically, we estimated a Multiple Correspondence Analysis MCA to determine the relationship between several ordinal and nominal measures of career preferences, followed by a number of t-tests and multivariate regressions on measures of initial career preferences, current satisfaction, and current collaboration patterns.

Findings:

Our survey sample reflects existing research that indicates that HBCUs have significantly more foreign born faculty than other institutional types. A little over half of all HBCU faculty who answered the survey reported being foreign born.

Regarding recruitment factors, foreign born faculty at HBCUs on average place greater importance on salary, benefits, institutional prestige and research emphasis than US born faculty at those institutions, even controlling for broad Carnegie classification types. Meanwhile, US born HBCU faculty tend to place comparatively more importance on geographic location, teaching emphasis, student and faculty diversity and institutional mission. Perhaps more interestingly, these differences are driven primarily by national origin, and not race or ethnicity. A series of logit models estimating the likelihood of a respondent to answer that certain factors were "very important" in their job search show that US born status is statistically significant in explaining these differences even when controlling for race and ethnicity.

MCA suggests that the profile with regards to career aspirations of foreign born HBCU faculty is closer to that of faculty at research oriented institutions, while US born HBCU faculty more closely resemble faculty at liberal arts institutions. 60% of HBCU foreign born faculty had research oriented tenure track positions as top career choice, 63% of US born faculty had teaching oriented tenure track positions as top choice.

While there are significant differences in terms of career preferences for US born and Foreign born faculty, these differences, surprisingly, do not lead to substantial differences in terms of job satisfaction. In terms of work satisfaction, foreign born faculty at HBCU are less satisfied with salary and sense of inclusion, but about the same regarding institutional reputation, reward system and rewards for research productivity, even as foreign born faculty at HBCUs self-report much higher levels of article productivity and grand funding. The finding regarding rewards for research productivity, in particular, seems to be in contrast with existing research (Mamiseishvili 2011).

One possibility that would explain this is related to the social networks of foreign born faculty, as HBCU foreign born faculty also report a substantially higher number of foreign based collaborators.

Conclusion:

Despite having a much more research oriented career preference, foreign born faculty at HBCUs report similar levels of satisfaction regarding the research emphasis of their institutions as US born faculty with teaching interests. One potential explanation might be the support obtained through social networks. More research is needed to determine the particular reasons for differences in preferences not translating to differences in satisfaction.

References.

Betsey, Charles L. 2017. Historically Black Colleges and Universities. Routledge.

Mamiseishvili, Ketevan. 2011. “Teaching Workload and Satisfaction of Foreign-Born and US-Born Faculty at Four-Year Postsecondary Institutions in the United States.” Journal of Diversity in Higher Education 4(3):163.

Perna, Laura, Valerie Lundy-Wagner, Noah D. Drezner, Marybeth Gasman, Susan Yoon, Enakshi Bose, and Shannon Gary. 2009. “The Contribution of HBCUs to the Preparation of African American Women for STEM Careers: A Case Study.” Research in Higher Education 50(1):1–23.

ABSTRACT. Reputation plays a key role in determining the allocation/division of reward among scientists pursing a competitive research career. Reputation, alongside individual ability, creates a strong path dependency in the trajectory of individual careers. We present a theoretical model in which the rewards of scientific production captured by an individual are proportional to the magnitude of that individual’s efforts. This deviates from the classical approach in which science is a pure public good, and in doing so may help account for the heterogeneity in scientific production we observe among peers arising from their career choices. Specifically we model science as a common-pool resource game, intrinsically connecting the appropriability of scientific output to a scientist’s optimal strategy. This simple model focused on the allocation of reward sheds new light on a variety of behaviours that have been observed amongst researchers, in particular ones often attributed to social network and community effects, but that have rarely or only approximately measured and accounted for.

Introduction: The rewards scientists receive for producing science arise both from creating a particular new piece of knowledge and from the public recognition that they were the person to do so (Stephan, 2012). Thus, in order to increase reputation, it is crucial that one must first win a discovery race, as priority is the basis by which one can legitimately claiming one’s contribution (Dasgupta and David, 1994). This fact is well established within the field of sociology of science where disputes over priority, and the incentives arising thereof, are understood to play a central role in the organisation of science (Merton and Storer, 1973). Canonical macroeconomic models, on the other hand, view science as a public good (Romer, 1990) — by definition non-rival and non- excludable. The pertinent result being that the economic payoffs of science are very hard to forecast, and more importantly, difficult to appropriate. Priority-based rewards, on the other hand, break with this by creating a privately-owned reward for scientific production — that the priority itself is the prize (Dasgupta and David, 1994). In this manuscript we model science as a local public good with appropriability, partially challeng- ing the assumption of non-rivalry in basic science. The beneficiaries of science production are the actors at the heart of that process, the contributing scientists themselves. Within this framework we integrate the invisible college as conceptualised Crane (1969).

Background: The literature on individual incentives and rewards for scientific production is vast. In recent years a steady stream of applied work have studied a variety of reward schemes paired with reputation and status. These papers often study the tradeoffs researchers face: explore vs exploit (Azoulay et al., 2011), specialise vs generalist (Teodoridis et al., 2018), and even with whom, and to what extent, to collaborate (Bikard et al., 2015). This sets a competitive environment which has some benefits, such as the efficient allocation of scientific effort amongst problems (Hagstrom, 1974) and some downsides when researchers incur into fraud in their research (Jin et al., 2013). Kealey and Ricketts (2014) follow the line of reasoning behind Dasgupta and David’s (1994) new economics of Science. Specifically they use a game-theoretic approach to model science as a contribution good, in which only contributors extract benefit from research. They call basic science a common resource, in a model that is, inherently, a local public good (Callon and Bowker, 1994). Our model extends this by incorporating private rewards to scientific production.

Main building blocks: This model is based on a Common-Pool Resource game. The stock of knowledge H is non-excludable to all players of the game. That means, they can easily access and build upon that stock of knowledge as part of the social group. This stock of knowledge generates, in turn, a set of rewards or outputs to society y. The aggregated yield of this knowledge stock H is thus y(H). In a standard public good/social dilemma game the rewards would be evenly split amongst all players such that r = y(H)/H. But in a common-pool game players capture individual payoffs as given by an individual distributional factor φ, thus resulting in a payoff r = φ·y(H)/H. This seemingly small modification actually introduces rivalry, inducing significant behavioural change (Apesteguia and Maier-Rigaud, 2006). In synthetic form, researchers will maximise their rewards minus costs: π=φ(y(H)/H)−C(φ)

Related results and discussion: Looking at different configurations of that basic equation allow us to make a number of claims regarding the effect of competition on science production. For example, by introducing a distribution of strategies that depends on the number of players, N, we extract results on the incentives to publish extensively (exploiting) in fields with large N. Similarly, we find results that imply that new fields or sub-fields are driven by researchers with poten- tially higher impact, in line with empirical results of Wang et al. (2016) Introducing researcher (natural) ability into the picture, we are able to detect the competition effects of the invisible college. Specifically our model predicts incumbent researchers of high ability will benefit from elitism within a given college, while new entrants of even higher ability are detrimental to the former. Challenges to status-quo are thus not welcome as also shown by Azoulay et al. (2015). We are currently working towards an empirical setting suitable for the full validation of these, as well as other, claims of the model.

ABSTRACT. 1. Introduction

Scientific publications co-authored by both universities and business companies in the private sector constitute the basis for one of the most popular and widely used proxy measures of public-private interactions. Many studies rely on university-industry co-publications to study different aspects related to interactions between universities and industry, such as the effect of these interactions on the scientific impact of publications (Lebeau et al, 2008), to examine the status of university-industry interactions in specific countries (e.g. Abramo et al, 2009; Olmeda-Gomez et al., 2015), or specific business sectors (e.g. Giunta et al., 2015). In the past years, the use of scientific publications to identify, trace, monitor and analyse university-industry interactions has been therefore limited to the identification of universities and business companies among the author affiliations contributing the publication. However, bibliographic databases are currently recording information on the funding acknowledgements made by the authors in their scientific publications, opening up new possibilities to identify university-industry interactions from a different angle on the basis of these publications, with the potential of offering new insights on such interactions. Often authors include in their scientific publications funding acknowledgements in order to indicate that the research being published was totally or partially supported by funds provided by specific organisations. This piece of information allows for the possibility of identifying business companies among the funding bodies. Since income from industry is also one of the indicators most frequently used in the literature on university-industry collaborations to proxy interactions between universities and companies, it seems reasonable and justified the used of acknowledgements to private funding in order to gain additional insights on public-private interactions.

In this study we consider both university-industry co-publications and university publications funded by industry in order to analyse and assess the possibilities and limitations of funding acknowledgements in order to gain additional insights on university-industry interactions through scientific publications. We aim at answering the following questions:

• What is the amount of both university-industry co-publications and publications funded by industry? • How do the number of publications in these two datasets evolve over time? • What is the overlap between university-industry co-publications and publications funded by industry? • Are university-industry co-publications and those including funding to private companies equally predominant across scientific disciplines? • Are university-industry co-publications and those including funding to private companies equally predominant according to the type of research (i.e. clinical/non-clinical)? 2. Data and method

Selected pharmaceutical companies

For this study we have selected 23 large pharmaceutical companies, which represents a small group of companies within the pharma sector but covers most of the companies with the highest levels of R&D investments in the sector worldwide. According to the EU Industrial R&D Investment Scoreboard produced by the JRC Economics of Industrial Research & Innovation, these 23 companies together invested more than 97 billion euros in R&D in 2016 and achieved sales for more than 600 billion in the same year. Looking at the countries where these selected companies are headquartered, 11 are located in the US, 9 in Europe, 2 in Japan and one company in Israel.

Source and data

We have collected from the CWTS in-house version of the Web of Science (WoS) all articles, letters and reviews published between 2009 and 2016 mentioning at least one of these 23 companies (or any of their subsidiaries or acquisitions) among the author-affiliations and in the funding acknowledgement section as registered in the WoS. However, we did not include in the analysis the recent acquisition of Monsanto by Bayer AG as well as Eli Lilly’s Elanco, as their main focus is other than pharmaceuticals In order to identify subsidiaries, mergers and acquisitions of these 23 companies, we relied on Moody’s Orbis which is the largest commercial database providing corporate information, supplemented with other written sources (mainly Wikipedia and Rafols et al (2014)). The universities considered in this study are those included in the edition 2018 of the Leiden Ranking. This edition covers 938 universities form 55 different countries. For some analyses (e.g. distinction between clinical and non-clinical research) we also rely on PubMed, by combining this database with the WoS. Type of research: clinical and non-clinical In order to distinguish between scientific publications related to clinical research and those dealing with non-clinical research, we rely on the classification of documents included in PubMed. A scientific publication has been classified as ‘clinical’ if it is classified in one or more of the following document types: Clinical Trial; Clinical Trial, Phase I; Clinical Trial, Phase II; Clinical Trial, Phase III; Clinical Trial, Phase IV; Controlled Clinical Trial; Pragmatic Clinical Trial or Randomized Controlled Trial. All scientific publications not classified as any of these document types as well as all the scientific publications retrieved from the Web of Science but not covered by PubMed, have been classified as containing non-clinical research. Funding acknowledgements processing Clarivate’s WoS started to include information on funding acknowledgements disclosed in scientific publications back in 2008. However, these acknowledgements as registered in the database, especially in biomedical areas, do also include conflicts of interest statements by the authors of the publications (e.g. Lewison and Sullivan, 2015). Companies are often mentioned in the context of these conflict of interest statements, as authors may indicate the participation in consultancy contracts carried out for specific pharmaceutical companies or membership of advisory boards in these companies, for example. This mix of funding acknowledgements and conflicts of interest in the same field of the database poses the challenge of distinguishing between the two. To this aim, we have been using text mining techniques to identify whether a specific company has been mentioned in the context of a conflict of interest statement or in the context of a ‘true’ funding acknowledgement. As it is possible to find both, conflicts of interest as well as true funding acknowledgements, in the same publication we have considered in the dataset of publications funded by pharma industry only those for which we found evidence of referring to a true funding acknowledgement, regardless the additional mentions to possible conflicts of interest in the same paper.

ABSTRACT. Background The knowledge exchange between public research organisations and industry has gained importance in the past years (Nonaka and Takeuchi 1995). The volume of necessary knowledge as such was growing, the complexity of many technologies increased, and boundary-spanning technological developments more and more demand interdisciplinary knowledge (Mugabushaka et al. 2016; Shafique 2010; Small 2010; Wang et al. 2015). At the same time, the benefits from innovation and the marginal effects of R&D investments seem to decrease, resulting in an R&D productivity slowdown (see for example Gordon 2012, Gordon 2014, Bloom et al. 2017, Syversen 2016). Scientific publications by companies might be seen as an indication of an intensified knowledge or science orientation of companies. In addition, science-industry co-publications might be interpreted as reflecting knowledge flows from science to industry, but also as outcomes of intensive project collaborations (Tijssen et al 2011). The numbers of company publications as well as science-industry co-publications have increased in recent years. Companies, however, might not have an interest to publish their research findings, at first sight. So what are the reasons for this increase? This is the core question of the research project presented here. Next to signalling (Hicks 1995, Simeth and Cincera 2013; Pellens & Della Malva, 2016), access to academic capital (Shapira and Wang 2012), increase of collaboration efforts (Vallas and Kleinman 2008, Chesbrough 2003, Narin et al. 1997), also technology transfer policies (Bozeman 2000, Perkmann and Walsh 2007, Frietsch and Schubert 2012) might contribute to the explanation. In the context of IPR management, scientific publications can be used for the documentation of the state of the art (Blind et al. 2013) and thereby prevent competitors from patenting crucial technologies (Della Malva and Hussinger, 2012). Collaborations with universities or research organizations may allow companies to outsource their research and (partly also) development processes. The externally produced knowledge can thereby be integrated in the internal R&D processes. Against this background companies strive to enlarge their collaboration activities (e.g. Liebeskind et al. 1996; Powell et al. 1996; Vallas and Kleinman 2008), while they are able to size-down their R&D departments (Chesbrough 2003). One outcome of this structural change could be an increase in science-industry co-publications. Currently, the scientific literature reflects on the question if the potential advantages of scientific publications by companies has changed in the recent years. Especially large and established R&D-intensive companies are more and more withdrawing from scientific publishing (Arora et al. 2015). As they were responsible for the majority of publications in the past, the question arises if young, technology-oriented firms are able to compensate for this (Drake 2014).

Methods The analyses are based on Scopus raw data of the years 2005-2017. Company publications are identified using a bag of words of companies' legal extensions. A manually coded benchmark group (gold standard) showed high recall at a reasonable precision rate. For the descriptive analyses the Scopus field codes (2-digit-level) and affiliation country information are used. As a source for patent data the EPO's Worldwide Statistical Patent Database (PATSTAT) is used and company information stems from a raw data file of BvD's Orbis Database. All three datasets will be matched using Levenshtein Distance as a matching procedure based on the affiliation, applicant and company names, respectively. An integrated dataset is foreseen for the majority of OECD and selected associated countries, while the method is tested and developed based on German company data. The matching procedure and the results will be compared with an integrated dataset based on the German Innovation Survey instead of Orbis data. By this comprehensive data matching approach most of the disadvantages and shortcomings of so far existing studies of company publication activities can be overcome, especially as these previous studies mostly focused on large and/or stock-market registered companies only (e.g. Simeth and Cincera 2016; Arora et al. 2015). In our dataset we will also include small and medium-sized companies so that comparisons between multinational, large and small companies will be possible. Interviews and a survey of company authors is employed to examine the motives to publish (or not publish), the publication strategies of companies and also the intellectual origins (e.g. internal versus collaborative projects etc.) of the published work. The survey results will also be used for the explorative derivation of further hypotheses for the quantitative analyses.

First Results The absolute numbers of publications by (German) companies have been increasing steadily over the past years. In the Scopus database, however, the share of company publications in total German publications is rather constant over the past years at a level of about 8%. The majority of these publications is in the medical research field, in biotechnology, and in chemistry, as well as in materials research and polymers. Large shares can also be found in engineering fields. In consequence, the majority of companies with substantial numbers of publications originate in the pharmaceutical and chemistry sectors. The pattern of the scientific fields is rather stable over time. The interviews turned out that in most cases no strategy or central management approach is driving and guiding these publications. It is rather the intrinsic motivation and the interests of the particular researcher that lead to an engagement in scientific publishing. More often than not these researchers (or managers) are also active in teaching at universities and universities of applied sciences, at least temporarily. In smaller companies, especially in science and technology-driven sectors, it is often the CEO who is the author of scientific papers. In addition, the first results show that several of the publications emerge out of collaborative projects with public research organizations and universities - either privately or publicly funded. In addition, the descriptive data will be enlarged to cover additional countries. First checks with China (it was expected to be a challenge) show reasonable results. The shares of company publications is much lower in China and reaches a level of about 5%. While the shares of national co-publications of Chinese companies with Chinese research organizations and universities is at a similar level like in Germany (about 4% of total publications), the share of co-publications with non-Chinese companies is much lower (1%). In France it is about 10% of the companies that are engaged in scientific publishing (Simeth and Raffo, 2013).

Outlook In the near future it is planned to include further countries, especially USA, France, and Italy as well as smaller European countries like Austria, Switzerland, or the Netherlands. The further methodological work of the coming months will mainly address the integration of different data sources, namely Scopus (bibliometric data), PATSTAT (patent data) and Orbis (company data), aiming for an integrated micro dataset of German companies. If possible companies from additional countries will be included and will act as a benchmark or comparison group. Based on this (quasi)panel dataset panel regression analyses will be employed to examine a number of questions on the structure of company publications. For example: Is there a structural difference in the publication behaviour of small and large companies? Do publishing companies originate in high-tech or low-tech sectors? Are publishing companies also engaged in patenting? Do citation rates of small and large companies' publications differ significantly? Are publishing companies economically more successful than non-publishing companies (requires a control group, to be extracted as matched pairs/statistical twins)?

ABSTRACT. Background: Researchers are somewhat like prospectors looking for gold. Both spend their time doing their work without knowing whether they will make a discovery. Researchers take their best results, publish them in high impact journals and promote them. Prospectors have it much easier. They take their ore to the assay office. The assayer determines the amount of gold. The prospectors’ pay is based on a fair evaluation of what they have discovered.

From this perspective, the methods used for ‘identifying research breakthroughs’ are, in essence, a description of assaying methods. The purposes are to reward researchers fairly and to show society that they are getting their money’s worth. Analysis of features such as bursts in citations, text, patents, downloads, social media and other scholarly communication serve this purpose. Methods that characterize changes in the knowledge network serve this purpose. Overall, the problem that is being addressed by this stream of literature is ‘fair evaluation’.

We are focusing on a different practical issue in this study. We are far more interested in research planning (where to prospect next) instead of research evaluation (whether you found gold in what you have already dug up). Predicting where discoveries are likely to occur is of central concern to practitioners. It is central to our understanding of how science evolves. But this topic has received very little attention in the literature. We are aware of very little previous work, other than some of our own, that explicitly focuses on testing large scale predictions about where researchers will be digging. We use funding by topic as a proxy for discovery under the assumption that funders will allocate their resources to those topics where they feel that discoveries will occur. Theory: We seek to predict research breakthroughs from a particular theoretical perspective. We are shifting from information theory (where papers represent information about what researchers have achieved) to a sociological lens (where papers can be used to identify the social group and the norms of that small group). Specifically, we are building on Kuhn’s (1974) thesis that (a) researchers organize into relatively small research communities and (b) these communities have different levels of paradigmatic maturity.

Hypothesis: Using the prospecting analogy, we posit that paradigmatically mature research communities are located over a vein of gold. A lot of people have congregated around the area because of past discoveries. Sophisticated equipment is being used to pull the gold out (e.g., one uses large databases and sophisticated methods for analyzing the data). Consequently, funders are more likely to direct future funds to paradigmatically mature research communities.

Paradigmatically immature research communities are, by Kuhn’s definition, composed of groups of people who don’t agree on how to frame the research problem. They don’t agree on what tools to use. This may be a consequence of not having found much gold in that area in the past. One could argue that maturity doesn’t come from agreeing on methods; it comes from a history of highly valuable discoveries that are proximate to each other and suggestive of future discoveries that can be made. Funders are aware of the general belief that there isn’t much gold in the ground in pre-paradigmatic communities, and are less likely to direct future funds to them.

Procedure: The first step is to identify Kuhnian research communities from the literature. Two separate models were created. For the first, and following Kuhn’s original suggestion, direct citation analysis was used to identify ~90,000 research communities using the Scopus database. For the second, we used textual similarity between documents to identify ~40,000 research communities using the PubMed database. The sizes of the research communities in these two models are similar. The difference in the number of research communities is mostly due to coverage. PubMed only covers the literature related to biomedicine, which is roughly 40% of the Scopus database.

The second step is to develop a set of metrics that represent paradigmatic maturity. In addition to network size and network density, we have identified and coded individual articles in terms of their type of contributions. Specifically, we identified method papers and discovery papers using citing sentences from PubMed Central. This allows us to determine the relative intensities of methods and discoveries in each research community.

Kuhn also suggested that mature paradigms follow norms such as accuracy, prediction and replication. We correspondingly note that these words are similar to those associated with male first authors in the research literature, and thus posit that the use of male vs. female language can be used as a signal of paradigmatic maturity.

We then allocated NIH funding (from the Star Metrics and NIH RePORTER databases) to research communities using grant-to-article linkages. Funding amounts were used as the dependent variable in our analysis, and the features described above –intensity of methods, intensity of discovery, and gendered language – were used as the independent variables in a regression analysis to see how well they could predict the funding amounts. All variables were normalized using log transforms.

Results: Initial results suggest that a combination of our three features is extremely effective at predicting future funding levels in both (Scopus and PubMed) models. We are currently double-checking these results because they are much stronger than we expected. Results from an earlier study, using different variables to predict funding, had an R-squared of around 36%. We are seeing a significant increase in this statistic to around 45%). Possible sources of this increase include focusing specifically on NIH rather than all funding agencies in the Star Metrics database, the identification of locations of previous (highly cited) methods and discoveries (i.e., existing gold veins), and the surprisingly strong relationship between paradigmatic maturity and funding.

Implications: This study has many limitations. It assumes that government funding policy is an indicator of where society believes there will be future discoveries. We also emphasize that we are only dealing with one funding agency (NIH) which has been criticized as following a ‘conform and be funded’ policy. In addition, we have difficulties in believing that paradigms are ‘immature’- this is an unfortunate turn of phrase that Kuhn used in the 1970’s. We therefore don’t yet know if we should claim that these results are positive (in showing that funding is going to the best ideas) or negative (in showing that funding is reinforcing the biases of elite researchers). But regardless of the interpretation, a model that can accurately predict where a funding agency allocates its dollars (and by extension where future discoveries will occur) is a significant contribution to the practical problem of deciding where to focus future research efforts and to our basic understanding about the science of science.

ABSTRACT. Abstract We investigate whether demographic differences among U.S. Doctorate Recipients impact publication output. This is achieved by matching the Survey of Doctorate Recipients (SDR) with the publication database Web of Science (WoS). Our research shows the most impactful determinants on the probability to publish are related to field of doctorate, employment sector and engagement in R&D activity. A doctorate recipient’s training is also significant, those who complete dissertations at high-research institutions are more likely to publish. After controlling for the dominate determinates, the demographic variables, race/ethnicity, gender and U.S. citizenship status at the time of graduation, still show impact on the probability of publishing. The ability to test a broad range of demographic variables is unique to the SDR-WoS dataset. Of the demographic variables, race/ethnicity has the strongest impact on likelihood to publish. Readers are cautioned that this summary represents a research-in-progress. As such the researchers are actively reviewing and revising the data it is expected that the numbers will change. Introduction Publication output in peer-reviewed science and engineering (S&E) journals and conference proceedings (publications) serves as an indicator of scientific research activity and knowledge generated by scientific research. In recent years researchers explored publication output itself to question what internal and external factors impact a researcher’s probability to publish, including author demographic characteristics such as gender, years since Ph.D., longevity, etc. Previous research has been limited by the available author’s characteristics using data from ORCHID, data collected with paper submittal (Scopus Author ID) or other data submitted by authors interested in explicitly linking their publications to a set of characteristics (Google Scholar, ResearchGate, etc.). These sources have gaps, for example an infrequent author may neglect registering or may choose to obscure demographic characteristics, such as not stating gender. In addition, the author databases typically include gender but not race/ethnicity or citizenship status. This paper overcomes author registration issues and develops a unique database by matching the Survey of Doctorate Recipients (SDR) to a publication output database. Data Survey of Doctoral Recipients National Center for Science and Engineering Statistics (NCSES) within the U.S. National Science Foundation conducts the longitudinal Survey of Doctorate Recipients (SDR) biennially since 1973, producing cross-sectional data on individuals who have earned a science, engineering or health doctorate degree from a U.S. academic institution and are less than 76 years of age (https://www.nsf.gov/statistics/srvydoctoratework/). The SDR provides data useful in assessing the supply and characteristics of the nation’s science, engineering and health doctorates employed in educational institutions, private industry, and professional organizations, as well as federal, state and local governments (NCSES InfoBrief, 2017, https://nsf.gov/statistics/2017/nsf17319/nsf17319.pdf). The SDR collects education history and demographics data, periodically questions on scientific collaboration and research outcomes are added to collect additional information on scientific productivity. Key SDR variables include demographics (e.g. age, sex, race, ethnicity, citizenship), employment status, field of degree and occupation. The SDR is weighted to reduce potential nonresponse bias by using the NCSES Survey of Earned Doctorates which is an annual census of all individuals receiving a research doctorate from an accredited U.S. institution. Beginning in 2001 the SDR was expanded to include those graduates from U.S. institutions who move abroad. The matching operation included the entire set of SDR respondents from the 1993 – 2013 surveys.

