previous day
next day
all days

View: session overviewtalk overview

08:30-10:00 Session 6A: S&T Policy Processes (SciSIP)


Location: GLC 233
A collaboratively-derived international research agenda on legislative science advice
PRESENTER: Karen Akerlof

ABSTRACT. Background and rationale Both in presidential and parliamentary systems of government, legislatures can play substantial roles in setting national policy, if with different degrees of power. In executing their functions, policymakers rely on receiving information from complex advisory systems: networks of expertise both within the legislature and externally. Scientific and technological information underlie many of the critical issues legislators face, whether cybersecurity, climate change, nuclear power, food security, or