Web of Science database The Web of Science (WoS), a large bibliographic database, was used as the source of scientific articles. Publications dated January 1990 to December 2012 were identified for potential matches to SDR respondents through research undertaken by Thomson Reuters, (now Clarivate Analytics). SDR – WoS Data Match Comparisons across SDR waves is tricky because in the longitudinal design there is a high level of sample overlap between waves. Also the survey coverage has expanded over time. The present analysis focuses on the 2013 SDR cross-sectional sample, which represent the latest sample with the most comprehensive coverage. In total there were 596,811 respondent/publication matches created corresponding to 26,445 authors out of the 35,265 SDR respondents to the 2013 survey. SDR respondents may be associated with one or more publication and potentially more than one SDR respondent maybe co-authors on a paper. For comparison purposes the doctoral recipients are grouped into two cohorts: those who receive their doctoral degree between 1995 and 2000; and those who received theirs between 2001 – 2010. This allows an appropriate period to pass between when a PhD is received and the peak publication period – 6 years post-doctoral receipt for the present data set.

The SDR-WoS data corresponds to other research showing a higher probability of publication in the fields of biology and physical sciences– specifically higher publication rates among the physical sciences and life sciences (including health and biology and agricultural sciences) and lower rates among social sciences and psychology. Focusing on the 4 largest groupings of degree fields: biology & agricultural sciences, engineering, physical sciences and social sciences and psychology shows changes in the race/ethnicity among PhD recipients when comparing 1995-2000 cohorts to those who obtained their PhD in 2001-2010. Across 4 largest fields of science race/ethnicity have increased between the two cohort groups but only in engineering has a minority (Asian) crossed over the 50% mark. The changes in part are due to the improved SDR coverage of U.S.-trained PhDs residing outside of the U.S. for the 2001-2010 cohort. Determinants of Publication SDR-WoS dataset permits a unique exploration into the likelihood of a U.S. doctorate recipient to publish in a peer reviewed publication or conference proceeding based upon not only training and employment, but also demographics of gender, race/ethnicity and citizenship status at the degree time. Training variables include field of study and whether the doctoral institution has a Carnegie classification as a very high research university. Employment variables include whether the primary work activities are in R&D (as determined by the respondent), institutional type of employment. These variables are explored across three different cohort groups depending on when they received their doctoral degree: 1995 – 2000; 2001-2005; and 2006 – 2010.

Analysis of the 2013 SDR-WoS dataset shows, after controlling for training and employment factors, significant differences in likelihood of publishing across cohort groupings by gender, race/ethnicity and to a lesser extent U.S. citizenship status. The analysis is based on the most typical 2013 SDR respondent who is a white male, U.S. citizen with a PhD in the field of engineering from a high research university who is currently working at a 4-year university, residing in the U.S. with primary work activities in R&D. Among the demographic variables, females, blacks and Hispanics have statistically significant lower odds of publishing. Interestingly the publication probability of an Asian PhD recipient is not statistically significantly different from white PhD recipients in the most recent cohort. Among the citizenship variables, some evidence of differences exists but the pattern is unclear. It is likely due to the association between race and U.S. citizenship.

Information is not available on submission rates by race/ethnicity, so we are unable to further explore potential bias among reviewers or publications. Although research of employment markets has shown name-based bias. Differences in the likelihood to publish potentially relates all the way back to racial and ethnic differences seen in primary, secondary and post-secondary educational pursuits.

ABSTRACT. 1. Background

Government Research and Innovation (R&I) programmes increasingly aim for deeper and wider societal impact. In Europe the idea that R&I policy should address ‘Grand Societal Challenges’ appeared in 2007, when the European Commission introduced the concept in its future vision for R&I policy. Two years later this was reinforced in the ‘Lund Declaration’, which stated that ‘European research must focus on the Grand Challenges of our time moving beyond current rigid thematic approaches.’ Challenges discussed included climate change, sustainable energy, health, cybersecurity and so on. Since then, a number of initiatives have been launched that explicitly focus R&I policy on societal challenges. These make new demands of the way we design, govern and evaluate government intervention in R&I. Of course, topics such as health and the environment are not new in R&I policy. In the new approach, they are tackled at a more systemic level, often aiming at transitions in socio-technical systems and regimes rather than focusing on piecemeal solutions. This requires more holistic, inclusive and multi-level governance than in the past as R&I policy starts to reach up from the micro to the meso level of the system of innovation. Such systemic goals are increasingly ‘mainstreamed’ in overall policy in several - mostly European - countries. Sweden, for instance, has embraced the UN’s Sustainable Development Goals in its national 2030 Agenda and Germany’s renewed High-Tech Strategy 2025 focuses one of its three Pillars on societal challenges. The UK Industrial Strategy Challenge Fund brings together research and business to tackle the big societal and industrial challenges. This focus on societal challenges coincides with increased interest in the growing literature on socio-technical transitions. We explore policy lessons from that literature but also its shortcomings for practice-orientated policy analysis. The research on which this paper builds used mixed methods including literature and document review, interviews, case studies and stakeholder workshops.

2. Complex Innovation and Transition Programmes (CITPs)

Implementing societal challenge focused R&I policy requires a new type of programmes. These are often large-scale, complex and multi-disciplinary, involving a wider set of stakeholders than the usual R&D performers, including regulators, regional authorities or representatives of civil society. They need long-term (financial) commitment from research funders and stakeholders as their complex goals are longer term and evolve over time. They often span policy domains beyond the R&I ministries and need multi-level governance, including regional and local actors. Such holistic interventions been extensively discussed in the transition management literature in the last couple of decades. However, despite the recent increased interest in politics and agency, most of that literature is remarkably silent about how to operationalise governance and to connect it to government. One way to implement societal challenge-orientated R&I programmes is via ‘Big Missions’ or ‘Moon Shots’. In the USA, ‘Big Missions’ usually involve large science (infrastructure) investments. However, the European big missions have taken inspiration from the US ‘Moon Shots’ such as NIH’s Beau Biden Cancer Moon Shot launched in 2016, which aimed to deliver a decade’s worth of cancer research in five years. The German 2025 High Tech Strategy has missions such as ‘Reducing plastics in the Environment’ and ‘Greenhouse gas neutral industry’ without quantifiable targets. An unofficial government ‘mission’ in earlier generations of the High-Tech Strategy was to ban fossil-fuelled cars by 2030. This sent a strong signal to stakeholders but lacked a dedicated implementation platform. The European Commission is including Missions such as such as Combatting Cancer, Adaptation to Climate Change and Healthy Oceans and Natural Waters in the next multi-annual R&I framework programme, Horizon Europe. In these examples, policymakers are still experimenting with the designing, implementing and governance of CITPs.

3. Policy implementation issues

We focus on two implementation challenges for CITPs: proposal assessment; and evaluation. CITPs need to select projects not only for quality but also for contribution to their complex goals. We ask whether traditional peer review is still adequate for this new type of programme. How should proposal assessment address the question of relevance to the societal challenge? Should a wider set of stakeholders be involved in project selection? Are practices used by US agencies applicable in Europe? This part of the paper is based on a world-wide review of proposal selection processes and criteria used by R&I funders in societal challenge and ‘objective-orientated’ programmes. One finding is that European countries experimenting with CITPs, are also adapting the selection of proposals for R&I funding to those new needs. The appropriateness of the project selection process very much depends on the preparation of the ex-post evaluation framework constructed during the design of the programme and is highly context dependent. The second implementation challenge we discuss is how to evaluate CITPs. There are well developed evaluation methods for tackling simple interventions that have a linear logic, but it is clear that we need a more sophisticated approach with CITPs. These are ‘complex’ in the sense that they deal with systems of people who learn and change their behaviour during the course of the programme. Many CITPs ‘experiment’ with different ways to do implementation. So, evaluation methods must understand causality in a changing reality with learning and sometimes changing goals. Some currently-used and generally quantitative evaluation techniques tell us little about the mechanisms of change. Some newer evaluation approaches focus on giving feedback to programme managers and stakeholders rather than to society. We need a ‘realist’ approach that develops theories of change that explain why programmes succeed or fail, uses robust, mixed methods and undertakes both the tasks of public policy evaluation: providing accountability to government and society; and feedback to those working with the programme to help them improve. The paper draws a recent study for the Swedish Agency for Growth Analysis on evaluating CITPs, including case studies in Germany, the Netherlands and the UK, and subsequent work in Sweden. One lesson is that not only governance is but also evaluation in CITPs should be multi-layered. New institutions may be needed such as ‘platforms’ to design and govern the programme, involving a wide range of agencies (or other agents such as cities) and other stakeholders are involved but which has staff and a leadership team of its own, to lead the programme. The platform would coordinate interventions managed by existing agencies (or combinations thereof) resulting in a hierarchical intervention and reporting structure.

4. Relevance of these issues and further work needed Several countries have started to implement these new types of interventions. Most are still experimenting and adapting their policy designs and governance. One result is the emergence of a spectrum of programme scale, ambition and complexity, including large-scale programmes but also smaller-scale interventions using new approaches to sub-problems within societal challenges. Examining project selection and ex-post evaluation opens up broader issues in R&I policy design, implementation and governance. Much more empirical evidence is needed on how the relatively conceptual literatures on transition management and mission-orientated policies can be operationalised in practice across different cultural and thematic contexts. This paper aims to provide some challenging thoughts based on policy practice and experimentation in this field.

ABSTRACT. Introduction

As the impact agenda grows to include more formal criteria to assess the broader value of research beyond academia, so too is the need to re-examine the implicit optimism that is embedded within expectations of the science-society relationship. Within this relationship, evaluations are expected to be conducted with the public’s best interests in mind, however this assumed orientation within the impact agenda towards positivity may act to blind reviews in the cases of grimpact (negative impact). Grimpact, although being a widely discussed issue in research evaluation in various forms, it is still poorly conceptualised, theorised and organised. This therefore limits the effectiveness of interventions on macro, meso and micro levels to combat its effects. This paper presents a discussion of the conceptualisation of grimpact and, using four vertical case studies as examples (Cambridge Analytica; MMR Vaccine; Economic crisis; and Karolinska scandal) further highlights the mechanisms underpinning grimpact. Using insights from RRI, extraordinary versus normal impact, technology assessment, normalised deviance alongside an exploration of the influence of post-truth in research’s role in society, this paper highlights a broken science system that is blind to the risk of grimpact. As such, grimpact thrives and feeding from existing evaluative frameworks is increased. Through an examination of grimpact, the paper also moves towards proposing a framework operationalised on the micro, meso and macro levels of research capable of recognising grimpact.

Background Recent political upheavals have the potential for wide-ranging effects on the public perception of the value of public services, including the societal impact of research and higher education. There has been a surprising public swing towards new, populist political movements that profess to represent “the real”, otherwise unrepresented and forgotten people, and this swing also has been felt in several nations that have hitherto resisted populism. There has been the rise of a new kind of citizen, what Reedy et al. (2014) called the ‘misinformed voter’ whose belief sets and voting behaviour may be impervious to rational arguments. This raises the question of how publics which are prone to confirmation bias, distrustful of public experts, and highly path-impregnated in their belief sets (Gastel et al., 2017) are prepared to value research that does not fit with their ideological conceptions. The rise of the “impact” agenda and its equivalents has in part provided a forum where the public value of research is discussed, weighed and promoted. It has been included as a formal criterion in many funding systems and mechanisms across Europe, UK and North America – all countries where the effect of these political changes are acutely felt. This has included the widening of research funding criteria to include conceptions of research excellence beyond academia, as well as more concrete actions such as including public members in extended peer review panels. Its inclusion reflects the “abstract faith” that public assign trust in science (Luhmann, 1979), and the potential it brings to improving their lives. In addition, the implicit positive message in most impact definitions globally, which includes specifications that impact can only have “benefits” and “value” beyond academia, and that its effects are part of an organised process from bench to bedside, is blind to the realities of research translation and unforeseen adoption. This betrays the science system’s devotion to a type of impact that is in line with promoting knowledge production that is done solely for the public good. Any claim for the wider public value of research depends on making claims on behalf of ‘the public’ and what might be regarded to create value for them. In this paper, a distinction is made between what is created beyond the academy and the value that society gives to that creation. Whilst capacities are neutral in that they exist or not, the value that particular publics attached to those capacities can be positive or negative depending on their ideological inclination or indeed the public mood of the day. In the long-term perspective, publics have been conditioned to regard valuable research as research that creates a positive economic impact, and in part this is because a value can be attached to the impact, that is of the economic value. The widespread acceptance of price serves to mystify the question of whether this is really valued by ‘publics’ or that value is an artefact of financial engineering (for example the market valuations of some university ICT spin-off companies created during the dot.com bubble). It is this issue and its partial solution through the use of market prices which creates the short term problem this paper envisions, which relates to the challenge of evaluating (meaning attaching to a value rather than applying an evaluation mechanism) the public value of research. In the absence of mystification, there are no generally believable claims for the public value of science to use as baseline indicators when particular political projects make populist claims about the positive or negative impact of specific branches of research. What is seen is that many claims are made for non-economic outcomes that are so extraordinary that they are indisputably good (what Sivertsen, 2018, calls extraordinary impact), and that they are unambiguously beneficial in not involving conflicting versions of societal process. This paper argues that what is missing is a deeper conceptual exploration of this politically contested version of impact in terms of its definitions, characteristics and precursors, and without that is it not possible to get beyond the domination of economic and non-controversial versions of impact. It contends that a useful starting point is to look at extreme examples of impact and public valuation of that impact, namely where there is a strongly negative impact, what we refer to in this paper as “Grimpact”. It presents three powerful cases (Siggelhow, 2007) of Grimpact to better trace out the core tensions, drivers and lines of force within this wider notion of public value. It uses notions from RRI, extraordinary impact, normalised deviance, post-truth and technology assessment to create and analyse grimpact typologies for the recognition of the concept more theoretically, as well as its identification during the practice of evaluation and valuation.

ABSTRACT. Rationale Transformative Innovation Policies (TIPs) postulate that addressing the key challenges currently facing our societies requires profound changes in current socio-technical systems (Weber and Rohracher, 2012; Schot and Steinmueller, 2018). To trigger such “socio-technical transitions” calls for a different, broad mix of STI policies (Kivimaa and Kern, 2016), with particular attention being paid to policy experiments conducted in protected technology niches. Therefore, TIPs can operate at different levels, but are often based on small scale experiments developed in protected niches (Schot et al., 2016; Schot and Steinmueller, 2018). Such experiments aim at both regime destabilisation (Kivimaa and Kern, 2016; Ghosh and Schot, 2019) and the process of accelerating and embedding niche innovations, including upscaling, replication, circulation and institutionalisation of the technology niches associated with the experiment (Turnheim et al., 2019). These policy approaches pose a substantial evaluation challenge: how can we evaluate small specific interventions or sets of interventions, with a narrow geographical and temporal scope, when the final objective are ambitiously systemic? How can we know whether a specific experiment has set ourselves up on the way to systemic transformation?

Background The literature on sustainability and socio-technical transitions has, so far, paid little attention to evaluation. Following the increasing interest in system evaluation and policy mixes (Magro and Wilson, 2013; Borrás and Laatsit, 2019), almost all the evaluation efforts have aimed to develop and test frameworks working at the macro-, or meso-level of socio-technical change. For instance, Janssen (2019) tried to assess the ‘value for money’ of policy mixes for transformative change using an approach based on Technological Innovation System. Yet, these contributions offer little help to evaluate the contribution of local interventions to systemic transformation. Turnheim et al. (2015) addresses this gap proposing “an integration strategy based on alignment, bridging and iteration” of learning-based evaluation of local initiatives with socio-technical analysis at ‘regime’ level, and quantitative system modelling at ‘landscape’. However, their proposal is very complex and articulated around the idea of aligning interventions that are inherently different as they operate at different policy levels.

The problem we face can be seen as a specific instance of the common challenge posed by the impact assessment of policies that occur a long way upstream from their intended final objectives, as for instance societal-challenge driven research policies, or local interventions aiming at socio-economic development. There are, however, specific characteristics of TIPs that influence the design of evaluation methods. TIPs proponents propose a set of policy characteristics that form part of the specific transformation logic they seek. For instance, policy interventions have to pursue changes in the structure and culture of governance (emphasizing inclusive participatory processes) and have to aim to generate “second-order learning”. We argue that these characteristics have to be extended to policy evaluation methods and practices, and imply crucial modifications in the roles of the actors involved in evaluation activities, as well as changes in the organizational routines within which these evaluation practices are inserted. This study will propose an evaluation strategy that addresses these challenges in a way that is consistent with the principles of TIPs as developed by TIPC: directionality, societal goal, systemic impact, deep learning and reflexivity, participation and inclusiveness (Chataway et al., 2017).

Approach The paper will develop an evaluation method building upon methods and techniques developed for the evaluation of policy interventions closely related to TIPs. We will identify and review a set of relevant evaluation approaches following a heuristic strategy based on a combination of search by keywords, and a “snowball” strategy following the references found in the initial research “corpus”. Evaluation work in these different policy fields has generated different evaluation communities and practices that have not been connected so far, including those developed for the evaluation of sustainable innovation, sustainable transition, responsible research to leverage sustainable transformations, innovation policy mixes for system transformation (Aranguren et al., 2017; Kivimaa et al., 2017; Janssen, 2019), EU climate policy (Hildén et al., 2014), interdisciplinarity of socio-ecological research (Holzer et al., 2018), or sustainability transition experiments (Taanman, 2014; Luederitz et al., 2017; Heiskanen and Matschoss, 2018). We will assess the suitability of the approaches and techniques developed in these fields for the evaluation of TIPs, and suggest a generic evaluation approach that is linked to extant evaluation practices. Initial versions of the approach were presented and discussed in TIPC workshops involving officers from science and innovation agencies in six different countries (South Africa, Colombia, Mexico, Norway, Finland and Sweden).

Anticipated results From our analysis we will conclude that TIPs evaluation should be “formative”; that is, it will aim to improve the definition and implementation of the interventions under evaluation, and to do so with the involvement of policy participants. This requires evaluation to be conducted in real-time, as a form of constructive monitoring. To be able to assess in real-time the degree to which the interventions are progressing towards the achievement of long-term systemic goals, the evaluation approach needs to be underpinned by both generic and specific “Theories of Change” (ToCs).

Following programme theory conventions, our generic ToC is formed by five elements: context, structure, processes, outcomes and impacts. Yet, we define each element to align it with transitions theory:

- Context: elements of landscape and socio-technical regimes - Structure: resources available to actors to enact change - Processes: the experiments and their activities - Outcomes: changes in people and organisations, including changes in networks, capabilities and learning, and expectations and visions - Impacts: long term effects produced by outcomes (related with big societal challenges, like those addressed by the Sustainable Development Goals)

The paper will present a generic ToC and guidelines on how to generate specific ToC based on the generic proposal. To enable second-order learning ToCs need to be flexible and should be revisited as part of the formative, real-time evaluation processes. The approach proposed includes a three-step process to build and revise the specific ToCs:

1. Identify the level of the TIP experiment (project, programme or policy mix) 2. Identify key evaluative dimensions (depending on the level) 3. Discuss the ToC and check its consistency with TIP criteria. Reformulate the ToC if necessary.

Significance We develop a strategy for evaluating TIPs based on a real-time, formative approach supported by flexible Theories of Change. Although ToCs are common in policy evaluation in other domains (for instance in development), they were seldom used in the reviewed literature. Our interaction with policymakers suggested the importance to anchor evaluation on a generic ToC that would help build a common rationale and theory-base justification for TIPs: a stylized view of the transformative change processes derived from transitions theory (Schot et al., 2016; Schot and Kanger, 2018). The resulting approach is innovative and provides an answer to the problem of assessing the downstream contributions and impact of current policy interventions, in a way that is coherent with TIPs principles.

ABSTRACT. BACKGROUND. Even a cursory glance at the evolution of the higher education sector in the United States in recent decades shows that university leadership and academics have been well aware of the need to adopt the “third mission” of service to society by directly engaging with the world outside college campuses. This recognition led not only to changes in universities’ mission statements, policies and public relation strategies, but also to the introduction of resources, support measures and dedicated funding for the social engagement activities.

Yet, the impact of these changes remains limited. Despite genuine efforts of individual faculty members, passion of students, commitment of leaders and pressure from the public and policymakers, societal engagement remains seen as the fringe academic activity both from outside and within universities themselves. Disciplinary research and to a lesser extent teaching still dominate the culture of academic excellence in top research universities that arguably have the best minds, the best ideas and most resources among all institutions of higher education to address societal problems and achieve public good.

One of the major challenges to the institutionalization of societal engagement in research universities is ambiguity of the concept of “engagement” itself. There are numerous contextually and institutionally dependent definitions of engagement between universities and society given to a multitude of relevant overlapping but not-quite-the-same terms such as “social”, “public” or “community engagement”, “outreach”, or “engaged scholarship” which are often used interchangeably within the same university and even more so in a broader academic community. However, without clear and shared understanding of what constitutes engagement by university faculty and administration it is difficult to establish clear and useful ways of measuring and reporting it and align corresponding internal processes and policies on all organizational levels. Resulting lack of evaluation, reward and recognition of academics’ efforts in societal engagement is one of the main reasons why faculty often abstains from participating in it.

Instead of making another doomed attempt to create an all-encompassing, one-size-fits-all definition or conceptualization of engagement that would reconcile different views on it within or in between universities, in this paper we approach the problem by suggesting a framework for analyzing a degree of integration of societal engagement in a research university that integrates the multiplicity of existing contexts, languages, definitions and conceptualizations of university engagement with society. In constructing our framework we focus primarily on ontological relations between interactions of university with “external” society and other principal university activities such as research, teaching and service. We argue that these ontological relationships are embedded both in the language of existing definitions and conceptualizations of social engagement and in the organizational structures and processes that govern all university and faculty activities.

METHODS AND RESULTS. Methodologically, our framework is constructed by applying semantic analysis to the existing scholarly and “grey” literature on university engagement with society. At the first stage, we constructed a vocabulary of various relevant terms for the phenomenon, including, among others, “community engagement”, “public engagement”, “engaged scholarship” or “university-community partnership.” The Web of Science (WoS) search conducted in February 2018 using the vocabulary yielded about 2300 documents. Narrowing the search by focusing on papers discussing theories, definitions, frameworks and concepts of engagement reduced the output to 675 documents. These papers were further analyzed for the presence of distinct definitions and conceptualizations of various terms for social engagement. Where relevant, contextually important cited references not indexed by WoS, such as policy documents, reports or book chapters were also added to the corpus. Eventually we identified 17 distinct definitions and concepts related to university interactions with society.

At the next stage, we applied semantic analysis to these 17 definitions and identified their six key common semantic components:

• Motivations for engagement, e.g. societal responsibility, civic duty or economic motivation to make knowledge useful for the economy
• Goals of engagement, e.g. provision of certain benefits or solution of societal problems
• Degree of integration of engagement in university activities: engagement is an independent university mission, or part of other missions, intersection of missions, or feature of a particular type of activity
• Target audiences, e.g. internal or external, specified by sector or geographical scale
• Emphasis on mutuality and reciprocity of relationships, or lack of it
• In some cases, the scope of covered activities, e.g. community service or economic development

At the third and final step of the analysis, we grouped 17 initial definitions into seven distinct models of university social engagement based on the degree of integration of engagement into the ontology of university missions and functions.

Model I is a “baseline model” in which engagement is a minor individual type of activity, often part of a broader service mission of the university. Engagement is more prominent in model II where it becomes a synonym for the service mission independent of research and teaching. In model III, engagement is a form of teaching and research that involves external audiences and societal goals. In model IV, engagement forms a subset of three university missions (teaching, research, service) that indicates their connectedness to external audiences and societal goals. In model V, engagement is an integrated “third mission” functionally overlapping with other two missions of teaching and research. In model VI, engagement is an attribute of any activity across three overlapping missions of teaching, research and service that indicates involvement of external audiences and societal goals. Finally, model VII marks full and comprehensive embeddedness of the university in society. In this case engagement is every activity in the continuum of scholarship that connects the university to the society in pursuit of societal goals.

SIGNIFICANCE. A framework that integrates these seven models can be used as a communication tool for developing common understanding of university social engagement between faculty and administration on the basis of institutional history, context and aspirations. It can also be applied as a diagnostic tool for identifying cases of misalignment of university policies, processes and organizational structures with the chosen model of social engagement embedded in the university strategic documents or mission statements. For example, it is often the case that societal engagement is still integrated in the language and definitions used in tenure and promotion policies or faculty activity reports according to models I or II (i.e., as a minor or secondary faculty activity), even if the university officially declares its intention to become more socially embedded and responsible, which corresponds to models V-VII. In such case there is a clear gap between declared university or faculty goals and actual university and faculty policies and practices. For these goals to become attainable, the university must close this gap by aligning the policy language, structure and implementation practice with societal engagement models used in strategic documents and mission statements.

More broadly, the framework explains why none of the research universities in the United States are truly successful in becoming socially engaged at the level of model VII or even VI in their operation. In their perception of societal engagement, university is the principal actor that somehow interacts with “external” society often described as homogeneous “communities” that are targeted by the university as the place of direct application of university knowledge and expertise. There is a distinct lack of the societal voice and authority in the university goal-setting. Thus, empowering society in the university decision making may be the best strategy to accomplish its social engagement mission.

ABSTRACT. Abstract

Entrepreneurship has become an important part of the mission of research university (Link, Siegel, and Wright, 2015), including patenting, licensing, and startup creation. All research universities have establishing technology transfer offices TTOs (Bercovitz & Feldman, 2008), as well as numerous programs and initiatives to promote entrepreneurship and the commercialization of university research (Siegel and Wright, 2015), demonstrating that the norm of academic entrepreneurship has been fully diffused. University technology transfer activities are now considered legitimate and taken for granted (Colyvas & Powell, 2006). However, there are huge variations in actual university tech transfer activities. Against this backdrop, researchers have asked the following questions: Why are the widespread adoption of TTOs and encouragement of university tech transfer activities not producing the expected outcomes? Why is there variation in technology transfer outcomes among the universities? Research suggests that institutional or macro-level explanations obscure the internal diversity of the university work settings and agency of the actors (Tuunainen, 2005) during the complex organizational change which permits the co-existence of conflicting institutional logics of academe and market (Lam, 2010). Given that institutional/macro variables only provide partial insight into academic entrepreneurship (Siegel, Waldman, & Link, 2003), more attention should be given to organization and individual factors. This study adopts a “micro-level” perspective on academic entrepreneurship, focusing on psychological and organizational factors that may affect this activity. Given that a successful university tech transfer is ultimately up to the active involvement of individual academic scientists, micro perspective can provide a better insight into academic entrepreneurship (Balven, Fenters, Siegel, & Waldman, 2018). Balven et al., (2018) propose three types of micro factors within academic entrepreneurship: 1) self-contained micro-processes that incorporate cognitive or affective phenomena; 2) relational factor focused on interaction with other individuals (i.e., department chair, colleagues, etc.); and 3) interaction between individuals and organization level factors (i.e., university tech transfer policies, organizational culture). Among three types of micro processes, we take the second and the third perspective of micro-processes focusing on leadership roles in reducing barriers to academic entrepreneurship. Specifically, this study examines whether the championing leadership mitigates the negative impact of 1) lower organization level receptiveness to academic entrepreneurship; and 2) lack of information - scientist's weak understanding and knowledge of how to initiate technology transfer processes. We theorize and hypothesize the links between organizational constraints to academic entrepreneurship, the role of championing leadership and technology transfer intention of university scientists. We specifically pay attention to and draw upon championing leadership literature that depicts innovation champions as risk-absorbers and informational bridge, hypothesizing the following: H1) Higher receptiveness to academic entrepreneurship will be positively associated with a scientist's intention to engage in technology transfer in the future; H2) Informational barriers to academic entrepreneurship will be negatively associated with a scientist's intention to engage in technology transfer in the future; H3) championing leaders will moderate the negative association between lower receptiveness to academic entrepreneurship and academic scientists' intention to engage in technology transfer in the future; and H4) championing leaders will moderate the negative association between informational barrier and academic scientists' intention to engage in technology transfer in the future. We test our hypotheses using longitudinal data from 391 academic scientists and engineers at 25 major U.S. research universities. Our econometric results indicate that championing leadership can have a positive influence of the propensity of scientists to engage in academic entrepreneurship. We find no evidence of an association between lack of receptiveness to academic entrepreneurship and technology transfer intention of university scientists. However, we find that informational barrier (i.e., confusion regarding commercialization process, lack of knowledge whether and how TTO can help them engage in academic entrepreneurship) is a strong factor that undermines academic scientist's intention to engage in technology transfer activities in the future. We find no evidence of any direct role of championing leadership. However, we find that championing leadership mitigates the negative relationship between the informational barrier and future technology transfer intention. The finding, in general, suggests that academic entrepreneurship is well received, at least in our study sample, and may no longer be a huge barrier for potential academic entrepreneurs. However, informational barriers such as scientists' confusion regarding the technology transfer process and their lack of awareness of TTO's role, may still be a huge barrier to academic entrepreneurship. There could be many ways to help potential academic entrepreneurship, and this study suggests the role of championing leadership as an alternative to foster academic entrepreneurship. We seek to make contributions to literature, practice, and policy. First, we seek to add to the leadership literature by investigating championing in the unique context of academic entrepreneurship where there are possible conflicts of interests between the traditional role of faculties (teaching and producing public knowledge) and commercialization activities. Second, this study adds to academic entrepreneurship literature by investigating the micro-processes of university technology transfer which has received relatively less attention from researchers in the field of academic entrepreneurship. Third, this study gives insight to policymakers regarding how to facilitate academic entrepreneurship and help potential academic entrepreneurs.  

ABSTRACT. Background: Research universities and Federal Research Labs (FRL) are the cornerstone of American innovation. The country’s national competitiveness depends on these institutions to increasingly perform, translating research into the innovative products the country needs. However, technology commercialization is a nonlinear process and difficult to achieve efficiencies and address gaps. To address this, it is necessary to understand best practices for high-performing universities. When best practices are documented and understood, it results in more information shared, gaps filled, commercialization sped up, more companies formed, and research more rapidly benefiting society. In addition to helping solve the nation’s technical challenges, universities are relied upon by their regions for economic health and market diversification.

Rationale: This study investigates the best practices of 59 Innovation & Economic Prosperity (IEP) designated universities in technology commercialization. The IEP university designation was created in 2013 by the Commission on Innovation, Competitiveness & Economic Prosperity (CICEP) at the Association of Public and Land Grant Universities (APLU). IEP universities are uniquely positioned to excel in technology commercialization with their institutional emphasis on innovation and economic development activities. Among a sample of 110 public doctoral universities in the U.S. with detailed technology commercialization output data(1) available between 2012 and 2016, those with the IEP designation produced a significantly higher mean volume of new disclosures, new patents, startups initiated, and exclusive licenses and options. This demonstrates the unique qualities of this study group with its intentional focus on economic development and innovation.

Overall, whether they are located in a federal lab or a research university, researchers are driven by solving the country’s and world’s problems. IEP university successes in translating research can be built upon, expanded, and utilized by federal research laboratories and other universities interested in expanding their lab to market activities.

Methods: A mixed method analysis was adopted to determine best practices. Qualitative data informed thematic groupings of the best practices, while quantitative data survey informed the validity of the finding.

The study utilized original data collected from interviews and survey as well as secondary data. This study’s participants were chosen based on their ability to speak to the most effective practices in the U.S. for bringing new technologies to market. The study collected and analyzed primary, original data from 261 participants involved in a variety of cross-sector clusters and collaborations: 51 interviews with IEP university faculty researchers, ten interviews with affiliates of the federal research laboratories, and 200 surveys with IEP survey panel members, with an average of three respondents per IEP university.

The study team conducted semi-structured interviews lasting between thirty and sixty minutes. These interviews were professionally transcribed. Researchers read and qualitatively coded each transcript to identify common themes and characteristics of the lab-to-market process to identify best practices.

For secondary data analysis, complete and comparable data was available for 48 universities. Most often exclusions were based on lack of data for a specific campus. The IEP designation is campus-specific, and some universities only report system-wide data to AUTM. AUTM data from 2012-2016 was averaged over the five-year period for comparison to other public doctoral universities as well as to other IEP universities. In comparing IEP universities to each other, a notable concentration of output volume from the AUTM data was noted in the top 35 percent (n=17) of the universities. This group is labeled “High Producers Group” throughout the report.

Results: Through a mixed methods study, four areas of best practices emerged: culture, champions, incentives and collaboration. Universities with a strong cultural emphasis on lab-to-market promote its value both internally to the university, as well as externally to the surrounding community. Strong technology ecosystems are dependent upon champions - experienced professionals assisting in the maturation of a technology through expert guidance and mentorship. Incentives are vital to motivate and reward new ideas, while resources provide the necessary environment for continued growth. Finally, key collaborations are necessary throughout the process to foster ideas and to access resources throughout the ecosystem.

Significance: This study provides foundational research for how universities effectively move innovations from the lab to the market and benefit society. Universities face increased demands for innovations that can serve the public good through commercialization or other access. These best practices form a foundation that can guide, grow, and evolve as IEP universities experiment and implement lab-to-market ideas. It is expected that this study will encourage more faculty researchers, university staff, and investors to lend their perspectives and ideas.

Endnote (1) Measures of technology commercialization volume were determined based on data from the Association of University Technology Managers (AUTM) Statistics Access for Technology Transfer (STATT) Database, which compiles the results from a survey of university technology transfer offices.

ABSTRACT. In the knowledge society, universities play a role as the engine of knowledge because they are generators of knowledge and increasingly they are part of the market. The evolution of the work of the university has incorporated the entrepreneurial approach; it is common to find initiatives that stimulate the transfer of technology, the creation of new firms and the interaction with the environment. This situation has forced the Universities to establish mechanisms and strategies that allow them to use their resources and capacities for the development of their research and extension functions and obtain economic benefits. The creation of firms spin-off from the university level has reached great relevance among the instruments of technology transfer from the university to society, compared to others such as research contracts or patents. Some of the reasons that have awakened the interest is because they tend to be located near where technology arises and facilitate the growth of the local economy, promote changes in the university, incorporate of graduates to spin-off, perform a better valuation of results of research and can generate income that benefits the founders and the university. The creation of spin-off university companies empowers business networks by encouraging the use of advanced technologies, encouraging the establishment of cooperation networks between the company and other agents. This activity has a drag effect as a whole because, through different mechanisms such as imitation, it incorporates good practices associated with the management of technology and other areas of the company. In this context, the university has had to find more direct ways of bringing its academic knowledge closer to the market, which constitutes a radical change for the universities in their functions, insofar as they have been induced to play an active role in the scene economic As a consequence, in recent years, the knowledge transfer mechanisms used by universities have evolved. The use of programs for the creation of spin-off in the university sphere has been extended, the investigations in this regard have an important role since the main weaknesses of spin-off companies can be known in the moments after their foundation and that help to prop up growing as a company. Although the entrepreneurial university opens new opportunities for social progress through a rapid and effective commercial application of scientific knowledge, it also poses hidden costs, which is why it is interesting to advance in the knowledge of the direct involvement of the university in the entrepreneurial activity, not all university spin-off companies manage to leave the scope of protection of the university and have a successful autonomous performance, this limits the expectation of job creation and economic development deposited in the creation of this type of companies, and has led to the realization of various studies that try to deepen the understanding of the phenomenon of the creation of spin-off. In Mexico, the influence of the successes obtained in other regions created the conditions for the emergence of different initiatives, where two phases can be identified: in the first, in the nineties there were some initiatives of university-company linkage and the emergence of some business incubators, where the efforts were isolated and led by the universities. In the second phase, starting in 2001, public policies designed to promote technology transfer and innovation based on a systemic approach began to be designed. Systemic strategies were created to support innovative companies through different mechanisms, such as the network of Business Accelerators, entrepreneur programs, seed capital, the National System of Incubators, clusters, technology parks, and Technology Transfer Offices, etc. In this phase, the creation and promotion of innovative activity become a fundamental axis of the public policy strategy of Science, Technology, and Innovation, in particular, to mention an incipient seed and angel capital industry to promote the creation of Technology-Based Companies. , as well as the incorporation of entrepreneurship issues and financing of Technology-Based Companies in the Science, Technology and Innovation policy agenda in the country. The incorporation of innovation as part of public policy faced barriers associated with regulations by the various figures that legally have the actors of the innovation system. Until 2015 that the legal barriers are eliminated, but its implementation within the universities has been very diverse. The objective of the research is to explore what factors are determinants in the transfer of knowledge within the university spin-off and understand the dynamics present from the perspective of university managers. The methodology followed is a descriptive-analytical method, where the literature is reviewed and entrepreneurs are interviewed with spin-off companies and university managers who contributed to its creation. However, the ability to transfer technology and innovation is still weak and largely the result of weak institutional capacities of universities, although there are entrepreneurs and creative proposals, not considering the commercialization of technology as a substantive activity in the substantive task of the academy, the few entrepreneurs who emerge who face an unfavorable context to consolidate their project. Although some authors mention that, an important barrier to entrepreneurship from the academy is the existence of the conflict of interest as an inhibitor of technology transfer and entrepreneurship to generate technology-based companies. The certain thing is that it is had in perspective that the changes emanated of the reforms to the Law of Science and Technology in 2015, the universities, allow surpassing these obstacles. It is relevant to mention that the legislative change of 2015 seeks to catalyze the necessary institutional and operational adjustments for the promotion of technology transfer, by encouraging the establishment and management instances for technology transfer and linking in all the institutions that carry out scientific activities, and innovation, as well as eliminating the legal impediment that public sector researchers had to participate in activities related to the private sector, and equip them with the power to form strategic partnerships, technological alliances, consortia, linking and transfer units of knowledge, Technology-based Companies and regional innovation networks; as well as companies spin off and participate as partners in the different modes of technology transfer, among other figures In addition to the reforms of the Law on Science and Technology, it is established that universities and entities must issue and make public their regulations within a period not exceeding six months, counted from the publication of the aforementioned decree. During this investigation, it was observed that the approval of the guidelines, regulations or any figure that regulates the creation of spin off companies within the universities has not had great repercussion even in the period stipulated by the Law of Science and Technology for the emission of the particular regulations or later. Among the results, we can mention that the organization management and the institutional framework within the universities are key because it becomes the first ecosystem that facilitates or inhibits the growth of entrepreneurial projects (infrastructure, support in the early stages of business development), as it matures, it gives way to prominence. From the empirical evidence in Mexico, it stands out that the spin off that currently exist are based on personal initiative, because the universities from which they emerged do not have promotion programs for the creation of spin-offs.

ABSTRACT. University commercialization support initiatives have evolved since the Bayh-Dole Act (Wright and Siegel, 2015). Approaches after the Bayh-Dole Act emphasized technology transfer offices and tended to be more centralized, intellectual property-oriented, and revenue seeking (Breznitz, 2011). Studies of these traditional technology transfer support programs have not been found to be significantly associated with positive commercialization outcomes such as new venture capital, companies, or jobs (Grimaldi et al., 2011). Methodological factors are an issue in these studies. There are few quantitative studies that are able to find comparison groups that can account for the effects of confounding variables such as the quality of the service, characteristics of the university and location, or attributes of the scientist. Individual-level characteristics also are not well captured. Another issue with these studies is that the commercialization support landscape has evolved toward accelerators and entrepreneurship training programs that tend to be more decentralized, emphasizing entrepreneurship capacity development (Clarysse et al., 2015).

This research will address these gaps by comparing the outcome of individual projects that received support through the US I-Corps program. I-Corps is a program that originated in the National Science Foundation (NSF) in 2011 to provide training in evidence-based entrepreneurship methodologies to accelerate commercialization research of its principal investigators (Youtie and Shapira 2017). The I-Corps training is based on an initial three-day bootcamp attended by teams of NSF principal investigators, entrepreneurial leads (usually graduate students or postdocs), and experienced mentors. The bootcamp provides intensive training in the use of the business canvas model (Osterwalder and Pigneur 2010) and customer discovery and strategic pivoting techniques (Ries 2011, Blank, 2013). The teams are encouraged to leave the laboratory, investigate the resonance of their application with roughly 100 potential customers and partners, and make changes based on the responses received. I-Corps includes six weeks of follow-up to the bootcamp to check on the customer discovery efforts and a final read-out in which the investigator-centered team decides whether or not to pursue commercialization.

I Corps training is provided through a network of nodes geographically distributed throughout the US. Georgia Tech’s I-Corps South Node was established in 2012 through the university’s VentureLab unit as one of the first three sources for the evidence-based entrepreneurship curriculum. VentureLab is a Georgia Tech program established in 2001 to assist faculty members through the commercialization process (Youtie and Shapira 2008).

Now that I-Corps has been operating for several years, it is reasonable to ask about its impacts. The trouble with existing efforts evaluate I-Corps is that results are found in descriptive annual reports published by VentureWell, the data collection and content dissemination arm of the program (Venturewell, 2018). Not only do the outcomes reported lack a comparison group, but they also do not take into consideration differences in local commercialization ecosystems (Clarysse et al., 2011).

This paper compares two entrepreneurship support efforts to accelerate academic entrepreneurship of Georgia Tech faculty projects: I-Corps services delivered through VentureLab (VentureLab+I-Corps); and similar services through VentureLab but outside of I-Corps (VentureLab-only). There are 70 VentureLab+I-Corps projects and 200 VentureLab-only projects. The comparison assesses the likelihood of commercialization outcomes such as attraction of substantial financial capital, new company formation, or jobs. The independent variable of interest is whether or not the project involves VentureLab+ I-Corps or VentureLab-only. This independent variable in essence represents whether there is something particular about the approach that I-Corps uses over and above the basic evidence-based methodology which has been widely disseminated through entrepreneurship education practices oriented around the use of the business canvas model and lean customer discovery methodologies. A significant consideration is the ability to identify factors that encourage investigators to select into the VentureLab+I-Corps versus the VentureLab-only service. A selection equation first presents significant variables that distinguish the two service groups. A second stage analysis presents outcome variables—financial capital, new company formation, jobs—as a function of the main independent variable of interest, and control variables for year of service, discipline, and characteristics of the investigator.

Information on these variables is obtained from the program’s customer relationship management system, which is based on VentureLab reporting; Crunchbase (investments); secretary of state incorporations for Georgia, Florida, Delaware, and California; Small Business Innovation Research; Georgia Research Alliance commercialization awards and jobs survey data (a state research capacity-building program at select Georgia research-intensive universities); and US Patent and Trade Office. The expected direction of the association is unclear. On the one hand, I-Corps is a structured national program that situates the Georgia teams in a peer network and operates with a defined outcome over a specified period of time. These characteristics could lead to more benefits associated with the I-Corp+VentureLab service. On the other hand, the VentureLab-only service uses comparable personnel and support approaches and does not have the specified time frame of the I-Corps program, which could enable benefits that require longer periods to come to fruition.

References Blank, S. (2013). Why the Lean Start-Up Changes Everything. Harvard Business Review, 91(5), 64-+ Breznitz, S. M. (2011). Improving or impairing? Following technology transfer changes at the University of Cambridge. Regional Studies, 45(4), 463-478. Clarysse, B., Wright, M., & VanHove, J. (2015). A Look Inside Accelerators. London: Nesta. Calrysse, B., Tartari, V., & Salter, A. (2011). The impact of entrepreneurial capacity, experience and organizational support on academic entrepreneurship. Research Policy 40(8), 1084-1093. Grimaldi, R., Kenney, M., Siegel, D. S., & Wright, M. (2011). 30 years after Bayh–Dole: Reassessing academic entrepreneurship. Research Policy, 40(8), 1045-1057. Osterwalder, A., & Pigneur, Y. (2010). Business model generation: a handbook for visionaries, game changers, and challengers. John Wiley & Sons. Ries, E. (2011). The Lean Startup: How today's entrepreneurs use continuous innovation to create radically successful businesses. Random House Digital, Inc. Siegel, D. S., & Wright, M. (2015). Academic entrepreneurship: time for a rethink?. British Journal of Management, 26(4), 582-595. Youtie, J., & Shapira, P. (2017). Exploring public values implications of the I-Corps program. The Journal of Technology Transfer, 42(6), 1362-1376. Youtie, J., & Shapira, P. (2008). Building an innovation hub: A case study of the transformation of university roles in regional technological and economic development. Research Policy, 37(8), 1188-1204.

ABSTRACT. Abstract

This article is motivated by an apparent innovation productivity paradox in India wherein there has been a remarkable increase in the contribution of Total Factor Productivity Growth (TFPG) to output growth without a concomitant change in the factors that are commonly considered as instrumental in technological change. The recent empirical evidence from India tends to suggest that the higher output growth in India, even surpassing China (IMF 2017) , has been contributed significantly by Total Factor Productivity Growth (TFPG), an often used measure of technological progress and innovation (Krishna et al 2017). A comparative analysis of China and India (Wu et al, 2017) argued that though China’s value added growth was 50 per cent higher than India during 1981-2011, the TFP growth in China was nearly 25 per cent slower than India (0.83% and 1.13% per annum respectively). Surprisingly enough, the higher TFP growth in India has not been associated either with internal R&D that draws on firms accumulated knowledge or imitation of the innovation of other firms which are often considered as source of innovation by scholars (Lewin and Massini 2003; Massini etal 2003; Nelson. 1993; Basant and Fikkert 1996 among others). The present R&D intensity in India (0.8%) is much low as compared to developed countries or even China (2.1%). Similarly, the number of patent applications in China in 2016 stood at 1.3 million as compared to only 46904 in India for the year 2015-16. According to the Enterprise survey of the World Bank (2012), while 18% of the Chinese firms reported technology licensing from foreign companies, the reported percentage in India was only about half of China (9.4% in India). Hence the crucial issue is how to account for the remarkable performance in TFG that appears to be an apparent paradox in the observed technology-productivity relationship in India.

.The literature on technological capability in developing countries have conceptualized technological change as an outcome of three important sources; involving technology import from developed countries, own R&D effort, mostly adaptive, and technology spillovers arising mainly from FDI and trade (Katz, 1987; Bell, 1984 2006; Kim 1987, 1997; Dahlman et al 1987; Fransman and King 1984; Lall 1992; Kim and Nelson 2000, Rijesh, 2015; Parameswaran, 2009; Goldberg et al. 2010; Topalova and Khandelwal, 2011; Siddharthan and Lal, 2004; Marin and Sasidharan, 2010; among others). . However, the studies have also noted that the bearing these factors do vary across industries/firms and is also governed by the form in which R&D and technology import takes place along with the nature of FDI. Here it needs to be noted that focus has been on technology whereas there is reason to believe that firm level productivity cannot be attributed entirely to innovations in the sphere of technology alone. Thus viewed, the apparent paradox calls for a broader perspective on innovation in a context wherein countries across the developing world, including India, the focus of policy is being shifted from science to technology and to innovation (STIP 2013). We approach this this problem using the lens of learning economy, which is at the core of innovation system perspective. From the very beginning innovation system perspective delineated two modes of interactive learning. The first one, often referred to as STI (Science Technology and Innovation) mode of learning (Lundvall 2007; Jenson et al 2007; Lunvall 2017) emanates from science and R&D efforts that leads to codified and scientific knowledge which Asheim and Coenen (2005) refers to as analytic knowledge. Such R&D efforts may be undertaken through in-house R&D units established by the firms – both local and foreign - , public research laboratories, universities and through their collaborative efforts. The second mode of learning discussed in the literature refers to Doing Using and Interacting (DUI) mode. This is based on the premise that not all the important inputs into the process of learning and innovation emanate from science and R&D efforts . In the real world, much of the learning is experience-based that takes place in connection with routine activities in production, distribution and consumption and produces important inputs to the process of innovation (Lundvall 1992 p 9). Asheim and Coenen (2005) argued that such learning activities leads to synthetic knowledge in contrast to the scientific knowledge disused above.

Using innovation system perspective, this paper analyses the relative role of scientific learning (STI mode) and experience based learning (DUI) in determining Total Factor Productivity (TFP). In order to capture STI mode of learning, we use four indicators of interaction representing intra-country interactions such as own R&D, purchase of technology from domestic sources, training workers and inter-country interactions like purchase of technology licenses from players outside the country. Similarly DUI is captured through, participation in Global Value Chains (GVC), import of capital goods, Foreign Direct Investment (FDI) and Outward Foreign Direct Investment (OFDI). The empirical strategy follows a two stage process. First, making use of the firm-level panel data from the Indian manufacturing sector during 2001–2002 to 2016–2017, TFP is estimated using semi-parametric method of Levinsohn–Petrin that accounts for the endogeneity bias in productivity estimation. In the second stage, we regress the estimated TFP scores on host of STI and DUI indicators. Since the focus of this paper is to analyse the impact of firm’s learning capabilities on TFP, the major concern in the analysis is to address the problem of endogeneity. The unobserved firm characteristics may affect both TFP and some of our regressors like R&D, technology purchases, participation in GVC, leading to spurious correlation between the two. Endogeneity and biased results may also arise when unobservable time-invariant firm effects are correlated with regressors in the empirical model. In order address the issue of endogeinity, this paper employs the dynamic panel data model based on the system GMM method initiated by Arellano and Bover (1995) and fully developed by Blundell and Bond (1998). Our results indicate the intra-country interactions within STI mode as represented by R&D, technology purchases and staff training are emerged as important factors in determining productivity. Among the interactions within DUI mode, inter-country interactions (GVC, OFDI and FDI) play an important role in determining firm’s productivity. Using the innovation system perspective, this study offers few additional insights. We found that domestic technology purchases, staff training and GVC along with overall institutional architecture (measured in terms of trade orientation, labour market regulations and product market regulations) play a significant role in productivity.

ABSTRACT. Introduction

This paper aims to estimate the relationship between innovativeness and public-sector support for different innovation types at the firm level. Current research on the efficacy of direct public sector support lacks in detail, variation between supported activities is not recognized in quantitative studies. Mainly because micro-level data about public sector support as an input is not detailed in innovation surveys. Specific surveys about public support as input are either with very small samples within a single scheme or without innovation output data. This paper combines two different data sources to account for the diversity of subsidies and innovation outputs.

A grouping of public sector support based on activities is presented to distinguish between possible inputs for innovative activities.

Possible types of interactions can be between targeted actors and activities across policy, time or geographic spaces. These interactions can also lead to conflicts between policy rationales, goals or implementation approaches.

Innovativeness is measured as an output of technological innovations divided into eight categories, e.g. new products, processes, services etc.

Data is Community Innovation Survey (CIS) covering eight years matched with an external database of all public support given to Estonian beneficiaries. External data about public sector support has been collected from relevant agencies

The full dataset used in the analysis covers Estonian firms between 2006 and 2012 that are included in the CIS. Total of 3204 unique firms with a total of 7408 observations. This is an unbalanced dataset, where firms are observed over time with gaps.

Random effects logit model suggests that there is a significant correlation between types of subsidies beneficiaries receive and different innovation categories.

Literature review

National innovation system view emphasises relevant institutions who have influence over innovative activities within the economy. These institutions are wide ranging: firms, industry, competitors, education system, legal framework, financial system, R&D infrastructure, standards and norms. National innovation system framework has been successful as a tool to discuss interlinking aspects and influences to innovation with policy-makers.

Innovation policy has become wider in the last decades. The second and third wave of innovation policy includes more actors and their relations. This also includes more rationales for intervention, whether it is direct to the firm or aimed at other institutions in the national innovation system.

Innovation policy is wider than instruments developed directly to affect firms. These can be regulations, economic transfers, networking instruments, awareness campaigns, education, consumer protection, to name a few.

In this paper, I concentrate on instruments which are directly aimed at firms, meaning they can be classified as monetary aid. Even if instruments are in-kind aid, these are often calculated in monetary terms due to EU State Aid rules. Instruments which are indirectly aimed at improving the performance of firms via improving national innovation systems are out of scope.

Data & Method

Data about innovativeness comes Community Innovation Survey (CIS) carried out by Statistics Estonia. CIS is carried out bi-annually, and four waves are used in this analysis, covering years between 2004 and 2012. Every wave surveys a representative sample of firms about their innovative activities, inputs and outputs.

Altogether, 7408 observations from 3204 unique firms are included in the analysis.

Data about direct business support instruments comes from all relevant public agencies on the local and national level in Estonia between 2001 and 2016. Two main public agencies dealing with public support supplied their full registers. Also, EU Structural Funds register and Estonian State Aid register was culled for analysis.

The model is specified as a GLMM with random intercepts.

GLMM specifies fixed effects for controls, such as time and industry. Random intercepts are estimated for every firm.

Classification of direct business support

Based on empirical observations from Estonia, direct business support is categorised into classes which highlight the activities support with instruments. Direct business support belongs mainly under STI policies, but also can appear in environmental, regional, social, transport policies among others. Therefore, we observe two findings. Firstly, the possible scope of supported activities, rationale and policy domains involved. Secondly, the range of policy mix in which firms have to operate.

I distinguish twelve distinct groups of public support based on activities supported. These are collaboration programmes; consulting; training & skills development; marketing & export promotion; innovation and R&D support; investments support; mixed support; labour support; financial guarantees; and direct subsidies.

Results

The average use of simultaneous policy instruments has between 2004 and 2012, two or three simultaneous policy instruments are not rare. The possibility of interacting effects when analysing policy instruments is relatively high.

Estimates for technological innovations suggest that investment, innovation and R&D, marketing and export promotion, and training and skills development instruments are positively linked with technological innovation outputs. Consulting, financial guarantees and mixed support are not.

Within product innovations, and its subgroups, new products and new services, some aspects are highlighted. For example, investments, training and skill development, and marketing instruments are positively related to product innovations but not with new services. Suggesting that firms whose innovative activities are mainly related to the creation or improvement of services are not more likely to be beneficiaries in these instruments.

In most cases positive and significant relationships are logical, suggesting that firms who participate in certain instruments are also innovative in similar outputs. Some instruments, such as consulting, labour support and financial guarantees are not related to any type of innovation outputs at all. These instruments are either not supporting any innovative activities, such as labour support, or not-innovative firms self-select into these instruments.

Process innovations, and its subgroups, new processes, new distribution systems, new support systems, show similar results. Firms are not uniformly related to policy instrument types.

Innovation and R&D instruments are not related to any type of process innovations at all. This suggests that firms who participate in these types of instruments are more focused on developing novel products or services only.

Investment instruments are positively related to new processes and new support systems, but not with new distribution systems. Investments are mostly the acquisition of new machinery and equipment, which should yield process innovations almost always.

Conclusion

The instrument mix can have overlapping effects; interaction effects should be accounted for in empirical studies. Many firms participate in two or more different types of public sector support instruments simultaneously.

Most modelling of innovative activities relies only on a binary value for all categories of public sector support. The model specified here shows that these instruments mainly work as intended, or at least are correlated highly with activities they should support. The effects are mixed, and sometimes contradictory. When we analyse this variance, it makes sense that investments in machinery aid in developing new processes. However, R&D instruments are not related to new processes. This variation is not available is most datasets and can lead to underestimation, or worse, wrong conclusions about the efficacy of public sector subsidies.

Results suggest that public sector support given to firms should include more variation in empirical studies to account for different effects which may not appear in homogenized datasets.

ABSTRACT. 1.Background and Goals Technological innovation can promote the technology, promoting the productivity of the entire society. Therefore, the government has increased its investment in financial technology to provide a strong guarantee for technological innovation. In addition, in order to deepen the fund management reform of scientific research projects and solve the fragmentation of science and technology plans, five major scientific and technological plans (special projects, funds, etc.) have been integrated. However, some scientific research management departments and scientific research personnel of universities and institutes have reported that there are some problems in the way of using funds for scientific research projects, such as too detailed formulation, inability to modify and attaching more importance to things than to people, etc. In recent years, the Central Government of China has issued a whole set of policies to reform the configuration and management of scientific research fund, which proposed a series of measures of “relaxation + incentives” in terms of the proportion of funds, the scope of expenditure, and the setting of subjects. However, how about the effects of these policy implementations? Has the implementation meet the original intention of the reform? How do the researchers think about the reform policy? In order to answer these questions, implementation evaluation for the policy reform of research fund management is conduct, to identify the problems in policy implementation. Based on this evaluation, policy suggestions to further promote the allocation and management reform of scientific research funds will be proposed. 2.Research Methods An evaluation model supported by literature review, data analysis, questionnaire and field interview is established, to conduct an implementation evaluation for the policy reform of research fund management. It is mainly carried out by holding symposiums and questionnaires among the relevant scientific research management departments, sampling interviews and questionnaires among the researchers, and consulting archives and materials of 10 universities or scientific research institutes. Following that, the results of the evaluation and problems in the policy implementation is proposed. Based on this evaluation, policy suggestions to further promote the allocation and management reform of scientific research funds are proposed. This study also provides a case study for the development of policy evaluation theories and methods. 3.Conclusions Through qualitative analysis of interview data, archives data, funding data at the institutional level and more than 3000 questionnaires collected by scientific researchers, the results are as follows: (1) The total investment in basic research gets less, and the proportion of basic research in the investment structure of financial science and technology is low; (2) The policy of canceling the proportion of labor costs and expert consultation fees in the direct budget of scientific research project failed to be effectively implemented;(3) The increase in the proportion of indirect expenses and the cancellation of proportional restrictions on performance expenditure was not implemented;(4) The use of fund and reimbursement is complicated;(5) The auditing of scientific research projects is frequent, and the inspection standards are not uniform.

ABSTRACT. Introduction:

The use of open-access publishing is an ongoing debate in academic research. There are studies examining whether open access increases, characteristics gleaned from bibliometric data on open-access publishing, predatory open-access journal publishing, and attitudes toward open-access. However, there appears to be a gap in the literature in examining the characteristics of individuals and institutions that are publishing in open-access journals beyond the characteristics available through bibliometric data. This study asks, are there unique characteristics of individuals and institutions publishing in open-access journals? We seek to examine individual career-level strategies that might promote the use of open-access as a production strategy. Further, we look to examine institutional differences within disciplines and institutional settings that might impact the choice of researchers to publish open-access.

Data and Proposed Methods:

This study uses data from the Directory of Open-Access Journals (DOAJ) and the NETWISE II project from the Georgia Institute of Technology to examine the characteristics of those publishing open-access in DOAJ journals. By using the DOAJ to create the list of open-access journals, we ensure a minimum level of quality to the open-access journals in our study as the DOAJ provides oversight to their directory and attempts to prevent predatory journals from inclusion in their list. The survey implemented by NETWISE II provides a greater level of detail into institutional characteristics than can be determined from scraping bibliometric data alone. By cross-referencing the list of articles published by respondents for the NETWISE II survey with DOAJ journals, we can see who has published in these open-access journals.

The proposed methods for this study are to run both logit models as to whether an individual has ever published in a DOAJ journal, as well as OLS regression models on the number of articles published in DOAJ journals by the respondents. The models will examine individual characteristics including gender, race, citizenship, and professorial position, as well as institutional characteristics including institution type (e.g., Research I, Research II, etc.) and discipline. The models will be run overall, as well as within institution types.

Discussion:

While other research has debated the merits of open-access publishing, this paper does not seek to promote or discourage open-access publishing. Rather, this study seeks to better understand whether those who choose to publish in open-access journals are distinct from other researchers in their field. Although this study is exploratory, it will contribute to the literature by helping to understand further the extent to which individual and institutional characteristics of academic researchers influence their choice to publish open-access.

ABSTRACT. Most Published Research Findings Maybe False, But Some Worth Doing: An Application to Public Health

1. Background Reproducibility and replicability are the two desired properties of correct scientific research (Leek & Peng, 2015). Previous studies about replications have shown that high-profile scientific findings cannot be reproduced with high-power experiment design (OpenScienceCollaboration, 2015). A study by Ioannidis (Ioannidis, 2005) showed that simple assumptions about how science is conducted lead to the striking conclusion that most published studies are wrong. Research is not all about the likelihood of finding true effects. The extrinsic value of research is also an important topic for funding agencies, which they consider on top of such likelihoods (Piwowar, 2013). Also, researchers assess how their research extends beyond their labs, into the real world (Mansfield, 1991; Piccoli & Wagner, 2003). In this study, we propose the question: ”should we continue working on research if the chance of getting correct findings is low?" To answer this question, we model research as two categories of a decision problem (adapted from Bayesian decision theory) and examine our proposed question with a public health application. 2. Methods In this section, we present details and an extension (a model of cost and benefit analysis) to the Ioannidis’s framework (Ioannidis, 2005). 2.1 A mathematical model of the correctness of research Statistical tests in scientific research often rely on establishing significance test thresholds such as α (for Type I error — the rejection of a true null hypothesis) and β (for Type II error — failing to reject a false null hypothesis). However, the probability of a true relationship existing depends on the prior probability of a relationship to exist in the first place. Ioannidis defined R as the prior ratio between the number of true relationships and the number of false relationships (Ioannidis, 2005). Combining certain values of the prior R, α, and β, we can arrive at the surprising result that many studies are more likely to be false than not. We now introduce the notation to show exactly why this occurs. The pre-study probability of a true relationship is P(TRT) as R, and the probability that researchers obtain a positive finding given a true relationship is P(RFP|TRT). There are two other factors — bias and multiple independent tests, which change the value of PPV (Ioannidis, 2005). Bias in research design, denoted by µ, can reduce the veracity of published research. Ioannidis defined it as "the proportion of probed analyses that would not have been ’research findings,’ but end up presented and reported as such.” After bias (u) is involved, PPV with bias is PPVµ = (Rµβ + R(1 − β))/(R + α − Rβ + µ − µα + Rµβ) (1) Now, we can consider the number of individual studies on the same research question. Let us assume that the number of independent studies with equal power is n and that at least of them finds a positive research finding. With this definition, PPV becomes: PPV = (1 − βn)R/(R + 1 − (1 − α)n − βnR) (2) Given the above mathematical framework and if we don’t consider bias and multiple independent studies, we can know “A research finding is thus more likely true than false if (1 - β)R > α” (Ioannidis, 2005). As most of the studies consider α is 0.05, then even if studies are using high power design (β is 0.2), then R is still a critical factor to consider. R should be greater than 0.0625 if there are more true research findings than false research findings. However, this Ratio R is unknown, and this left us the question of how to decide research with such uncertainty.

2.2 Two-Category Gain Estimation with Bayesian Decision Theory When we make decisions under uncertainty, scientists do not just consider probabilities alone. Scientists consider the expected value of one outcome from the decision as the multiplication of return and the probability of if people will have the outcome as they thought. Therefore, it is possible to answer the proposed research question as a decision-making problem under uncertainty. Therefore, we apply Bayesian decision theory to answer this question. Bayesian decision theory is one way to evaluate different classification decisions with probability and cost associated to them – then measure the expected value for each possible outcome (Duda, Hart, & Stork, 2012). Two-category gain estimation is a special adaption of Bayesian decision theory. In general, we define two states of nature w1 and w2 which denote two categories that we cannot observe. We have priors about them P(w1) and P(w2). Therefore we cannot make estimation only based on P(w1) and P(w2) because we have some observations as data. We denote the observation with a feature x associated with the two categories. Here, the feature x is the given research findings. Therefore, with the feature x, we have two conditional probabilities P(w1 | x) and P(w2 | x), since we can find the conditional probabilities with the training data. Decision theory, then, defines actions that we can take. We will define two actions a1 and a2 which denote the action of choosing category 1 and 2, respectively. The posterior estimation of the class is related to different gains. These gains are defined by λij and represent the gain of taking action ai if the category we would be true. Using Bayesian inference, we can estimate the expected gain of each action as follows gain(a1|x) = λ11P(w1|x) + λ12P(w2|x) (3) gain(a2|x) = λ21P(w1|x) + λ22P(w2|x) (4) By trade-off with the two gain functions (comparing the gain), we can find the optimal decision for the test data with the training data. 3. Results: An application to heart disease research We now examine how these statistical and cost-benefit analyses translate into heart disease research. We collected data from the National Institute of Health (NIH) and the Centers for Disease Control and Prevention (CDC). According to these data, heart disease cost the U.S. around 200 billion dollars each year, including the cost of health services, medications, and missed productivity. Meanwhile, the NIH spends 1.37 billion dollars on heart disease research (CDC, 2017; NIH). We use these data to simulate our classification model to find the best trade-off point of R if we let whether the relationship is true to be wi and whether we do research to be ai in our proposed model. The best trade-off points of R in this simulation are 0.000830 and 0.00637 for the simulation with the best research practice (small bias µ = 0.05 and high power and no other independent studies n = 0) and the worst research practice (lower power, and many independent studies, n = 10), respectively. These results indicate that even though published studies in the area are likely false, researchers should still work on heart disease research.

ABSTRACT. The US National Park System (NPS) is an important part of American natural, historical and cultural heritage. At the core of the NPS mission is to conserve protected areas, ecosystems and scenic landscapes across the country for present and future enjoyment. The NPS and other protected areas also provide many services to society - from educational and recreational spaces to the supply of drinking water, in addition to supporting regional tourism economies that attract more than 300 million visitors a year. In addition to these roles, the NPS are also the grounds for a diverse array of scientific activities that are directly impacted by dynamic US policy. In this work we evaluate how external political and economic policy changes affect the direction and scope of science conducted in the NPS, which is relevant to the environmental and biological science, eco-tourism, resource management, parks planning and stewardship. Background Through the more than 100 years of history of NPS, US parks stewardship practices have evolved and many of them have been used as a model emulated by other countries (especially developing ones) for managing and protecting their own natural resources. The idea of centering around a crown jewel, e.g. Yosemite National Park, have been replicated in other countries, and entire national systems of protected areas have been drawn in the image of NPS. Thus, decisions made around NPS not only have implications in NPS but also could derive in the countries that constantly mimic the NPS [1]. Although changes in practices or organizational structure of parks could be assumed as coherent improvements, some changes might produce the opposite effect. Indeed, recent controversies such as the PADDD (Protected areas downgrading, downsizing and degazettement) have pointed out how changes in the political status of protected areas could dramatically affect the ecological and social integrity of the protected area and its surrounding landscapes [2]. One of the most conspicuous political changes in NPS is the 1963 Leopold report, which reoriented the entire system towards a more scientific decision making process increasing the capabilities of the entire system; other policies such as the Endangered Species Act could constrain the abilities of NPS to face disrupting scenarios [3]. This is to say, policies that are in principle aligned with the mandates of the NPS can nevertheless have decohering impacts. Currently, one of the main assumptions in NPS managerial practices is the use of adaptive management in order to identify practices that are producing undesired effects and take actions to correct them. Science is essential for implementing and re-designing management plans in natural parks or other types of protected areas. In principal, scientists develop baselines, assess the impact of managerial strategies and provide most of the knowledge required in the decision making process. An important part of variation is the background and training of park scientists, which may have different sources such as universities, private and public. Cognitive, psychological and physical distance between scientist and decision makers represents another significant hurdle, one that has been shown to affect the knowledge transfer, acquisition and implementation into managerial plans [4]. Consequently, the ability of entities (parks, managers, scientists) to adapt their behaviors is a relevant topic worth exploring using data driven empirical methods.

Against this background, we performed an extensive analysis of the collaborations between academics and NPS decision makers and practitioners. Methods In the half century since the Leopold report, scientists have been hired by NPS and played an important role in NPS managerial activities. These NPS scientists are bridges that facilitate knowledge flow between academia and parks, thereby facilitating the diffusion of scientific knowledge into and out of NPS domains. This particular configuration between academia and decision makers is an effective structure for ensuring the use of the best available science due to the shortness of cognitive and physical distance. With this in mind, we study the historical dynamics of scientific knowledge production in the NPS. We base this study upon research outputs published between 1990 and 2018 contained in the Web of Science (WOS). All together we gathered 36618 documents published worldwide, of which 17044 at least one author is affiliated to an US institution. We assessed the dynamics of number of publications, topic diversity and asymmetries in collaboration for different scales of aggregation (Global, USA, NPS). Results and discussion Our results assess the degree to which policy changes and other external factors affect scientific productivity and collaboration rates relating natural park science. First, we show that during organizational changes in the NPS the realized rate of scientific production was lower than the expected and after the restructuration the NPS productivity grown slower. Productivity associated with different US is variable and does not appear to be strongly correlated with political changes. Second, we find that budgetary factors are more strongly related to changes in the topical diversity of park science. In this sense, our results speak to the agenda prioritizing that is a notorious direct effect of budget shortcuts. Moreover, some of the effects of budget can impact non NPS institutions reducing their capabilities to conduct science in the parks. Finally, we measure the degree to which political changes do affect the network of collaboration that is accessible to official park scientists within the NPS. Political changes that reduced the social capital in parks were associated with more uneven distribution of collaborations and consequently a smaller landscape of knowledge accessible for decision makers and practitioner in parks. Our data-driven approach addresses the implications of some policies in the research associated to NPS indicating that some policies have impacted negatively the scientific productivity slowing the rates in which science is produced and the collaboration patterns of parks as well as other institutions as universities. These results suggest that NPS restructuring have constrained the opportunities for conducting research in the national parks, making the knowledge needed in the decision making process harder to reach. A lack of information or access to the best science available could negatively influence the adaptive management in the NPS might produce negative unintended consequences for conservation, thereby reducing the ability of the NPS to preserve these lands for future enjoyment.

References [1] Schelhas, J. (2010). The US national parks in international perspective: the Yellowstone model or conservation syncretism. National Parks: vegetation, wildlife, and threats. Nova Science, New York, 83-103. [2] Kroner, R. E. G., Krithivasan, R., & Mascia, M. B. (2016). Effects of protected area downsizing on habitat fragmentation in Yosemite National Park (USA), 1864–2014. Ecology and Society, 21(3). [3] Wagner, F. H., Gill, R. B., Foresta, R., McCullough, D. R., Porter, W. F., Pelton, M. R., & Salwasser, H. (1995). Wildlife policies in the US national parks. Island Press. [4] Lowell, N., & Kelly, R. P. (2016). Evaluating agency use of “best available science” under the United States Endangered Species Act. Biological Conservation, 196, 53-59.

ABSTRACT. Enormous attention has been given to family-friendly policies in higher education as the boundary between work and home has been blurred. Family-friendly policies are expected to create a healthy work environment supporting the careers of scientists in higher education (Feeney & Stritch, 2017; Mayer & Tikka, 2008). Scientists working in higher education institutions are faced with demanding and taxing workload (Post et al., 2009; Soat, 1996). This is especially for female faculty due to the expectations for women to meet traditional gender roles (Blackwell et al., 2009; Matos, 2015). Those expectations induce female faculty to strategically hide family commitments (Drago et al., 2006), to change positions (Xu, 2008) and eventually to discontinue their investment in career advancement.

Work-home conflict occurs when demands from the work domain are incompatible with demands from family. Failure to maintain work-life balance (hereafter, WLB) increase faculty’s intention to leave, reduce the productivity and organizational performance (Beauregard & Henry, 2009). Therefore it is critical in understanding what leads to WLB and how these determinants vary across genders to manage STEM workforce and to promote gender equity among scientists. Prior scholars have explored individual and organizational factors that influence WLB (Pichler, 2009; Shrestha & Joyce, 2011). Yet, few studies have looked at how the composition of faculty's social network can affect their WLB. Depending on how the network is composed of, network can mitigate work-family conflict by reducing stress and eventually have a positive impact on work outcomes such as scientific products.

Drawing from social network literature and social support literature, our study addresses two research questions: (1) Does having multi-functional social network improves WLB? and (2) Does the relationship between multiplex ties and WLB differ between female and male scientists?

Using social network theories, this study takes the characteristics of network into account and specifically focuses on the implications of multi-functional network. Multi-functional networks are composed of multiplex ties. Multiplex ties are defined as ties the existence of multiple relationships between two individuals (Ferriani, Fonti & Corrado, 2012). In this paper, we specifically focus on three types of support that scientists can receive from their networks: support for teaching, support for research and support for career development. By looking at the teaching advice network, the research collaboration network, and the professional development network, we assume that overlapping ties in those networks will make the individual’s network multi-functional.

To manage the balance between work and home, having multiple supports from professional network is ideal. For instance, research support provides opportunities for submission of grant proposal and production of patents (Cross & Sproull, 2004; Ingram & Roberts, 2000) which can decrease stress and burdens stemming from publication or grants requirements. Teaching support offers advice related to teaching activities such as student interaction and course management. This type of support alleviates the demands related to teaching performance. Therefore, scientists who have more multiplex ties can gain more professional resources and support from their networks, reduce the substantial level of burden coming from work which in turn enables individuals to manage a balance between work and home.

While there is no big difference for number of multiplex ties across gender (Ibarra, 1992), the impact of multi-functional network on WLB can vary by gender since women are more vulnerable to balancing work and home. It is harder for female faculty to keep WLB since they face the same pressure from university to improve performance (increase productivity) as male colleagues while they are expected to take care of home and family when male counterparts are not. Hence, female faculty are more likely to benefit from multi-functional network for it provides more professional support that resolves work-home conflict.

Using a 2011 NSF-funded national survey on academic scientists in four STEM fields (biochemistry, biology, civil engineering and math), we explore the interactive effects of multifunctional network and gender on WLB. Using linear regression model, we test hypotheses controlling for individual characteristics (personality, marital status, rank, nationality, etc.), work characteristics (work hours, workload, etc.) and organizational characteristics (working environment, institutional type, field, etc.). We expect that faculty who have more multiplex ties within their social networks will have better WLB and that the effect of multi-functional network will be more salient for female faculty.

The results have both theoretical and empirical contributions. First, we contribute to WLB literature by integrating theories from the fields of social network, gender equity and personnel management. Our research provides insight regarding how the characteristics of social network can affect WLB and how the influences differ between male and female. Second, this study can provide implications for family-friendly policies in higher education. We propose that the existence of multi-functional networks can advance WLB, especially for women. Higher educational institutions, therefore, can adopt relevant policies or practices to enhance scientists’ WLB and promote gender equity, which eventually can help organizations attract, recruit and retain employees as well as enhance scientists’ productivities.

This research was funded by the US National Science Foundation (Grant # DRL-0910191), (Co-PIs: Dr. Julia Melkers, Dr. Eric Welch, Dr. Monica Gaughan).

ABSTRACT. Over the last decade, social innovations have gained significant popularity. The core ideas of the social innovation concept – ‘being social’ and ‘being innovative’ – are usually contrasted to the business innovations (so-called ‘market innovations’) that possess only ‘innovativeness’ feature. Academic and non-academic literature have described numerous cases of social innovations all over the world (for example, see Howaldt, Kaletka, & Schröder, 2016; Pisano, Lange, & Berger, 2015; Rehfeld, Terstriep, Welschhoff, & Alijani, 2015). To my knowledge, there are no quantitative studies that apply a range of possible socio-economic and political factors and explain the differences in social innovation activity across the regions. A vast body of academic research and research projects highlighted the roles of policy, universities and active citizens as the primary drivers of social innovations (for example, Davies & Simon, 2012; Jurgen Howaldt, Kaletka, Schröder, Rehfeld, & Terstriep, 2016; Moulaert, 2005). Despite this, the preferred method of social innovation analysis among researchers and practitioners is case studies, and the quantitative attempts are limited to the descriptive analysis. In contrast, the geography of business innovation has a long history of studies that investigate why innovation emerge in some places but not the others, and what factors are associated with the innovative output: for example, university and industry R&D expenditures, social capital and human capital, networks and connections. There are several gaps not addressed in the literature. First, there have been limited attempts to test empirically what factors are associated with social innovation activity – such as the drivers identified in the social innovation reports, factors derived from the studies of relevant organisations such as non-profits and social enterprises; or the findings of the business innovation literature outlining the roles of R&D, human and social capital. Also, there is limited understanding of whether the factors relevant for social enterprises, social innovations and non-profits (presence of social needs and institutional gaps such as the rise of poverty and unemployment) have any influence at the business innovation activity.

Methodology

The current paper attempts to answer the question of whether spatial distribution of social innovations and business (‘market’) innovations is explained by similar economic and societal factors. The literature on business innovations has shown how human capital, social capital, R&D expenditures, cities or industries characteristics influence the magnitude and concentration of innovation activity. In turn, there is rich empirical evidence produced on the location of social enterprises and non-profit organisations, and many detailed case studies outlining the drivers of social innovation activity. To address these gaps, the current study introduces six regression models that contain two dependent variables reflecting innovation and social innovation activity accordingly and different combinations of independent variables derived from the literature on business innovations and social innovations (and related concepts such as social entrepreneurship and non-profit organisations). Due to the data availability, 65 NUTS 2 regions were analysed in the current study. These regions belong to eight countries: Austria, Czech Republic, Denmark, Spain, Finland, Italy, Sweden and Slovakia. The remaining EU states were excluded due to shortage of data. Business innovation related variables include university and industry R&D expenditures, human capital (measured as a number of people with higher education degree), social capital indicators (trust to other people, institutional trust, associational activity - share of people participating in various organisations such as sports clubs, voluntary organisations). Variables related to social innovation, non-profits and social enterprises include political activity, associational activity, GDP level, share of people employed in public services, growth of unemployment during 2012-2016, growth of GDP during 2012-2016, poverty level.

Results The findings indicate that patent activity and social innovation activity tend to concentrate in more advantaged, heterogeneous and economically developed regions. The research findings have also revealed that regional social capital has a greater impact on both the patenting activity and social innovation activity, compared to other factors. As expected, patent activity was likely to be higher in more demographically diverse places with the growth of GDP per capita, population and fewer poverty rates, better water, medical and sanitation conditions. Unlike previous studies ( Barro, 1991; Dakhli & De Clercq, 2004; Mankiw, Romer, & Weil, 1990), University R&D expenditures and share of people with higher education turned to be rather insignificant. Business sector R&D expenditures showed a significant positive impact; these effects are consistent with the dataset which included many regions hosting high-tech and low-tech industries such as mechanical, automotive, transport-related production, engineering, and others. The variables indicating Social needs turned out to be moderate predictors of innovation activity when ‘business innovation’-related variables are not taken into account. One of the social capital indicators - associational activity showed much stronger influence on the patent activity compared to the R&D expenditures and human capital. These findings are consistent with the research highlighting the role of networks for knowledge dissimilation and innovation. R. Burt (1992, 2000), Coleman (1994), Granovetter (1973) and several other authors offered a network perspective on social structure and social relations, where the number of connection and actors’ placement in the network play an important role. The propensity of the firms to engage in the business networks might be therefore caused by or related to the overall tendency of people to join formal and informal groups, clubs and associations. Social innovations location patterns revealed some similarities to business innovations. Similar to business innovations, social innovations tend to concentrate in wealthier and heterogeneous areas that accommodate a higher number of foreign-born populations, and near the sources of industrial R&D rather than knowledge sources (e.g. universities). Similar to patent activity, R&D expenditures in the higher education sector showed no significant relationship to social innovations. Contrary to the expectations, social innovation projects are more attracted to the areas with low levels of interpersonal trust, and where people both politically active. The influence of the associational activity was insignificant and institutional trust did not reveal any substantial impact on social innovations. Also, unlike business innovation, social innovation showed positive linkage to the population. Unlike patent activity, the variables denoting the rise of unemployment and population as well as poverty levels were rather insignificant for social innovations. The research also did not present any evidence that social innovations locate in the areas of a shortage of public servants, rather areas with more people employed in the field attracted more social innovations. Among all factors, political activity, GDP and population showed the strongest impact on social innovation. In that sense, social innovations resemble non-profit organisations and social enterprises that are also attracted to wealthier rather than deprived places (Bielefeld, Murdoch, & Waddell, 1997; Joassart-Marcelli & Wolch, 2003; Katz, 2014). However, not all wealthy and developed places would offer a fertile environment for social innovation development, but only the ones with a politically active civil society. The regional socio-economical context matters for social innovations and business innovations to a different extent. The expected triggers for social innovations (poverty, unemployment) showed greater relevance to business innovations than social innovations. On the contrary, factors usually associated with business innovations such as heterogeneity, wealth and industrial R&D were significant for social innovations. While current research shed some light on the spatial location of social innovation, further data collection and research is needed to investigate the influence of the variables on extended geographical area.

ABSTRACT. We exist in an era of “big science” with continual year-over-year increases of scientific publication (Price, 1963; Price, 1983; Larsen and Von Ins, 2010). Because scholarly publishing follows the inverse square law, meaning that works that enter publication benefit from cumulative advantage processes reflected in ever-growing citation and breadth of use, a diverse array of subsequent uses for that knowledge become theoretically possible and evident. However, such access and use of peer-reviewed scientific and technical information (STI) has traditionally been relegated to the “invisible college” of scientific communities of practice and scholarship. Scientific communication is nothing new. What is relatively new, however, is an accelerating trend towards "open access" for these works. Once only available to those with university or scientific affiliation, this peer-reviewed literature is increasingly globally accessible to anyone with internet access. Open Access policies and mandates such as the NIH Open Access policy and the European Union’s nascent “Plan S” typically rely upon first principles including legalistic and economistic arguments to advance the cause of open access. Potential, or imagined, uses of open access materials are imagined as under-resourced scholars operating in the Global South or intended to speed the pace of discovery and innovation within the triple-helix of university, industry and government sectors. Left out of these imagined uses are the huge swath of potential users who are not scientists, professionals or scholars, often referred to as "lay" persons.

The proposed research poster explores an understudied interpretation of “open access.” What makes high-quality STI accessible - not simply technologically - but “legible” to the non-scientist or the average lay person who is not steeped in the discipline? Who are the imagined users of open access scholarship?

As part of this effort to understand lay person use of scientific information I leverage a unique data set obtained from Harvard’s DASH Open Access repository. The data consists of 3567 comments from users who downloaded works from the repository. This is a significant data set since users are prompted to “tell their story” about how “Harvard DASH or Open Access has affected you” (Harvard, 2019). Initial analysis suggests interesting trends. First, there are many expressions of gratitude to Harvard University as an institution for making these works accessible. Second, approximately one quarter (24%) of a sample of the 3500 comments indicate a “lay” quality within their comments, with 8.5% indicating more obvious everyday use of STI (for example: “I am just citizen of the US, living in CA, confronting racism every day as a white person, wanting to understand the very earliest origins of slavery in what became the US, especially the economics of it” in reference to the downloaded article “The Slave Trade and Origins of Mistrust in Africa”). Notably, 1.5% of the sampled data set include some variation on the word “curious” or “curiosity” in their stories signaling everyday curiosity and lifelong learning as a small but important quality in laypersons seeking out high-quality scientific research.

Building upon the Harvard DASH comments I also conducted an exploratory analysis of comments on MIT’s OpenCourseWare YouTube channel. Taking a random sample of 14 videos with a total of 807 comments indicates that expressions of gratitude occur in this realm as well. Interestingly, many gratitude expressions are personally directed towards the professor leading the course rather than the institution as was evident with the Harvard DASH data. I performed an OLS regression and identified a correlation between “likes” (e.g. up-voting) and expressions of gratitude in comments. I conclude there is a statistically significant relationship between “likes” and expressions of gratitude in YouTube comments within the MIT OpenCourseWare videos (p < .05). A fuller empirical analysis and interpretation of both the Harvard and MIT datasets will be completed prior to the Atlanta Conference.

Non-scientist use of scientific information is an understudied phenomenon. This research poster advances understanding by analyzing user comments from open access repositories and social media to better understand motivations and dynamics of lay person use of high-quality scientific information. As scientific elites and the science base of society comprise a relatively small percentage of the total global population, lay persons make up the vast "dark universe" of information-seeking behavior. Thus, the proposed research may have significant implications in advancing knowledge about public understanding of science, as well as studies related to open access and science and technology policy.

References:

Harvard University. DASH: Your Story Matters. (2019). https://dash.harvard.edu/

Larsen, P., & Von Ins, M. (2010). The rate of growth in scientific publication and the decline in coverage provided by Science Citation Index. Scientometrics, 84(3), 575-603.

Price, D. D. S. (1976). A general theory of bibliometric and other cumulative advantage processes. Journal of the American society for Information science, 27(5), 292-306.

Price, D. J. D. S. (1963). Little science, big science. New York: Columbia University Press.

ABSTRACT. 2019 Atlanta Conference on Science and Innovation Policy

Summary for Poster /Extended Abstract

Topic: Science and Technology Human Capacity Building in Developing Countries:  Challenges and Opportunities for Guatemala, El Salvador and Honduras

Track: Scientific workforce/ diversity

Key Words: [Scientific workforce, Central America, Developing Countries, Science and Research Human Capacity Building, Guatemala-El Salvador-Honduras]

Developing countries face numerous challenges in the process of building their scientific and technological human capacity; particularly in relation to the training and accumulation of human resources or specialized human capital in science, technology and innovation (ST&I). The lack of organized and sustainable higher education options (master's and doctoral programs), non-existent or low-quality academic programs, lack of research-oriented study options, and other factors are strong contributors to the emigration of talented students from underdeveloped countries to developed countries. At the same time, the consolidation of a global knowledge-economy, the internationalization of higher education and the competition to attract foreign talent in the industrialized countries represent a difficult test for poor regions to keep their already scarce qualified human resources.

In the case of countries such as Guatemala, el Salvador and Honduras –which are considered to be underdeveloped not only in the social and economic aspects, but also in the areas of ​​the science, technology and innovation – the challenges are even greater, since more urgent problems related to poverty, diseases of underdevelopment (malaria, maternal and child high morality), precarious productive and economic systems, natural disasters and climate-change emergencies and others, give more urgency to pressing problems and, therefore, public policies in ST & I are postponed.

Excluding societies, with large inequalities brought in since the colonial era, results  in large portions of their most vulnerable population being excluded from any possibility of academic and professional development. Armed long-lasting internal conflicts -in the cases of El Salvador and Guatemala- have created strong barriers for different groups in rural areas to access to education and reduced the possibilities of training and accumulation of highly qualified human resources to build up a critical mass.

On the other hand, the aforementioned countries have incipient democracies whose civil governments have created fragile institutions in general, and in particular there are very few initiatives to support the governance of science, technology and innovation sectors. As a result of this scenario, we can verify the insufficient existence of highly educated human resources with abilities in academic and scientific research.

A wider research topic undertaken by the authors of the present work studies the process of building capacity in science and technology in Guatemala, El Salvador and Honduras with the support of international cooperation, and as part of this inquiry a particular approach attempts to systematize the challenges and opportunities for building science, technology and research human capacity.

As proposed by Milèn (2001) capacity building is a continuous and dynamic process, builds on what already exists, the construction process has an intrinsic value in itself, allows dealing with continuous changes and is carried out in a holistic or integrated way. It has implications for medium-long term. It has different levels and for the purpose of the present work the focus is place in the micro level: individual - science and technology human capacity.

Human resources or individual focuses on how people are educated, how their knowledge and skills fit with the rigorous scientific research. The focus is on technical, professional, managerial and communication and networking knowledge and skills. It also deals with attracting people for the utilization of their knowledge and skills.

Research Questions to be addressed

1. Which are possible reasons explaining the shortages of S&T human capital in Guatemala, El Salvador, and Honduras?

2. What capacities should be built in S&T human capacity in Guatemala, El Salvador, y Honduras?

3. What support can international cooperation provide in the process of capacity building in human S&T capacity in Guatemala, El Salvador and Honduras?

Methodology and Data Source

The geographical scope addresses three countries located in the Central American sub-region: Guatemala, El Salvador and Honduras. These three nations compose the northern part of the isthmus, sharing common features in their socio-economic development stage, pervasive poverty indicators, inequality and other problematic such as violence and organized crime. Expressly, three other countries of Central America were not considered for this analysis: Nicaragua, Costa Rica and Panama. The similar characteristics in terms of science and technology governance, public policy trajectory, as well as economic structures and challenges common to Guatemala, El Salvador and Honduras allow the investigators to identify similarities and differences.

Through the application of various methods in data collection, the robustness of the process is strengthened and the triangulation gives more validity to the study. Interviews with key actors in the public and private sectors relevant to science and technology. Also specific data collection took place with actors from international cooperation and the research communities in the three countries.

On-line comprehensive survey on S&T human capacity building – research communities: A mapping exercise carried out to identify databases composed by researchers in the three countries to apply surveys for greater representativeness and strengthen the quality and depth of the analysis. In this process it was determined that while Guatemala and El Salvador have registries of researchers with nationwide coverage, Honduras has only partial catalogs of researchers such as that of the obtained from the National Autonomous University of Honduras.

The number of responses to the online survey varied, as multiple obstacles were encountered. From the total number of responses 148 were done from the Guatemala research community, 151 from El Salvador and 56 from Honduras.

References:

Milèn, A. (2001). What do we know about capacity building? An overview of existing knowledge and good practice . Geneva: World Health Organization

ABSTRACT. According to the World Health Organization, 7 million people die prematurely every day in the world due to air pollution. The global cost caused by these deaths is estimated at $225 billion a year. Clearly, the high levels of atmospheric pollution in some areas of the world present a high risk to the health of the population. In particular, Chile has of the most polluted cities with particulate material (PM) in Latin America and the Caribbean. In the cities of south central of Chile, the main sources of PM emissions come from biomass fuel burning for household and commercial heating and cooking. Although there are environmental policies that seek to reduce PM emissions, these policies have not been effective in reducing concentrations of PM10 and PM2,5, being the most critical periods of contamination in winter. This poster seeks to quantify the health effects of air pollution in the three most polluted cities in Chile: Padre Las Casas, Osorno, and Coyhaique. Additionally, the positive impact of the implementation of new technology for the reduction of emissions coming from biomass fuel burning for household and commercial heating and cooking. The technology to be evaluated will be the one that we have developed with a group of entrepreneurs during the last two years. This technology consists of an electrostatic filter that is installed directly in biomass combustion stoves, and its main characteristic is to reduce the emission of PM by up to 90%

Previous studies that show the relationship between the increase in health problems and exposure to high concentrations of particulate material will be used, considering the specific characteristics of each geographical area. A dose-response function will be estimated for each of the three most polluted cities in Chile in other to quantify the health impact of air pollution on these zones. The main externalities that will be quantified are mortality, cardiovascular disease and admission, respiratory disease and admissions, labor absenteeism, and restricted activities. Also, experiments will be carried out that allows having a distribution of the MP emission reduction of the filter, analyzing different types of conditions. Subsequently, the new scenario will consider the estimated reduction to the exposure of the particulate material of the population produced by the installation of electrostatic filters. Several numbers of devices in each city will be considered to know the curve of increase of the welfare of the society according to the number of artifacts that will be installed. This curve will be compared with the public investment that the state must make when implementing an environmental policy that subsidizes the filters.

The expected results will allow us to know an estimate of the health impact that is being generated by air pollution compared to the effect of the implementation of a new technology to reduce emissions of PM, analyzing this situation in the most polluted cities in Chile. These results will allow finding the equilibrium point between the increase in the welfare of society by implementing a new technology to reduce MP emissions and the cost of public investment in environmental policies. Finally, this poster will be useful for the generation of new environmental policies that allow reducing efficiently and in the short term the high levels of concentration of particulate matter in the atmosphere, a problem that still does not have an efficient solution in the cities of the south-central of Chile.

ABSTRACT. Science plays a crucial role in environmental collaboration and governance. As such, science can inform policymakers about the nature of a problem and possible solutions. In many cases, however, science that is conveyed to support decision-making is constrained by the institutional configuration within a given political context thus impacting the effectiveness of the science that is used to address environmental problems.

The Guangdong-Hong Kong- Macao Greater Bay Area consists of 11 major cities with two distinctly different political-economic systems sitting at the outer estuary of Pearl River Delta, South China. Although the area only makes up to in total 56,000 square kilometres accounting for 0.6% of the total land area of China, its combined GDP of US$1.64 trillion has surpassed Russia, which is the 11th largest economy in the world. At the end of 2018, the population of the Area was approximately 70 million, accounting for only 5% of China’s total population. In addition, a variety of natural ecosystems converge in the Greater Bay Area including aquatic ecosystems such as rivers, lakes and oceans; wetland ecosystems such as estuaries, mangroves, islands and fish ponds; and terrestrial ecosystems like farmland, forests, and cities.

The fact that the Greater Bay Area is one of most developed regions in the world as reflected from its large-scale and fast-speed industrialization and urbanization has placed enormous pressure on the environment resulting in disturbing and even destroying these interconnected natural ecosystems. Protecting the ecological environment and restoring the balance between nature and human in this densely populated, highly industrialized but ecologically complex Area do not only require collaboration across cities, levels, and systems but also the critical input of scientific research in order to understand the nature of these issues. Given these, environmental collaboration and governance in the Greater Bay Area is both a scientific and public policy challenge and of great significance to the sustainable development of China and the world.

The environmental collaboration and governance in the Greater Bay Area emerged from the changing social, economic and political environment and strengthened through continuous cross-region interactions. As the governance structure within the region has evolved over the decades, the forms of collaboration and governance have also been adjusted accordingly. Nonetheless, the puzzle that what role has science played in this process against the changing instructional settings remains largely underexplored. In this research, we will address this knowledge gap through tracing the development of science in the process of environmental collaboration and governance. We will answer the following questions: (1) to what extent has science been integrated to the regional environmental collaboration and governance over time; (2) what political and economic institutions within the Greater Bay Area have restricted or enabled such integration of scientific advice in policy over time.

We propose to adopt mixed methods combing large N and small N studies to implement our study. The quantitative approach aims at finding out the trends of the scientific literature that has been published to support the environmental collaboration and governance in the Greater Bay Area over time. A database of scientific literature on environmental collaboration and governance and a database on environmental policies of the region will be built for this purpose. Meta-analysis will be performed to identify the key themes in these databases. The co-occurrence of the words and themes in each database will be analysed. The resulting network will be visualized, showing semantic maps of the scientific and policy discourse surrounding environmental collaboration and governance in the Greater Bay Area. This process makes it possible to address the question to what extent scientific research has been an integrated part of policy discussions and. In the second stage of the study, we will carry out surveys and interviews with authors whose scientific research was closely linked with the policies identified. The survey aims at finding out whether scientists themselves were aware of the incorporation of their results into policies and to what extent they are involved in the process of policymaking. The survey will also help us identify potential candidates for in-depth semi-constructed interviews. For the interview data, we will employ the Institutional Development and Analysis Framework (IAD) as our analytical framework. IAD takes a multi-tier perspective to identify the major types of structural variables that influence the action situation and will guide us to identify the institutional arrangements or factors that have influenced on the science-policy relationship on environmental collaboration and governance in the Greater Bay Area.

So far, there is no research that has addressed the role of science in policy in an environment that is as complex as the Greater Bay Area. Our study will be the first of its kind. It will clary the role of science in this case while contributes to the understanding of the interplay between science and environmental collaboration and governance in many other similar contexts within or outside China.

ABSTRACT. BACKGROUND OF THE STUDY Ever since Independence, India’s encounter with gnawing poverty and stark deprivation, particularly of the weaker and the marginalised sections of society, cajoled India’s planners to moot the development policies with the sole objective to exacerbate growth with equity. That’s why since the beginning of the planning era, the stress had been laid on strengthening and expanding the social sector with the premise that it would boost the inclusive growth agenda, manifesting in equal access to employment and economic opportunities; equal participation in decision making and reduction in poverty and inequality. Human Development is concerned with equity and human welfare base of economic development and it is enhanced by strengthening the social sector which provide basis for economic growth. Social Sector develops human resources by empowering them through education, health, sanitation and community development in conjunction with risk management; promote investment, entrepreneurship, women empowerment, employment generation and poverty elimination. Inclusive Growth incorporates equity, equality of opportunity, protection in market and more productive employment opportunity consider being fundamental constituent of successful growth strategy for India. High and increasing levels of inequality could hamper poverty eradication, which in turn could hinder the growth process. Deciphering the concept of inclusive growth, precisely it implies the reduction of poverty and inequality, and equal access for all in social and economic opportunities. In other words, inclusive growth will manifest in employment generation and participative decision making both at micro village level and the macro level of whole economy. Inclusive innovation is considered when a positive impact is observed on the livelihoods of the excluded groups using technology. A strategy of inclusive innovation encompasses the emphasis on education, health, and other basic public facilities along with policies which aim at improving livelihood support and increasing employment. The present study endeavours to examine the impact of three components of social sector development viz. education, art and culture; medical and public health and family welfare on inclusive growth in India. India's policy for social sector development related with inclusive growth is intended to be investigated. MODEL ESTIMATION Data for the present study have been taken from various reports of World Bank, Planning Commission, Economic Surveys, Annual Budgets, newspapers and magazines. The study covers the period from 1985-86 to 2015-16. All the monetary values have been expressed in constant prices of 2004-05 and all the values are deflated using GNP deflator. In the present study, the concept of inclusive growth is represented by three indicators viz. (a) poverty reduction, (b) equal access to employment opportunity and (c) social and economic inclusion in the growth process. The reduction in poverty was represented by declining absolute poverty and dissipating income inequalities in India. Equal access to employment opportunities indicated by rising ratio of female male in labour participation rate, universal employment generation through MGNREGA and ascending percent share of employment of SC and ST classes. Rising social and economic participation is designed by escalating gender parity index, increasing female to male ratio, rising gross enrolment of reserved category and growing political participation of reserved categories. After construction Index of Inclusive Growth, backward stepwise multiple regression analysis is deployed by regressing various indicators of education, art and culture; medical and public health and family welfare in India on inclusive growth index as dependent variable. In this, the formulation of model is based on log of both independent and dependent variables. The independent variables as elaborated above are presented in equation are Expenditure incurred on Education as proportion of Total Central Government expenditure; Gross Enrolment Ratio; Pupil Teacher Ratio; Expenditure on Health and Family Welfare as proportion of Total Central Government expenditure; Number of Hospitals Beds available per thousand population; Number of Doctors available per thousand population; Number of Health Centers available per thousand population; Percent of Patients recovered under Epidemiological diseases out of total population and Percent of child immunized out of total child population. RESULTS KMO measure is 0.829 indicating a good measure and Bartlett’s test of sphericity shows a significance level of 0.00, a value that is small enough to reject the null hypothesis. Both measures indicated the good model. The first two principal components explained more than 80 percent of the variation with 73.243 percent and 10.943 percent. Using the proportion of these percentages as weights on the component score coefficients, an inclusive growth index was constructed from two extracted components ranges between 0 to 1. The Inclusiveness Index of growth reveals increasing trend initially till the year 2008-09 where again it starts plunging. The analysis shows a slower trend growth rate of inclusiveness in India. Inequality of opportunities among different sections of society is the need of an hour and hinders the growth to make it inclusive. Equalizing of opportunities improvise the equality in access of benefits to strengthen livelihood strategies through empowerment. One of the major reasons for this slow growth is empowerment of marginalised sections including women. This involvement of including the excluded section would indicate the inclusive innovation if technology used. There is need to re-conceptualisation the inclusive growth as complete inclusion is absent in all spheres of human society and far from the balanced development. Once we have constructed a composite index of inclusive growth, we would regress the components of social sector with this composite index to access the relative contribution towards achieving the proclaimed objective of inclusive growth. Adjusted R2 of model is 60.1 percent of the variation of inclusiveness. The DW statistic is 2.032 and therefore the data is not auto correlated. Anova test of Model explained that model is well fitted as the p value is highly significant. It was found that the variable “Number of Doctors available per thousand population” exudes negative association while the variable “Gross Enrolment Ratio” contributes positively on inclusiveness. Since the beginning of the planning era, the stress had been laid on strengthening and expanding the social sector with the premise that it would boost the inclusive growth agenda, manifesting in equal access to employment and economic opportunities; equal participation in decision making and reduction in poverty and inequality. The analysis shows a slower trend growth rate of inclusiveness in India. India is in dire need of innovative and advanced strategies especially for rural health services. Investment in doctors and nurses may potentially improve the services rendered to patient. There is need to ensure the greater effectiveness of the existing social sector programmes launched by government on the inclusiveness. The gaps have to be found out and eliminated to facilitate inclusive development in future with the use of technology. Inclusive innovation is still a new concept in India.

ABSTRACT. The objective of the research is to compare the Canadian aerospace industry, linking business models (BM), intellectual property (IP) and innovation management, a situation little explored in the literature. This exploratory study presents how companies in the sector develop their strategies for open innovation processes. With this perspective and according to the Chesbrough's taxonomy (2006) , three types of open BM can be found: a BM conscious of its environment, which seeks new ideas, innovations, technologies or personnel that can improve its internal processes; an integrative BM, in which internal and external knowledge is an integral part of the daily activities of the company; and finally, an adaptive BM, which responds to the dynamism of the market, adapting to the different circumstances that arise. To characterize the companies throughout the three categories, a questionnaire was constructed that included the dimensions mentioned above. The companies surveyed were those that participated in the production processes, belonging to the aeronautical, space or defence sub-sectors, excluding airlines and airports. The questionnaire was answered by 71 companies that represent 10.9% of the database of companies that were considered viable to answer the questionnaire. For most questions, the seven-point Likert scale was used. The companies are distributed in four regions: Ontario (42%), Quebec (37%), Western Canada (14%) and Atlantic Canada (7%). Of the total number of companies, 85% practice R&D to some degree. The results show a clear dominance of product innovations (73%) and process innovations (72%). Results in accordance with the high-tech companies that focus their efforts on diversifying their products, penetrating new markets, reducing costs and production times. A large proportion of companies (70%) use some type of OI practices for the exchange of value. For this group of companies, the importance given to inbound practices (3.4 / 7.0) was greater over outbound practices (2.9 / 7.0). Of 17 OI practices analyzed, the best score was obtained by informal knowledge networks (5.2 / 7.0), because companies consider informal relationships with members of their network, an effective and reliable source to obtain new knowledge. Co-creation carried out with support of customers and consumers (4.8 / 7.0) ranked second in this category. This result shows that the integration of these actors into the innovation process allows companies to generate and develop new knowledge that allows them to obtain economic benefits as strategic and, at the same time, reduce the uncertainty and risk associated with development. Aerospace companies consider IP as an important mechanism against competition. 84% of companies use at least one method to defend their IP, giving relevance to strategic methods over formal methods. Industrial secret is the most used method (5.4 / 7.0), followed by lead-time advantage of competitors (5.2 / 7.0) and complexity in the design (4.9 / 7.0). Through the statistical analysis “Two-Step Cluster”, it was possible to identify the natural grouping of open business models that existed in the sector. The representative themes for the analysis identified: the practice of R&D within the company (one dimension); the way of working with open innovation (one dimension); OI practices used, inbound or outbound practices (17 dimensions); as well as, how to manage intellectual property (22 dimensions). In these last two cases, due to the large number of dimensions in relation to the number of respondents, it was necessary to reduce the dimensions using a principal component analysis using a varimax rotation. Of total companies that answered the questionnaire, 50 had some degree of openness in the development of their BMs. The result showed that 38% of companies operate under a BM aware of their environment, 62% implement an integrating BM and no company develops an adaptive BM. The two resulting groups are significantly different, companies that have an integrating BM obtained a systematically higher score in all the dimensions on which they developed BM aware of their environment. For the first case, companies are conscious of their environment and their innovations, synchronize strategic planning with OI practices, especially the inbound ones. That is, external ideas, new technologies, suppliers and customers are considered new collaborators in the generation of knowledge. These companies take advantage of the IP and direct their BMs to a more open direction. For the second group of companies, these characteristics take on more strength, the bonds between the internal and external actors of the company become stronger. Now, the company is aware of the factors that affect its productivity, taking advantage of them in a more strategic way. To better understand the panorama of companies that use open BMs, the resulting groups were compared with the dimensions of external financing resources, protection methods, as well as internal and external partners. Companies with an integrated BM give greater importance to the mechanisms that contribute to the generation and exchange of value on companies that have a BM aware of the environment. In the case of financing, government subsidies for the development of internal and/or collaborative projects are considered very important for the sector. These financing methods allow companies to obtain complementary resources to develop their R&D. Companies are more open, they need collaborative structures that are economically sustainable. For this reason, companies develop collaboration agreements and solid strategic alliances, which allow operating in an open manner, establishing trust between the parties involved. On the side of protection methods, the predominance of strategic methods over formal ones is still evident. But it is noteworthy, the great importance that patents win for companies with an integrated BM with respect to those who have a BM aware of their environment. Finally, the dimension of internal and/or external partners, presents internal employees as the most important partners for the company when developing innovations.

ABSTRACT. In this paper, we explore the impact of the economic cycle on university-industry scientific knowledge co-creation output. According to our university-industry cycle theory, there are reasons to believe that economic growth will either encourage or discourage firms to co-create with universities, but the former is more likely to occur in crises and the latter in expansions. To verify this, we use data on Spanish firms’ co-publications with universities from 2000 to 2016, which includes the Great Recession started in 2008. Our results agree with the theory, so that when the economy grows fast, firms co-publish less with universities and when the economy grows slowly or contracts, firms co-publish more with universities. Policies to promote university-industry scientific knowledge co-creation output could adapt to the phase of the economic cycle.

ABSTRACT. Background

The deployment of new technologies appears to be an important part of the solution to deal with some of the grand societal challenges recognized by the UN Sustainable Development Goals such as climate action and sustainable cities and communities. The role of policy makers in pushing the development and deployment of these technologies is unclear. A new paradigm of science, technology and innovation policy making indicates that the design of interventions has to combine top-down and bottom-up approaches. This mission-oriented approach involves the participation of different sectors of society. In this context, it is important to understand the motivations behind the private and civil sector’s decisions and interactions.

The transportation sector is one of the main contributors to greenhouse gases emissions. Alternative fuels (AF) and advancements in vehicle technologies could decrease transport emissions by 35-57% relative to 2010 levels (Howey, 2012). This study uses an evolutionary game theory approach to model the network of stakeholders involved in the process of adoption of alternative fuel vehicles for freight transportation. The objective is to understand the effect of different regulation alternatives taking into account the responses from the vehicle manufacturers, fuel providers, and consumers.

Methods

The use of evolutionary game theory to model the interactions between different sectors of society was proposed by (Encarnação et al., 2016). We consider three populations vehicle manufacturers, fuel providers, and consumers that are initially under a technology lock-in, preferring fossil fuel technologies for their operations. This model simulates random interactions between individuals in each one of the populations. Over time, each individual can decide to change its strategy based on two aspects. First, the probability of an individual opting for a new strategy is frequency dependent. This means that it is more likely that someone decides to implement a new technology if a higher fraction of that population Is using that technology. Second, the probability of an individual changing its strategy also increases when the payoff received by using the new technology is higher than the old one. This means that successful behaviors spread faster. When building the payoffs table, we consider that the public sector can choose to implement diverse incentives such as subsidies, providing refueling infrastructure or R&D support. Also, the model considers the possibility of the civil sector boycotting the private sector in case their strategies do not align.

Preliminary results

The first set of preliminary results indicate that, without policy intervention, it is not possible to escape the technology lock-in state. Therefore, the public sector has to move first and provide some incentives for the private and civil sector to adopt the new technology. Then, we implement a set of combinations of policy interventions and observe the fraction of each sector that decides to support alternative fuel vehicles over time. The results from the simulation indicate that diversified and targeted incentives increase the proportion of freight transportation firms adopting the new technology. Particularly, implementing a subsidy that targets these firms instead of equally distributing the subsidy between the three sectors achieves a more effective outcome.

Next steps

The next steps in this project involve two aspects. First, we create an indicator of efficiency for these policy combinations. We compare the cost of each of the scenarios to find the most efficient alternative to achieve the adoption of alternative fuel vehicles. Subsidies are compared with alternative policy instruments. Second, we generalize this approach to address the role of the public sector in other technology transitions such as the development and deployment of autonomous vehicles.

References

Encarnação Sara, Santos Fernando P., Santos Francisco C., Blass Vered, Pacheco Jorge M., & Portugali Juval. (2016). Paradigm shifts and the interplay between state, business and civil sectors. Royal Society Open Science, 3(12), 160753. https://doi.org/10.1098/rsos.160753

Howey, D. A. (2012). Policy: A challenging future for cars. Nature Climate Change, 2(1), 28–29. https://doi.org/10.1038/nclimate1336

ABSTRACT. The 40 years of reform and opening up is not only a process of major historical transformation of China's economy and society, but also a process of achieving a historic leap in China's scientific and technological innovation. After four decades of baptism in the tide of reform and opening up, China's science and technology innovation system has become an important part of the international science and technology innovation system. The "Several Opinions of the State Council on Comprehensively Strengthening Basic Scientific Research" promulgated in January, 2018 pointed out that a new round of scientific and technological revolutions and industrial reforms are booming, along with the accelerating scientific exploration, more closely integrated disciplines, and plenty of major breakthroughs being made in some basic scientific issues. The major developed countries in the world have generally enhanced the strategic deployment of basic research, and global scientific and technological competition has continuously moved forward to basic research. Basic research is a scientific research with a basic mission of profoundly understanding natural phenomena, revealing natural laws, acquiring new knowledge, new principles, new methods, and cultivating high-quality innovative talents.Basic research policies are an important part of national science policy, including policies that promote the development of basic research, and policies that governments use to develop basic research results for the country's overall goals.At the same time, as an important part of technological innovation, the scientific knowledge output of basic research has public goods attributes and a strong positive externality, which requires governments’ related policies and financial support . Thus, It is especially important to review and summarize China's basic research policies in the 40 years of reform and opening up, in which technological innovation has become the fundamental driving force for national transformation and development and the key node of competition among countries. Based on the recognition of the important position of basic research in China's science and technology innovation system, this study proposes the following considerations based on policy changes and inter-governmental cooperation: What are the variations of governments’ attentional allocations in basic research in different periods? What is the motivation for change? Which government departments are concerned about and support the development of basic research? What is the relationship between the departments? Therefore, this study is based on the basic research policy text data of China's central government level from 1978 to 2018. Through policy literature quantification, case analysis and expert interviews, it constructs a framework for policy change and inter-governmental relations analysis, and sorts out the changes in China's basic research policy since the reform and opening up. The purpose of this study includes summarizing the context of policy changes of basic research in China since the reform and opening up, identifying collaborative networks among government departments for basic research policy and providing reference for improving the scientific and rational basic research policy system, enhancing the strength of scientific and technological innovation and promoting the construction of a world science and technology power. This study is based on the basic research policy text data of China's central government from 1978 to 2018.This study divides the policy into the stage of reconstruction exploration (1978-1985), the initial stage of reform (1985-1992), the gradual recovery stage (1992-1998), the development acceleration stage (1998-2006), and the deepening reform stage ( 2006-2013), strategic development stage (2013-2018).Specifically, based on the division of policy stages, this study develops co-word analysis, cluster analysis and social network analysis from the perspective of “policy change” and “intergovernmental relations”. In conclusion, firstly, according to the division of stages, the number of basic research policies in China from 1978 to 2018 showed an upward trend. Secondly, through the classification and discrimination of policy keywords at different stages, the policy focus of China's basic research policy system has been shifted in different aspects during the 40 years of reform and opening up.Thirdly, in the past 40 years of reform and opening up, the number of departments involved in supporting the development of basic research in China has been from small to large. The departmental cooperation network has moved from “a dominant” to “multi-pronged”, from “divide-and-conquer strategy ” to “cooperation”.

ABSTRACT. Introduction

This poster presentation builds upon the work done in my PhD monograph (in progress). The first chapter estimates firm innovation strategies based on empirical observations. The second chapter estimates the relationship between firm innovativeness and public support. These are the preliminary results from the third chapter where shifts between firm innovation strategies are estimated together with direct public sector support to innovation.

Firm innovation strategies are based on firms’ choices in the innovation system: sources of information, partnerships, funding for different types of innovative activities, strategic goals, and appropriation methods used.

Firm innovation strategies are estimated based on two-step analysis: exploratory factor analysis and k-means clustering.

Results show that there are five distinct patterns of innovation: science based, open innovation, internal strategies, supplier based, and market oriented. In addition, a sixth category is available when firms indicated no innovative activities during the period.

Data comes from the Community Innovation Survey (CIS) in Estonia, a standardised innovation survey curated by Eurostat and carried out by Statistics Estonia. Altogether 9155 observations for 3502 unique firms between 2002 and 2012 is used.

External data about public support is culled from EU Structural Funds register and EU State Aid register, besides two main agencies supporting firms provided their full internal registers for the analysis.

Data about public support has been categorised into ten groups based on supported activities. These are collaboration programmes; consulting; training & skills development; marketing & export promotion; innovation and R&D support; investments support; mixed support; labour support; financial guarantees; and direct subsidies.

Main research question is following – are firms with certain public support more likely to shift their innovation strategies compared with previous periods?

The method is multinomial logit model with state dependency.

Background

Two different literature streams are relevant for this analysis. Firstly, firms’ innovation process and its dynamics. Secondly, public intervention to firms’ innovation process.

Firms’ innovation process and its dynamics expect the firm to balance between managerial capabilities and pressure from outside. The former consists of capabilities built within the firm and how they grow over time. These can include many possible paths, such as learning capabilities, organisation culture, ambitions, strategy creation, reaction to market forces, etc. The latter emphasises the surrounding environment which constrains the firm at the same time. Firms are bound by the technology, innovative potential and knowledge base available in their main activity. The influence creates industries which develop innovations in similar patterns.

Dynamics of this process is mainly developed based on products, which are narrower than firm-based theories. Most firms produce more than one product. Nevertheless, there are life cycles visible. The start of the life cycles is product innovations, development and finalising its properties. They are followed by process innovations, economies of scale and cost-efficiency.

Firms innovation process and choices about innovation strategies have many influencers. Some theories suggest that best behaviour should create path dependencies that are relevant for all within an industry. Some indicate that differentiation strategies are the key to long-term competitive advantages.

Innovation policy has been continuously emphasised for almost a century. In brief, the main idea is that there are suboptimal innovative activities without public intervention. The main outcome is, to increase innovativeness of firms, and therefore increase economic growth. First, second and third wave of innovation policies builds upon each other, widening the scope of activities and rationale for intervention. With this, the scope of policy instruments has also increased over time.

Methods

The first step was estimating firm innovation strategies with exploratory factor analysis and k-means clustering. This yields a single category for every firm in every period.

The second step was document analysis for all available public sector support given to the firms in the CIS sample. 146 policy instruments were categories into ten groups based on supported activities and policy rationale.

The third step is estimation with multinomial logit models with state dependency, firm controls and time controls.

Results

Preliminary results are currently available.

The models can be interpreted as such, coefficients estimate whether firms are more likely to shift from one innovation strategy to another if they receive public support. The coefficients are comparisons between one innovation strategy and base category.

The base category in all models is being not innovative (firms declared themselves not innovative in the survey during the period, therefore no innovation strategy).

Preliminary results indicate that there is variation between different types of public support and shifts between innovation strategies.

Firms are more likely to be science based compared to being not innovative when they receive innovation and R&D support even when their previous period innovation strategies are considered.

Firms are more likely to be supplier based compared to being not innovative when they receive investment support even when their previous period strategies are considered.

Firms are more likely to be open innovation compared to being not innovative when they receive training and skills support even when their previous period strategies are considered.

Main results show that there are significant differences between the uptake of different policy instruments when innovation strategies are taken into account. Firms with science based strategies, i.e. relationships with universities and R&D activities are more likely to participate in innovation and R&D instruments, similarly, with other categories and respective behaviour.

Results suggest that there is positive self-selection into policy instruments; firms acquire capabilities necessary to participate in certain instruments.

For some instrument types, there is no difference between innovation strategies and firms without any innovative activities.

Conclusions

Preliminary results suggest that there is some variation in innovative behaviour of firms and activities we wish to support with policy instruments. Causality is still an issue, but there are positive and significant correlations.

These results suggest that policy instruments might shift firms into behaviour that is intended. However, currently, it is difficult to estimate whether it is based purely on self-selection.

Results so far are also relevant for policymakers, that current financing has been at least somewhat useful, providing for firms who have necessary capabilities to undertake intended behaviour.

ABSTRACT. As China’s prominent weakness in the process of building a well-off society in an all-round way, the construction of ecological civilization and environmental protection have attracted wide attention from society and the academic community. In 1991, American economists Grossman and Krueger proposed the famous Environmental Kuznets Curve (EKC), pointing out that the pollution level increases with the increase of per capita income at the initial stage of development. However, In 1991, Michael Porter of Harvard University put forward the Porter Hypothesis (PH), which provides a theoretical possibility for seeking a win-win development model that takes into account both economic development and environmental protection. One of the important revelations from the PH is that innovation is an important way to achieve both development and protection at the same time. In this context, this study explores under what circumstances environmental regulation would promote the level of technological innovation in a region. The PH asserts that environmental regulation might promote enterprises’ technological innovation, thereby improving their competitiveness (Porter, 1991). Well-designed environmental regulation may promote enterprises’ technological innovation and stimulate the “innovation compensating” effect, which can not only make up for enterprises’ environmental compliance costs (Porter et. al., 1995).The PH has aroused heated discussions in the academic community. Palmer and other scholars believed that enterprises are rational entities pursuing maximized benefits. If enterprises could improve production efficiency and profit rate through technological innovation, then governments’ environmental regulation measures would be totally unnecessary (Palmer et. al., 1995). Generally speaking, the existing empirical studies mainly focus on three variants of the PH, namely the “weak,” “strong,” and “narrow” versions of the PH (Jaffe et. al., 1997).Western scholars carried out some extensive empirical studies(Jaffe, 1997 ; Brunnermeier et. al., 2003). In China, some analysis show that the relationship between environmental regulation intensity and the rate of enterprises’ technological progress presents a U-shaped curve. In the eastern and central regions(Zhang Cheng et al., 2011). Others analyzed panel data from 1999 to 2007. Results show that the “weak” version of the PH has been confirmed in the eastern region, but not in the central region (Wang Guoyin, 2011). Based on the above theories, the following two hypotheses are proposed: Hypothesis 1: After the technology intensity of a province reaches a certain threshold, environmental regulation will have an impact on technological innovation; Hypothesis 2: When the technology intensity reaches a certain threshold, the higher the intensity of environmental regulation, the more significant the impact of environmental regulation on technological innovation.In this study, 30 provinces in China were selected as the sample (data of Tibet Autonomous Region are largely insufficient and therefore excluded from the sample). All the data examined in this paper are from China's Science and Technology Statistics Yearbook, China's Environment Statistics Yearbook, and China's Industrial Economic Statistics Yearbook. Based on above theories, the output function of technological innovation in different regions is expressed as follows: Innovation=f(RDI, Personnel, ERI). This formula indicates that the level of innovative activities equals to a function of research and development investment intensity, the number of research and development personnel, and environmental regulation intensity.As the number of patents is a counting variable, simple OLS regression is not good enough to estimate the influence of each variables on dependent variables. Based on judgment, the unconditional variance of the sample dependent variable is approximately equal to the mean value. The panel Poisson regression model with fixed effect is used to estimate the equation.(PATENTS_(i,t) )=β_0+β_1 (RDI_(i,t) )+β_2 (Personnel_(i,t) )+β_3 (ERI_(i,t) )+β_4 〖(ERI_(i,t))〗^2+u_i+E_(i,t), Where, PATENTS_(i,t) represents the number of patent application in Province i in Year t; RDI_(i,t)represents the intensity of research and development investment in Province i in Year t; ERI_(i,t) represents the intensity of environmental regulation Province i in Year t. Personnel_(i,t) represents the full-time equivalent of research and development personnel in Province i in Year t. u_(i )represents the stochastic parameter of fixed effect. E_(i,t)is the residual. The results show that: (1) nationwide, the relationship between environmental regulation and the level of technological innovation presents a U-shaped curve. With the increase of environmental regulation, the level of technological innovation firstly decreases and then increases; (2) environmental regulation in the eastern region has a significant impact on technological innovation, while the impact in the central and western regions is not significant; and (3) when the technology intensity in a region (province) reaches a certain threshold, environmental regulation would have a significant impact on the level of technological innovation. The higher degree of technology intensity in a region (or province), the more significant the impact of environmental regulation on technological innovation.

ABSTRACT. A firm’s research publication activity cannot be easily understood from the theoretical lens of neoclassical economics, or even strategic management, mainly because firms wary of the possibility of imitation by competitors. Nevertheless, scholars in innovation studies have tried to suggest reasons why firms might make their research outcomes public, despite the high risk associated with this course of action. Notwithstanding the clear existence of advantages and disadvantages of firms’ publication activities, prior studies have usually focused on one of these, with little effort being made to provide a balanced perspective. This study aims to fill this gap in the literature by providing empirical evidence on factors influencing the potential advantages and disadvantages of firms’ publication activities, especially by focusing on emerging science-related technologies. In particular, this study relies on the concept of knowledge spillover, which has mainly had a negative connotation in the past; however, it has recently been noted that firms can learn vicariously from external knowledge that is shaped by knowledge spillovers from themselves. In this paper, the potential advantages and disadvantages of publishing a paper are defined with respect to the originating firms’ and external actors’ gaining of patents (i.e., proprietary knowledge) that are directly influenced by ideas within the knowledge spillover pool (i.e., papers citing the originating firm’s paper). Based on this definition, we investigate the dynamic influence of the evolving characteristics of knowledge spillover pool on the probabilities of patent generation by originating firm and external actors. Research questions are driven by the following three aspects that are still controversial in terms of the originating firms’ advantages and disadvantages: 1) the size of knowledge spillover pool; 2) the proportion of industry papers in the knowledge spillover pool; and 3) the similarity between the knowledge spillover pool and revealed knowledge. The research questions are examined using the data on publications and patents in artificial intelligence over the last ten years by 117 firms that have pursued publication and patenting in this emerging science-related technology area. By doing so, this study is expected to contribute to the ongoing discussion on the mixed effects of firms’ publication activity.

ABSTRACT. Science awards are highly institutionalized with many governments around the world operating a science award program to recognize or promote scientists’ performance. This paper addresses how the design of a scientific award system shapes the recipients’ impacts within the scientific community, focusing on the way the award system is structured to nurture the ability of early-career scientists rather than acknowledging the performance of established scientists. A scientific award is one of the central factors underlying the motivations of scientists to pursue and advance their research careers. A traditional form of the award has been a prize to honor the accumulated achievements of eminent scientists, as seen in the Nobel Prize or the Copley Medal created and managed by the private sector institutions. More recently, scientific awards are introduced and operated by governments as a tool to not only recognize renowned scientists but identify emerging scientists with great potentials. Especially latecomer countries in the global scientific scene – notably South Korea and China – have developed science awards targeting early-career scientists. It turns out that the performance of the recipients after the award as measured by standard indicators such as publications is generally higher for early-career awards than for traditional types of awards given to late-career scientists. Therefore, governments facing increasing pressure and competition with the rise of new public management appear to be strategically utilizing science awards as another mechanism to boost national scientific performance. While previous research has found the positive effects of science awards providing early recognition and status to young scientists on their future research productivity, relatively little attention has been paid to how those namely “strategic” awards targeting early-career scientists shape the recipients’ research behaviors and impacts on the scientific community. Our study aims to fill this gap by developing a theoretical framework to understand the operation of the strategic science awards based on the principal-agent theory and analyze the effects of such awards on extrinsic incentives based on the motivation theory. We further utilize the egocentric network analysis to examine the research network of the recipient before and after the bestowment. Our preliminary finding based on the data collected for the past eighty-three awardees of the Korean Young Scientist Award (given to scientists under the age of 40) since its inception in 1997 suggests that the research clusters of the recipient tend to be differentiated across heterogeneous groups after the bestowment. A more diversified network implies greater social capital of the awardee that can be used as a resource for enhancing research performance. Our study provides a new light on the science award as a differential mechanism in scientists’ collaboration network. Especially, our finding shows that the strategic award can function as an external incentive mechanism for early-career scientists to open a new window of opportunities to expand their research network dramatically. The study concludes with a few policy implication about possible outcomes expected from a specific award program to influence the behavior and performance of scientists.

ABSTRACT. Background and rationale - A development within the last century in scientific research has been the need for very large apparatus to explore new experimental fields, notably within high-energy physics. These ‘megascience projects’, which have a minimum budget of one billion US dollars, are generally undertaken as cooperative ventures by countries seeking to exploit scientific opportunities. Such projects are characterized by high levels of technological uncertainty, because success will likely depend on the development of new highly-advanced technologies. However, there is a notable lack of research into the leadership of megascience projects.

Methods and Research Questions - The projects investigated were the Tevatron at Fermilab, near Chicago IL, and the Large Hadron Collider (LHC) at CERN on the border between France and Switzerland. This research used a combination of archival and interview-based research to develop two case studies that answered three research questions: (1) What are the characteristics of those who lead megascience projects? (2) Where were their leadership skills developed? (3) How were their leadership skills developed?

Results and significance - The most important finding was the tailoring of senior leadership selection according to the needs of specific project phases. Four phases were identified: initiation, approval, construction, and exploitation. During the project there was a transition in senior leader characteristics from a transformational autocracy to an increasingly laissez-faire style. The characteristics of successful leaders of megascience projects at all organizational levels include 1) the primacy of technical competence, 2) strong management ability, 3) trustworthiness, and 4) team empowerment. This is somewhat unusual compared to other projects on this scale. The experiential nature of leadership training within megascience projects is also critical for success, with formal leadership training programs acting in a support role at most. This work also has implications for the next generation of megascience projects which is addressed as a conclusion.

ABSTRACT. Background and rationale Reliable information about doctoral candidates and doctorate holders are a critical factor for achieving high-confidence, evidence-based higher education and science policy decisions. Assessing the quality of training and the professional and personal determinants of highly educated talent for the academic and non-academic job market is essential to secure a leading position in the global education market and fulfill the demands of a strong innovation-based national economy. Young researchers are also expected to make considerable contributions to meet Germany’s and the European Union’s innovation and development targets. This generates interest from a diverse group of stakeholders from the national policy level down to individual universities that require data for their policy and governance making processes. In this presentation, we introduce and discuss a set of indicators which aims to deliver selected information about early career researchers in Germany. The National Academics Panel Study (nacaps) is a longitudinal study that generally surveys doctoral candidates and doctorate holders in about the study conditions for their doctorates, career aims and professional development, as well as their general life situation hence closing research blind spots and building the data basis for the associated indicator model. The first collection wave started in 2019 and will be updated in bi-yearly intervals. The model will prospectively be revised with future advancements of nacaps’ survey instruments. The set of indicators aims to inform and support stakeholders with differing information demands. The indicator-based display of data will be hosted on an interactive online data portal. Hence, three separate sites will be provided for the following user groups: 1) The German Federal Ministry of Education and Research as the contracting authority of the nacaps project 2) The general interested public, which includes current and future doctorate holders 3) The 56 participating universities, of which each one will receive an individualized website with (additional) personalized data. First advances in the design of indicators of early career researchers have already been made with the 2014 indicator model for the National Report on Junior Scholars, proposing a large variety of indicators for the populations of doctoral candidates and doctorate holders in Germany. Secondary sources of references include common indicator sets developed by the Federal Statistical Office, the OECD, the “UniWiND Coordination Office for Advanced Graduate Information (UniKoN)” and the “Research Core Dataset (KDSF)” as well as the nacaps predecessor, “Pro-File”. These indicator models offer a wide range of indicator sets but largely remain on a conceptual level. In addition, the nacaps indicator model focusses its development on a stable continuous panel data set, enabling a stakeholder-specific presentation. In the presentation, we will focus on the indicator model in light of its development process.

Methodological Approach The critical challenge in the design process of the indicator model was the harmonization of the information demands of the different stakeholders involved in the nacaps project. As a result, the indicator model is a product of both a deductive bottom-up, as well as a top-down development procedure.

Strict quality criteria including target orientation, repeatability, applicability, reliability, validity, and international comparability have been applied to the selection of each indicator.

Considering the variety of data sources and the mentioned quality requirements for an indicator model, we followed a “system modeling approach”, by focusing on the general and multidimensional presentation of the performance of the German doctorate system with respect to some important policy objectives (e.g. gender equality or reconciliation of family and professional career). A further selection stage involves the available set of survey constructs which are rated under the criteria of reference frame fit. Commensurability is a prerequisite for organizational competition. The frame fit requirement criterion enables us to provide benchmarks for individual universities, delivering further contextualization hence improving policy and governance making.

Results The result of the design process is a model which consists of circa 30 indicators covering a handful of topics. While there is substantial overlap of stakeholder demands, some indicators exclusively address specific stakeholders, e.g. universities display high demand for the evaluation of certain programs and courses whilst the BMBF be provided with additional information following an issue-centric approach.

The indicators are categorized into eight distinct topics related to the conditions of doctoral candidates and doctorate holders in Germany, i.e. 1) working and employment conditions 2) scientific supervision 3) type and structure of the doctorate 4) doctorate motives and career intentions 5) career paths and perspectives after the doctorate 6) international mobility 7) contribution to research, teaching and scientific transfer 8) private life situation and personal background

Significance The reported indicators are intended to improve the strategic management capabilities of universities on the one hand and will have a significant impact on the policy-making process of government actors on the other hand. Prospectively, the indicator model will be developed further, as the nacaps project proceeds with the launch of new collection waves, providing panel data. These yet to develop longitudinal indicators will ultimately provide the possibility to evaluate the effectivity of intermediate policy outcomes. This will add complexity, especially to the mode of demand coordination which so far has some to-be-addressed shortcomings in terms of structuredness. For discussion will be the indicator set draft with a focus on their categorization.

ABSTRACT. Manufacturing remains a critical force in both advanced and developing economies. The sector has changed over the years, bringing new opportunities and challenges to business leaders and policy makers. The degree to which countries are able to prepare for these changes will determine whether they thrive. The manufacturing sector plays an important role in the national system of innovation (NSI). The objective of the Next Industrial Revolution is to play a key role in enhancing South Africa’s manufacturing capacity so that it can compete in high-value products in global markets. Therefore, in the context of government efforts to ensure a growing role for science, technology and innovation (STI) in a more prosperous and inclusive society. This study aims at providing new insights regarding the pattern of innovation in the South African manufacturing sector using the results of the country’s three national innovation survey Data (2005, 2008 and 2012) in order to identify the main patterns of innovation in the manufacturing sector. Our study of the pattern of innovation in the manufacturing shows that progress has been made, with companies making visible efforts to engage in innovation activities. Enterprises with successful innovation observed to have similar pattern and innovation activity was positively related to innovation which arises from the application of technologies that were developed and are already in use elsewhere. The results reveals some weaknesses in the manufacturing sector innovation such as inadequate collaboration between firms and other firms, government or public research institutes as well as higher education. External collaboration can bring improvements to the production process of firms. In other words, there is a need for the manufacturing sector to engage in external collaboration during the innovation process to better improvement in their products and markets.

ABSTRACT. Public policy subsystems often involve a series of diverse types of actors from distinct territorial and governmental levels who play different roles and seek to have an incidence in order to fulfill their interests (Sabatier, 1993). In this process, actors interact and become interrelated, shaping policy networks, i.e. “stable patterns of social relations between interdependent actors, which take shape around policy problems and/or policy programs” (Klijn, Koppenjan, & Termeer, 1995, p. 439). The case of innovation policy and instruments is a relevant example of how a policy network can operate within a policy domain, since science, technology and innovation (STI) activities involve the collaboration, communication and interplay of different actors, particularly within innovation systems (Freeman, 1987; Lundvall, 1992; Nelson, 1993; Edquist, 1997). Innovation policy networks are built on the assumption that actors cannot keep pace with the changes in an innovation systems if they do not access to external sources of knowledge (Pyka, 2002).

Networks are a relevant feature in innovation studies regarding innovation systems (Weber & Truffer, 2017). Whether too weak or strong, network linkages arise as a systemic problem that can hamper innovation by hindering interaction among actors and, therefore, require attention by innovation policy (Carlsson & Jacobsson, 1997; Woolthuis, Lankhuizen, & Gilsing, 2005). Networks are a relevant issue for innovation policies in both mature and emerging innovation systems. According to Chaminade, et, al. (2009), on one hand, network problems in a mature innovation system of a developed country can be the “lack of dense interfirm networks” as well as “weak university-industry research networks” (p. 372). On the other hand, the main network problem in emerging innovation systems of developing countries like Colombia is weak linkages in firms (indigenous–multinational), market agents (customers–producers), and local needs (universities­–industry) (Chaminade, Lundvall, Vang, & Joseph, 2009).

Innovation policy networks in Colombia have been studied from a governance point of view that considers different levels of analysis: strategic, programmatic and social levels/networks (Orozco, et al., 2015; Pantoja & Moreno, 2018; Orozco, Villaveces, Ordonez-Matamoros, & Moreno, Forthcoming). The ‘strategic’ or regulatory policy network refers to the instances and councils that have the legal mandate to define innovation policy orientations; the ‘programmatic’ policy network is composed by those organizations that participate in the execution of actions for the implementation of innovation policies and programs; and, finally, the ‘social’ level have been regarded as the virtual space for deliberation on innovation policy issues, including public agencies and various actors, shaping a governance network (Orozco, et al., 2015; Pantoja & Moreno, 2018; Orozco, Villaveces, Ordonez-Matamoros, & Moreno, Forthcoming).

The literature on innovation policy recognizes the existence of a variety of actors involved in policy within an innovation systems, particularly from a functional point of view (Bergek, Jacobsson, Carlsson, Lindmark, & Rickne, 2008). However, it has focused mainly on those leading actors and organizations that have a visible role in R&D and innovation activities, and neglected those that have a less visible role by not being stricto sensu STI actors/organizations but are relevant in sustaining the enabling environment for STI activities. Here, STI is regarded as a ‘practice field’ that involves those actors with a leading role in R&D and innovation activities; sustained by ‘practice work’ repertoires, i.e. “those arrays of activity enacting, making possible, sustaining on time and shaping the rationale and values of a practice field” (Ramos-Mejía & Balanzó, 2018, p. 5). This enabling activities are carried out by actors that have been understudied by innovation policy literature.

In this regard, this chapter examines the policy network shaped by a particular STI policy instrument in Colombia: the royalties fund for STI in Colombia, implemented with the purpose of increasing STI funding at the territorial level and, therefore, to strengthen the capabilities of territorial governments in STI activities. We argue that between 2012 and 2018 this particular policy instrument has led to the configuration of a policy network at an operational level, involving a variety of actors that do not specialize in R&D or innovation activities, but enable the operationalization of STI by mobilizing resources at the territorial level. What are the main features of this operational policy network? How is it different from a traditional policy network? Who are the most relevant organizations in this operational policy network? What is the role of non-STI organizations in this operational policy network? These are the research questions that this chapter addresses.

Drawing inspiration from the work of Orozco et, al. (2015; Forthcoming), the research method here is a social network analysis that focuses on an ‘operational policy network’ composed by actors involved in royalties-funded STI projects (nodes) and the contracts (edges) that relate them to each other in the implementation of projects. Centrality metrics like degree and betweenness are analyzed in order to give an account of the dynamics and characteristics of the network. An operational policy network approach is built regarding innovation systems approach, policy instruments literature, policy networks approach and functions of innovation systems.

One of the main arguments for the recent 2018 reform to the royalties for STI funding scheme was that political not-STI-specialized actors were responsible for the projects. However, we find that STI actors like universities were the most influential at the operational level given their connection to the rest of the organizations that participated in the mechanism. This universities even have the potential of ‘systemic intermediaries’ (van Lente, Hekkert, Smits, & van Waveren, 2003) since they are the ones that keep the network together through intermediation. Also, non-STI actors are usually connected to the network by one single tie, which is a characteristic of their role as hard and soft intermediaries (van Lente, Hekkert, Smits, & van Waveren, 2003) that are important in resources mobilization (Hekkert, Suurs, Negro, Kuhlmann, & Smits, 2007) which is in turn ‘practice work’ to sustain the particular ‘practice field’ of STI policy at the operational level (Ramos-Mejía & Balanzó, 2018). This practice work therefore consists in mediating between their own source of knowledge and their clients in providing resources, technical services and management support.

We argue that operational policy networks emerge out of narrow policy instruments implemented at the micro or operational level, with the participation of a high diversity of actors that fulfill systemic functions like entrepreneurial experimentation, development of positive external economies, resource mobilization. This actors are mainly territorial public universities, R&D firms, SMEs, support / supply firms, civil society organizations, local entrepreneurs, among others. They interactions are mainly practice-based: project execution, procurement, knowledge, goods and services supply and resources mobilization. This make operatoinal policy networks different from broader policy networks regarding their source, level, scope, type of actors involved, the interactions between the actors in the network and the governance ‘dancing partners’ that drive the interactions.

References

Bergek, A., Jacobsson, S., Carlsson, B., Lindmark, S., & Rickne, A. (2008). Analyzing the functional dynamics of technological innovation systems: A scheme of analysis. Research Policy, 37(3), 407-429. Obtenido de https://www.diva-portal.org/smash/get/diva2:267496/FULLTEXT01.pdf

Carlsson, B., & Jacobsson, S. (1997). In Search of Useful Public Policies — Key Lessons and Issues for Policy Makers. En B. Carlsson (Ed.), Technological Systems and Industrial Dynamics. Economics of Science, Technology and Innovation (págs. 299-315). Springer.

Chaminade, C., Lundvall, B.-Å., Vang, J., & Joseph, K. J. (2009). Designing innovation policies for development: towards a systemic experimentation-based approach. In B.-Å. Lundvall, K. J. Joseph, C. Chaminade, & J. Vang (Eds.), Handbook of Innovation Systems and Developing Countries. Building Domestic Capabilities in a Global Setting (pp. 360-387). Edward Elgar.

Edquist, C. (1997). Systems of Innovation. Technologies, Institutions and Organizations. Londres: Pinter.

Freeman, C. (1987). Technology and Economic Performance: Lessons from Japan. Londres: Frances Printer Publishers.

Hekkert, M., Suurs, R., Negro, S., Kuhlmann, S., & Smits, R. (2007). Functions of innovation systems: A new approach for analysing technological change. Technological Forecasting and Social Change, 74(4), 413-432.

Howlett, M., Mukherjee, I., & Woo, J. J. (2018). Chapter 9: Thirty years of research on policy instruments. In H. Colebatch, & R. Hoppe, Handbook on Policy, Process and Governing (pp. 147–168). Edward Elgar Publishing.

Klijn, E.-H., & Koppenjan, J. (2000). Public management and policy networks. Foundations of a network approach to governance. Public Management: an international journal of research and theory, 2(2), 135–158.

Klijn, E.-H., Koppenjan, J., & Termeer, K. (1995). Managing networks in the public sector: A theoretical study of management strategies in policy networks. Public Administration, 73, 437-454.

Lascoumes, P., & Le Galès, P. (2016). Instrumento. In L. Boussaguet, S. Jacquot, P. Ravinet, J. I. Cuervo, J.-F. Jolly, & D. Soto Uribe (Eds.), Diccionario de Políticas Públicas (A. C. González, J.-F. Jolly, V. Herrán Ocampo, D. Soto Uribe, C. Soto, C. Isaza, . . . J. Ayarza, Trans., 2 ed., pp. 342-350). Bogotá: Universidad Externado de Colombia.

Lundvall, B.-Å. (1992). National Systems of Innovation: Towards a Theory of Innovation and Interactive Learning. Londres: Pinter.

Nelson, R. (1993). National innovation systems: a comparative analysis. Oxford: Oxford University Press.

Orozco, L. A., Cancino, R., García, M., Moreno, G., Petit-Breuilh, J., Goñi, J., . . . Ordóñez-Matamoros, G. (2015). Redes de política y gobernanza de los Sistemas Nacionales de Innovación: una comparación entre Chile y Colombia. In Mirada iberoamericana a las políticas de ciencia, tecnología e innovación: perspectivas comparadas (pp. 221-258). Buenos Aires; Madrid: CLACSO; CYTED.

Orozco, L. A., Villaveces, J. L., Ordonez-Matamoros, G., & Moreno, G. (Forthcoming). Innovation policy and governance networks on national innovation systems. Retrieved from https://www.academia.edu/38972783/Innovation_policy_and_governance_networks_on_national_innovation_systems?email_work_card=view-paper

Pantoja, L. M., & Moreno, L. G. (2018). Evolución de la red de política para la innovación en Colombia: el fenómeno emergente de conformación de redes de gobernanza mediante el análisis de redes sociales (Twitter). Praxis Sociológica, 23, 65-87.

Pyka, A. (2002). Innovation networks in economics: From the incentive-based to the knowledge-based approaches. European Journal of Innovation Management, 5(3), 152-163.

Ramos-Mejía, M., & Balanzó, A. (2018). What It Takes to Lead Sustainability Transitions from the Bottom-Up: Strategic Interactions of Grassroots Ecopreneurs. Sustainability, 10(7), 2294.

Sabatier, P. (1993). Policy Change over a Decade or More. En P. Sabatier, & H. Jenkins-Smith (Edits.), Policy Change and Learning: An Advocacy Coalition Approach (págs. 13-39). Westview Press.

Scharpf, F. W. (1978). Interorganizational policy studies: Issues, concepts and perspectives. En K. Hanf, & F. W. Scharpf (Edits.), Interorganizational Policy Making: Limits to Coordination and Central Control (págs. 345–370). London: Sage.

van Lente, H., Hekkert, M., Smits, R., & van Waveren, B. (2003). Roles of Systemic Intermediaries in transition processes. International Journal of Innovation Management, 7(3), 247–279.

Weber, M., & Truffer, B. (2017). Moving innovation systems research to the next level: towards an integrative agenda. Oxford Review of Economic Policy, 33(1), 101–121.

Woolthuis, R. K., Lankhuizen, M., & Gilsing, V. (2005). A system failure framework for innovation policy design. Technovation, 25, 609–619.

ABSTRACT. Non-R&D intensive firms are defined in the innovation literature as firms that lack substantial investments in research and experimental development (R&D). In developing countries, over the last few years, non-R&D-intensive firms have played important roles in their respective sectors’ value chains and have contributed to economic growth. Despite their growing contribution to innovation outputs, non-R&D intensive firms have largely been relegated to the bottom of the industrial value chain. This is evidenced by a large majority of innovation studies that have predominantly focused on R&D intensive firms especially in high-tech industries.

Considering the growing importance of external sources of knowledge for innovation, the ability to absorb, assimilate and replicate new knowledge acquired from external sources has become essential for firms to innovate. This ability is referred to as the absorptive capacity. Prior research has shown that R&D intensive firms tend to possess significant stock of cumulative knowledge resulting from their R&D investments. As a result, these firms are more likely to have a higher absorptive capacity, which in turn benefit their innovative capabilities.

Contrary to early innovation studies that have considered R&D investments as a perquisite for innovation, recent studies have shown that firms do not necessarily need to invest in R&D to become innovative. This implies that non-R&D intensive firms have an absorptive capacity that contributes to their innovation capabilities. As such, this study aims to explore how the absorptive capacity of non-R&D intensive firms influences their innovation performance.

The study develops an econometric model that relates absorptive capacity of non-R&D intensive firms to innovation performance while controlling for appropriate factors. The developed model is tested using firm level data of non-R&D intensive firms in South Africa’s manufacturing and services sectors to explore how the absorptive capacity of these firms facilitates their innovation output (i.e. development of products and services).

ABSTRACT. Basic research performed by universities is one of the most important drivers of technological development and economic competiveness. To maintain their competitive edge, especially industrialized countries have implemented novel university research funding programs during the last decades. Aiming at selectively developing internationally leading universities within Germany, the Excellence Initiative was launched by the German federal government in 2006. The initiative added 4.6 billion euro to the German university system from 2006 to 2017, making up four percent of its total research spending (IEKE, 2016). Furthermore, it was subsequently used as a role model for funding programs in various other countries, such as France, Japan, Malaysia or Spain (DFG, 2015).

To spur technological development, scientific knowledge generated by universities needs to be transferred to the private sector. R&D cooperation is one of the most intensive knowledge transfer channels. Research on the impact of public university research funding on university-firm R&D cooperation, however, is still rare, especially compared to the amount of literature on other transfer channels, such as scientific publications and patents by universities.

Firms typically participate in R&D cooperation with universities to capitalize on their tacit knowledge and already built equipment. Universities usually cooperate with firms to raise additional funds for their research activities or to complement their research with relevant findings gathered throughout R&D cooperation. The literature on the impact of public university research funding on R&D cooperation between firms and universities identified two opposing funding effects. On the one hand, public funding increases the attractiveness of the funded universities to private firms. Funded universities have an increased ability to employ high-performing scientists and to purchase equipment or material. Also, their improved reputation fosters the visibility and scientific legitimacy of joint R&D cooperation, independently of their actual research capabilities. On the other hand, public funding might decrease the incentive of universities to cooperate with private firms. Public funding fosters the financial independence of university scientists and decreases their need to acquire additional third-party funding from the private sector. Funded scientists are thus able to allocate their resources more freely according to their research interests. If their research does not profit from R&D cooperation with private firms, receiving public funding will decrease their incentives to participate in them. The net effect of public funding on university-firm R&D cooperation, hence, depends on the magnitude of the opposing effects.

I investigate the net effect of additional public university research funding provided by the Excellence Initiative on local university-firm R&D cooperation and, thus, on knowledge transfer. The paper takes a firm-level perspective and analyses a firm’s probability of participating in an R&D cooperation with its local universities. Its focus is on the change in a firm’s likelihood to cooperate if its closest university was awarded research funding by the Excellence Initiative.

This paper contributes to the literature on university-firm R&D cooperation by using research grants from the Excellence Initiative as quasi-experiment to overcome the endogeneity of public funding in a difference-in-differences setup. Public funding is typically allocated to universities employing high-performing scientists. This selection process generates a bias, as universities employing high-performing scientists are also the ones most attractive to cooperate with in the first place. The difference-in-differences analysis allows to control for this selection by simply comparing the pre-treatment trend of the treatment and the control group. Only Blume-Khout et al. (2015) and Lanaham et al. (2016) tackled endogeneity issues in the related literature, but with different instrumental variable strategies, whose identification assumptions are harder to verify.

Besides its main contribution of adding quasi-experimental evidence, this paper is the first to investigate changes in the composition of R&D cooperation. Former research only empathized on the aggregated amount of R&D cooperation. However, an increased attractiveness of funded local universities could trigger additional R&D cooperation between local firms and universities, but it could also cause firms to substitute non-local cooperation in favour of local ones. To examine the composition of local and non-local R&D cooperation, this paper also studies the impact of public funding on the probability of cooperating with non-local universities.

Differentiating between public funding in engineering, life and natural sciences further adds to the existing research. The theoretical model of Jensen et al. (2010) suggests research funding in applied science fields to foster university-firm R&D cooperation, while funding in more basic science fields rather reduces cooperation. The rational: research in more basic science fields profits less from cooperating with firms and, hence, university scientists in these fields are more likely to stop cooperating after receiving additional public research funding and being able to allocate their resources more freely. Empirically, however, Lanaham et al. (2016) are the only ones directly considering heterogeneous effects.

Finally, the analysis is highly relevant for German and international policy makers. The German Excellence Initiative is representative for the design of several newly implemented university funding programs in different industrial countries. This paper is the first considering its impact on the private sector. It adds to the evidence on the net impact of public university funding on R&D cooperation and on knowledge transfer. Finding a negative net effect would especially challenge the established assumption of public university funding being a simple tool to foster regional technological development and economic competitiveness. Also, the same holds for getting to know that additional funding only changes the composition of R&D cooperation instead of generating new cooperation and more knowledge transfer.

The firm-level outcomes of the difference-in-differences estimations confirm the results of prior research conducted on the country-, university-, department-, or researcher-level: The net effect of public university research funding on R&D cooperation is positive. Firms whose closest university received a research grant from the Excellence Initiative are more likely to participate in an R&D cooperation with their local universities. The impact of public funding, however, varies with the funded science field. Following the empirical results from Lanaham et al. (2016) and the theoretical model of Jensen (2010), the outcomes indicate the most pronounced effect for funding allocated to engineering sciences, the most applied science field. Funding in engineering sciences consistently increases a firm’s average cooperation probability by three to seven percentage points, whereas no consistent impact across all estimations has been found for funding in life or natural sciences.

Indicating the causality of the estimations, I find a common pre-treatment trend in all robustness checks. Moreover, cooperation with non-local universities remain unaffected by the Excellence Initiative. Therefore, public university research funding distributed as part of Excellence Initiative in engineering sciences seems to trigger additional R&D cooperation instead of only changing the composition of local and non-local cooperation.

References

Blume-Kohout, Margaret E.; Kumar, Krishna B. and Sood, Neeraj (2015), “University R&D Funding Strategies in a Changing Federal Funding Environment”, Science and Public Policy, 42, 355-368.

DFG (2015), “Bericht der Gemeinsamen Kommission zur Exzellenzinitiative an die Gemeinsame Wissenschaftskonferenz”, Deutsche Forschungsgemeinschaft DFG, Bonn (Germany).

IEKE (2016), “Endbericht der Internationalen Expertenkommission zur Evaluation der Exzellenzinitative”, International Expert Commission to Evaluate the Excellence Initiative IEKE, Berlin (Germany).

Jensen, Richard; Thursby, Jerry and Thursby, Marie C. (2010), “University-Industry Spillovers, Government Funding, and Industrial Consulting”, NBER Working Paper #15732.

Lanaham, Lauren; Graddy-Reed, Alexandra and Feldman, Maryann P. (2016), “The Domino Effects of Federal Research Funding”, PLoS ONE, 11(6), e0157325. doi:10.1371/journal.pone.0157325.

ABSTRACT. Scientific and technological (abbr. S&T) achievements and the transformation of S&T achievements are the special terms of government in China. In the Law of the PRC on Promoting the Transformation of Scientific and Technological Achievements, it is said that S&T achievements refer to those applicable ones produced through scientific research and technological development, while transformation of S&T achievements means the entire process of the follow-up tests, development, application and widespread use of the applicable S&T achievements，through to the final creation of new technology, new techniques, new materials, new products and new industries -- all for the purpose of enhancing the productive forces. The definition in this law presupposes a linear process of the transformation of S&T achievements, which is “produce achievements before transformation”. It is a mistake to regard the transformation of S&T achievements as a linear process, and it stands in the way of solution to the two-skin problem between technology and economy. Due to the linear mode, innovation policies are made with a bias to the supply side, and it is believed that if the output of achievements is increased, the transformation of achievements can be increased. As a result, a lot of incentive policies focus on increasing the number of patents, and motivating the enthusiasm of univeristies, research insitutions and their researchers, such as “delegating the disposition right, the right of use and the usufruct of the S&T achievements”, boosting the proportion of rewards, and allowing researchers to start businesses without resigning. Actually, in most cases, these policy interventions only bring the activities of “treading a fine line” and “walking in the legal grey zone” into the track of standardization and institutionalization, but without significant improvement on the overall data of transformation of S&T achievements at the macro level of China. Under the guidance of this kind of incentives, the transformation efficiency of the S&T achievements has not been significantly improved, and the reason is attributed to the fact that universities and research institution do not produce good and valuable achievement. As a result, there is a point of view that transformation of S&T achievements is a false proposition. In fact, this is a double misunderstanding of S&T achievements and transformation. With the linear mode, S&T achievements are mostly limited to those technologies in the form of patents, trademarks, licenses, etc., which can be directly used for transactions and quickly transformed into products. In addition, with the narrow-minded understanding of viewing S&T achievements as technology, the local government promote the transformation of S&T achievements limited to “gain technology and gain projects”. The criteria to assess the success of the transformation of S&T achievements is whether the project is implemented, whether the enterprise is established, whether new products and new technologies are generated, etc.. For instance, the mayor of Wuhan city, Hubei Province, proposed in 2017 that “local transformation propotion of S&T achievements in universities and research insitutions in Wuhan is supposed to account for 80% within next five years”. As innovation grows more networked and goes digital, the concept of S&T achievements has become more extensive. It not only includes encoded patents, technologies, papers, licenses and others explicit knowledge, but also includes tacit knowledge, skills, know-how and other non-coded achievements. Therefore, the forms of transformations of S&T achievements are also more and more diverse. It has two types of transformation. One is so-called “contract-type”, including contract research, collaborated research, intellectual property rights transaction, spin-off, etc.. The other is so-called “collabrated-type”, including training, communication, student internship, staff mobolity and so on. Usually, the second is more important. It is not difficult to find that knowledge flow and achievement transformation are carried out through learning, communications and exchanges each other among these organizations in most cases. The crossing network of knowledge flow among universities, research institutions, enterprises and intermediary institutions, as well as the dynamic relationship network interwoven with the situation composed of policies, institutions and markets, are the network model of the transformation of S&Tachievements. Although this concept is proposed for the first time, the network model of transformation of S&T achievements already exists. Translational medicine is an innovative concept that emphasizes the combination of theory and practice and accelerates the transformation of basic research into clinical application. The United States, the United Kingdom, Japan, Singapore, China, the European Union and other countries and regions have established a number of translational medicine centers, which connect the government, universities, hospitals, enterprises and other institutions, forming an interdisciplinary, multidisciplinary and integrated development of a wide range of collaborative network. The goal of building this network is not just to turn technology into drugs or devices, but to focus more on how to facilitate the flow of knowledge across the network, thereby promoting overall innovation efficiency and disease treatment. At the same time, translational medicine research also provides important support for countries to improve public health policies. The Chinese government has also started to consider the support policies for innovation from a systematic perspective, such as promoting the collaborative innovation system and the construction of the industry-university-research cooperation network. However, the specific operation is still in a linear mindset, and the focus of most policies is still projects. So, generally speaking, the end of the project period means the end of the cooperation. In the future, we should re-recognize the transformation of S&T achievements, uphold the concept of network model in policy making and work, strengthen the connection between the government, universities, scientific research institutions, enterprises, intermediaries and other subjects, establish the corresponding communication mechanism, and promote the flow of knowledge in the network.

ABSTRACT. Introduction The Strategic Technology & Innovation Management (STIM) Consortium is an engagement platform, established in 2013 by the Institute of Manufacturing of the Cambridge University, to develop and implement innovative approaches between academics and industries sectors. The purpose of this study is to show how STIM Consortium offers a broad and practical approach to academic engagement (AE) activities creating a unique model of University-Industry collaboration. STIM developed an Open Innovation Engagement platform for the knowledge sharing in which researchers propose to companies the themes of the engagement. This uniqueness makes STIM a novel model of AE and an exclusive tool for public-private engagement research driven.

Research Methodology We gather information from companies and scholars involved in the past STIM programs (2013-2018). The study uses a qualitative content analysis methodology (Eisenhardt, 1989; Yin, 1994) for the examination of a phenomenon in its natural environment. The method is appropriate for understanding the "how" and "why" of underexplored phenomena. Primary data were collected during semi-structured interviews to bring out contents without influencing the interviewee; the interviews were conducted in person and lasted on average one hour (min. 30 and max. 90 minutes). Secondary data about activities and companies were collected from official reports and databases. We run two rounds of interviews, the first concerned companies and the second the scholars. In setting the interviews for defining of the AE activities, we followed the Four Central Measures approach (Perkmann et al., 2013; Vick and Robertson, 2018) that identifies the most relevant aspects of the engagement: 1. Activities, 2. Motivations, 3. Barriers, 4. Outcomes. We assumed events happened following the generic AE model (Perkmann et al., 2013) that counts five steps: identification, initiation, coordination, execution, and evaluation.

Results STIM Consortium is a peculiar model and a unique experiment of open innovation AE platform for its academic relevance, the participation of multiple stakeholders from research, education, and industry, and its focus on knowledge exchange mechanisms from academia to industry. 1. Activities: STIM is a “platform/marketplace” and “self-organizing design”. With its low subscription policy, STIM manages companies’ expectations fitting the nature of industrial research need and offering a “knowledge buffet”. Scholars propose research projects, and managers engage those they like the most. 2. Motivations: Scholars and managers engage for very different reasons, the most important is the self-interest. Managers’ reasons are corporate social responsibility, talent scouting, problem-solving, or networking. Scholars’ reasons are research projects funding, intellectual property development, or executive educational engagement. 3. Barriers: a. Trust: STIM does not promote short-term work relations. Its added value lies in the fact that managers and scholars build long-term relations in which both can better meet the reciprocal demands. b. Firms’ Absorptive Capacity: Managers join meetings as a “knowledge buffet” seeking inspiration, but the lack of the necessary absorptive capacity might lead them to not adequate projects wasting time, resources and opportunities. 4. Outcomes: Identifying the right metrics to determine the level of engagement is still an ongoing activity. First results suggest that a useful indicator for companies is the length of membership and their propensity to renew it. It is crucial to measure the general satisfaction level that companies experience in joining STIM because relevant engagement activities take place after the meetings where scholars and managers get involved in more-in-depth, parallel research activities.

Implications of our analysis are significant both at theoretical and managerial level. On the first side, we identified a new path in the Open Innovation paradigm that goes from university to industry. Scholars propose what research activity is attractive to develop and industries join them. On the managerial aspect, it is essential for firms this sort of AE to ease the exploration phase, to identify new research trend and to evaluate new opportunities. A long-term goal of STIM’s AE Open Innovation platform is to extend the application of the model to other universities.

Limitations At the present the major limitation is that the study will be ongoing until June 2019, we count to provide a definitive draft of the complete article at the end of August. Currently, we are analyzing the parallel activities developed by scholars and companies. This is helping us in defining STIM as a new model of engagement. The ultimate goal of the study is presenting a new format of engagement. We are working to complete the benchmarking study here described looking at the full project development to identify the novel approach and to strengthen the theories about the AE. This will enable STIM to be considered an innovative and unique experience. For benchmarking the activities, we are also seeking both to improve the existing assessment measures and to find new metrics.

References Eisenhardt, K.M., 1989. Building Theories from Case Study Research. Acad. Manag. Rev. 14, 532–550. Perkmann, M., Tartari, V., McKelvey, M., Autio, E., Broström, A., D’Este, P., Fini, R., Geuna, A., Grimaldi, R., Hughes, A., Krabel, S., Kitson, M., Llerena, P., Lissoni, F., Salter, A., Sobrero, M., 2013. Academic engagement and commercialisation: A review of the literature on university–industry relations. Res. Policy 42, 423–442. Vick, T.E., Robertson, M., 2018. A systematic literature review of UK university–industry collaboration for knowledge transfer: A future research agenda. Sci. Public Policy 45, 579–590. Yin, R.K., 1994. Discovering the Future of the Case Study. Method in Evaluation Research. Am. J. Eval. 15, 283–290.

ABSTRACT. Background The economies aiming to develop a solid scientific labour force find in international mobility a way to enrich the creation of research capacities by sending nationals to study abroad. Mexico has been funding doctoral training in foreign universities for over four decades. However, different from the Asian cases, which accumulated significant R&D capacities through their internationally trained people, Mexico’s scientific workforce is still small and faces emigration of the most talented. The main assumption in this context is that the policy needs to be implemented in a more collaborative manner with national actors, namely implied employers. This is because if the policy is not reflecting its expected effects, this may be due to the need to aligned this policy to national needs. In this context, the main motivation of this research is to explore how the programme for doctoral formation corresponds, or not, to the needs of the employers for doctorate holders. This study looks into the involvement of the different actors, and how these actors perceive the value of international mobility in doctoral training and the effects of the interaction between actors in the absorption of doctorates and creation of research capabilities. This research focuses on the programme by which Mexico funds its citizens to pursue doctoral education abroad. The nanotechnology sector is used as an empirical case in order to provide more in-depth details on how such policy delivers the expected outcomes in the system.

Doctoral training in S&T is key in the creation and consolidation of the scientific and technological capacities in a country. Governments consider doctoral training as a policy priority, and although the rationales for sending people to pursue doctoral training abroad may differ across countries, there are key motivations for which developing countries invest public funds into sending nationals to study abroad, namely learn the latest techniques, and research problems that the global scientific community is addressing. The number of doctoral students from developing going to a different country –most commonly a developed country to pursue doctoral education has risen in the past decades and continues to show an increasing trend, which indicates the interest of governments to incorporate international mobility as essential in doctoral training. Mexico is one of the countries that have been relying on international mobility for doctoral training. The formation of highly specialised people, namely doctorate education in STEM fields is a national priority as stated in the S&T policy documents and programmes of the last three decades. The design, implementation and funding of the doctorates formation strategy in Mexico are led by the Mexican Council of Science and Technology (Conacyt). Additional funds come from other national and international sources but most are allocated through Conacyt for its implementation. Doctoral training in S&T is key in the creation and consolidation of the scientific and technological capacities in a country. Doctorates represent the future scientific labour force, they embody the technical knowledge and talents by which the existing knowledge is renovated and new ideas are developed. The benefits of exploiting the human capital embodied upon the highly specialised/doctorates, and also the disadvantages that the shortage of doctorates generate have been widely studied. In doctoral formation international mobility represents a human capital formation mechanisms that enrich the skills of those that study abroad, creating at the same time new career development opportunities. For the sender countries, international mobility for doctoral training is an opportunity to learn and to be present in the global research community.

Methods This study explores the Mexican strategy for doctoral training abroad by a set of research questions that aim to gain some understanding of the rationales for doctoral formation between the different actors in the national system. The overarching question is how the policy maker’s rationale for doctoral formation corresponds, or not to the needs of the employers for doctorate holders. Additional questions look into the involvement of the different actors in the S&T policy –of which doctoral formation is part, design, funding and implementation, and how they perceive the value of international mobility in doctoral training, and the strategies they have developed to capitalise from the capital embodied in doctoral holders after return.

The questions of this study are addressed by adopting a qualitative approach framed under a diverse set of literature, namely innovation studies, human capital formation theory, innovation systems theory and migration studies. An essential assumption that emerges from these sets of theories is that doctoral training is a key element for the development of technological capabilities and innovation, but that the inefficacies in the domestic conditions to absorb the doctorates after completion, and further their career development promotes long term emigration or the under-capitalisation of the capital built abroad during doctoral training. Thus, the aim of this study is to provide in-depth understanding about the rationales for sending nationals to study abroad, and about how the interactions and level of involvement of different actors in the system affect the capacities and conditions by which doctoral training can be translated into research and innovation capacities in Mexico. In order to explore this assumption and provide answers to the research question in this study, this research focuses on the policy programme by which Mexico funds its citizens to pursue doctoral education abroad. The scope of the research focuses on the actors in the nano sector. In the first data collection stage of this work, interviews were conducted with the policymakers responsible for that programme; with higher education institutions and company as implied sectors that value and demand doctorates. This data collection yielded 52 interviews in total. To capture the complexity of the phenomenon of doctoral training and international mobility as a human capital formation mechanism, in a second data collection stage, a qualitative online survey was conducted for Mexican nationals who pursue doctoral education in a foreign university funded by the Mexican government. Having an analytical framework the concepts of capitalisation of doctoral training and policy involvement data was analysed under a systemic approach. By bringing together the data from diverse actors, this work contributes to understanding how the valorisation and influence of doctoral formation policy through international mobility differs between actors in the system, and how the dissimilar interest and lack of cohesion of rationales triggers long term migration of doctorates and underdevelopment of those that stay, hindering the creation of conditions to foster the capitalisation of doctorates into domestic research capacities.

Findings

Findings suggest that the programme furthers opportunities for the doctorates sent abroad, but that those opportunities are not capitalised in the national arena, thus remaining only as individual benefits. While on the side of the employer, findings suggest that when actors do not see their interests reflected in policies, their involvement is limited, affecting the possibilities for capitalisation of doctorates, which promotes long-term emigration or the under-capitalisation of the human capital built abroad. The implications of these findings are that S&T policy might reconsider the focus on international mobility and that governments must promote the participation of actors in the policy process in order to develop strategies where they share a common interest.

ABSTRACT. Aim of the proposal.  In this study, we intend to investigate two research questions: How do subsidy policy affect the allocation of R&D resources? Should resources be allocated towards lowering the cost of R&D or broadly lowering distortionary barriers?

Background: The subsidies policies have caused controversy among both policy-makers and scholars on whether they are effective in stimulating innovation to what degree government should intervene the innovative behavior of firms. A major challenge facing late-comer economies, including China is how to distribute the innovation resources efficiently.

Studies on the effectiveness of R&D subsidies tend to explore its impacts on R&D investment, namely, crowding out effects and crowding in effects. However, evaluating the effects only on innovation inputs provides a partial assessment of the impact of the incentives because the goal is to increase firms innovative capabilities and ultimately, more innovative economy as a whole (Bronzini and Piselli, 2016). Wei et al. (2017) argue that targeted R&D subsidies and industrial policies can increase aggregate R&D but possibly induce the wrong firms to invest in R&D. Misallocation of innovation resources, especially between state-owned and private firms causes the inferior performance of Chinese firm in indigenous innovation. Studies have documented that subsidies policy in China is biased towards certain firm characteristics such as ownership, size, and sectors (Brandt et al., 2013; Boeing, 2016). However, the potential results caused by such bias is under-explored.

Furthermore, previous studies identify distortion of production resources across heterogeneous firms is one major cause for productivity losses and for the persistent differences in productivity across countries (Hsieh and Klenow, 2009; Banerjee and Moll, 2010; Ranasinghe, 2014). Misallocation of resources are potentially relevant to slow diffusion of frontier technologies to less-developed countries (Restuccia and Rogerson, 2017). However, the cause of misallocation in innovation resources is under-explored (Li et al., 2017). This study investigates the potential misallocation of R&D resources caused by subsidy policy and further explores its impacts on firms efficiency in innovation.

Methodology and empirical base: The data are taken from a firm survey in Zhongguancun Science Park (ZSP) in Beijing, China. Most of Chinese high-tech firms locate in the Science Parks. Founded in 1988, ZSP is the largest Science Park in China in terms of both the number of firms and their gross outputs. As of the mid-1990s, all firms identified as high-tech in the Park had to compulsorily submit their annual financial statements to the Administrative Committee of the ZSP. Therefore, the sample firms feature more active R&D activities. The dataset provides detailed information about, among others, identification, financial performance, operations, R&D activities, and innovation for around 13,541 to 18,433 firms across different years.

We build a model to measure R&D resource misallocation across firms by applying Hsieh and Klenow (2009) approach to a knowledge production function setup. The differences in innovation efficiency are the result of industrial discrepancies in the R&D input and output market distortion. Panel threshold model is adopted to estimate the heterogeneous effects of subsidies on R&D resource distortion.


Results : We find that innovation efficiency in China increased for the period of 2005 to 2009, while it fluctuates during 2010 to 2015. R&D subsidies generate heterogeneous impacts on the innovation efficiency for firms. Subsidies on lower productivity firms and state-owned firms generate higher dispersion on TFPR, and therefore cause bigger resource distortion. Resources should not be allocated to lowering barriers but rather towards lowering the cost of R&D. 
References  Banerjee, A. V. and Moll, B. (2010). Why does misallocation persist?, American Economic Journal: Macroeconomics 2(1): 189-206. Boeing, P. (2016). The allocation and effectiveness of china's r&d subsidies - evidence from listed firms, Research Policy 45(9): 1774-1789. Brandt, L., Tombe, T. and Zhu, X. (2013). Factor market distortions across time, space and sectors in china, Review of Economic Dynamics 16(1): 39-58. Bresnahan, T. and Yin, P.-L. (2010). Reallocating innovative resources around growth bottlenecks, Industrial and Corporate Change 19(5): 1589{1627. Bronzini, R. and Piselli, P. (2016). The impact of r&d subsidies on firm innovation, Research Policy 45(2): 442 - 457. Li, H.C., Lee, W.-C. and Ko, B.T. (2017). What determines misallocation in innovation? A study of regional innovation in china, Journal of Macroeconomics 52(Supplement C): 221-237. Howell, A. (2017). Picking `winners' in china: Do subsidies matter for indigenous innovation and firm productivity?, China Economic Review 44: 154-165. Hsieh, C.-T. and Klenow, P. J. (2009). Misallocation and manufacturing tfp in china and india, The Quarterly Journal of Economics 124(4): 1403-1448. Ranasinghe, A. (2014). Impact of policy distortions on firm-level innovation, productivity dynamics and tfp, Journal of Economic Dynamics and Control 46: 114-129. Restuccia, D. and Rogerson, R. (2017). The causes and costs of misallocation, Journal of Economic Perspectives 31(3): 151-74. Wei, S.-J., Xie, Z. and Zhang, X. (2017). From “made in china" to “innovated in china": Necessity, prospect, and challenges, Journal of Economic Perspectives 31(1): 49-70.

ABSTRACT. Background and Rationale. Technological advancements have historically been important forces in reshaping work environments and the workforce, with significant implications for economies and societies. Recent developments in automation and artificial intelligence promise rapid and possibly unprecedented levels of change in the labor market, with impacts ranging from how basic tasks are performed to more fundamental changes to the nature of employment itself.

Recent discourse surrounding these emerging technologies has focused largely on labor displacement, arguing that labor substitution is a problem while labor complementarity can be construed as a positive force. In this paper, we look beyond wages and employment rates, arguing that labor complementarity may not be uniformly positive. In particular, increasing levels of automation in work may impact workers’ well-being in the present as well as their expectations about the future. Technology can impact worker stress, difficulty of tasks, levels of monitoring, autonomy, and job security, among other impacts.

Methods. To examine this concern about the impact of automation on worker well-being, we rely on two data sets. First, we examine a 2017 Pew Research Center survey on Automation in Everyday Life, which considers how individuals perceive automating technologies have affected their work, as well as their future work expectations. We restrict our attention to employed individuals, and contrast how individuals in manual labor roles are affected compared to those in management labor roles. This analytical approach follows theory emphasizing the difference in technology impacts across ‘low skill’ and ‘high skill’ work (Griliches, 1969). Standard OLS is used to estimate differential effects for workers across industries and demographic characteristics along a range of outcomes related to worker well-being and future expectations.

Second, we use the 2012 and 2013 American Time Use Survey (ATUS) Well-Being Module to ascertain whether workers in highly automated jobs experience different levels of well-being (measured through scales of happiness, stress, tiredness, meaningfulness, etc.,) when performing their jobs. The data set allows us to directly measure well-being of workers while working. To identify the degree of automation in each occupation, we use the Skill-Biased Technological Change literature (Autor, Levy and Murnane, 2003; Acemoglu and Autor 2010; Jaimovich and Siu 2014) to classify workers by their occupation’s task content, comparing routine work to nonroutine work and cognitive to noncognitive work.

As an alternative approach, we adapt the methodology of Frey and Osborne (2017). Their approach uses a Gaussian process classifier and expert determination to classify 700+ occupations by their susceptibility to computerization, leveraging detailed occupation data from the Current Population Survey and O*NET. Jobs with significant elements of physical manipulation and social and creative intelligence are classified as less automatable. Both classifications are used to evaluate impacts on well-being, again using standard OLS with demographic and industry controls. These alternative classifications of automatability serve as robustness checks on our findings and additionally help to assess the reliability of the automation-based occupation classification approaches themselves.

Preliminary Results. Our initial results from the first data set indicate that workers perceive meaningful impacts on their well-being as result of automation, and that these results differ across manual and managerial (cognitive jobs), ranging from around .1 to .2 standard deviations across several outcome variables. Manual work appears to be associated with negative impacts on outcomes including how interesting work is, opportunities for advancement, and job security. Standardized effects are reduced but remain significant in most cases when control variables for education, industry, and prior knowledge about automation are included.

Additional results from the second data set suggest mixed and weaker associations between well-being (happiness, meaningfulness, stress) and future susceptibility to automation. These results may be explained by individuals’ lack of knowledge about the “automatability” of their work, especially given that these jobs have not yet been (fully) automated. Additional work is needed to contrast present-day impacts on well-being caused by automation already in place versus impacts on well-being due to future expectations about automation. Moreover, more work is needed to develop 1) the relationship between the three theories that connect work activities to technological change (manual vs. cognitive, routine vs. non-routine, and Frey and Osborne’s newer classification) and 2) the relationship between these theories and well-being.

Significance. In view of impending technological change due to automation and artificial intelligence, this research sheds light on emerging questions about technological complementarity in the workplace. By expanding the scope of discourse beyond displacement and pay to address the critical issue of worker well-being, we provide a more robust understanding of the impacts of automating technology on the lives of workers. In turn, worker well-being has significant implications for labor, education/training, health, welfare, and socio-political stability, and is therefore worthwhile of attention by researchers and policymakers.

ABSTRACT. Background Since the late 1980s, dual-career couples have accounted for over one-third of the academic workforce, yet they remain an understudied population. The majority of studies focuses on the two-body problem and with that on a narrow set of questions revolving around geographic (im)mobility, career outcomes, and family life. In an attempt to recruit and retain qualified faculty, institutions have responded to the two-body problem by implementing dual-career policies that assist both partners to find positions. Much less is known about how dual-career couples navigate academic life, including how well they are integrated socially and professionally at their institutions and how they build and maintain collegial ties. These questions are critical considering that social relationships in academia and within the institution in particular play an important role for professional identity, career advancement, and workplace satisfaction. Social networks serve as important channels of information about collaboration opportunities, funding, and scientific discoveries, which affects the scientist’s scholarly performance and visibility. Networks also offer guidance, advice, and social support, affecting scientists’ psychological well-being and feelings of inclusion within the institution and their field.

Research Question Informed by social capital theory, we ask: In what ways, if any, do dual-career couples’ collegial ties differ from those of non-dual-career couples? We are interested in structural aspects including institutional network size and network composition as well as non-structural aspects such as patterns of network resource use (collaboration, career advice, teaching advice). We focus on collegial networks that cross-cut traditional boundaries between purely professional and social relationships and enable socializing and friendship building. On a conceptual level, collegial-friendship ties (i.e., faculty who are considered friends) and collegial-socializing ties (i.e., faculty with whom someone socializes outside of work) demonstrate close, trust-based relationships that may enable access to and exchange of resources that otherwise would not exist or occur. Further, we distinguish between dual-faculty (i.e., couples in which both partners are in academia) and dual-PhD (i.e., couples in which both partners have an academic background but work in different industries) scientists. Both groups are similar in the way that they are characterized by low educational and occupational disparity, fairly equal access to resources, and egalitarian values resulting in collaborative and supportive partnerships. However, in contrast to dual-PhD scientists, dual-faculty scientists have formed a joint commitment to the academic lifestyle in which social relationships form within the same workplace setting. As such, dual-career faculty may be able to piggyback on their spouse’s network in ways that dual-PhD and other groups cannot.

Data & Methods We use data from the 2015 NETWISE II study, an NSF-funded national survey of tenured and tenure-track faculty members in STEM disciplines, which includes a series of questions about a scientist’s collaborative and advice network within their institution. We conduct an extensive descriptive analysis to understand the structural characteristics of socializing and friendship networks and how scientists use those networks for instrumental or psychosocial support. Next, we use negative binomial regression to estimate network sizes and logistic regression to predict the proportion of socializing and/or friendship ties to a scientist’s total on-campus network, controlling for gender, race, academic age (years since PhD), rank (Assistant, Associate, Full), academic discipline (biology, biochemistry, civil engineering, and mathematics), and institutional type (Research Intensive, Research Extensive, other).

Preliminary Results Network size: We find that socializing networks are larger than friendship networks, both for dual-faculty and dual-PhD scientists. However, friendships with faculty do not necessarily subsume socializing ties. This means that faculty socialize with some colleagues outside of work but do not consider them friends, and vice versa. Second, our regression analyses show that the social capital of dual-faculty scientists differs from those of dual-PhD and non-dual-career scientists. Specifically, dual-faculty have more socializing and friendship ties, both in absolute terms and in proportion to their overall on-campus network. Resource use: We find that faculty gain resources from their on-campus colleagues in a number of different ways. A) Collaboration: Dual-faculty have more socializing ties within their collaborative network than dual-PhD and non-dual-career scientists. In contrast, dual-PhD academics seek more collaborators outside their socializing network than dual-faculty and non-dual-career academics. However, we do not find any differences between the three groups in terms of collaboration within or outside friendship networks. B) Teaching Advice: All faculty seek teaching advice more from colleagues they socialize with and/or colleagues they call friends than from people who fall outside those close circles of ties. However, dual-faculty have proportionately higher numbers of socializing and friendship ties within their teaching-advice network than dual-PhD and non-dual-career scientists. C) Career Advice: The proportion of socializing ties within someone’s career-advice network is higher for dual-faculty than it is for dual-PhD scientists. In contrast, dual-PhD seek proportionately more career advice from non-socializing and non-friendship ties than dual-faculty.

Significance Our analyses indicate that dual-faculty scientists are deeply embedded in their institutions. This finding is particularly interesting in the light of qualitative studies suggesting that institution caution to recruit academic couples as they might negatively impact departmental climates and harmony.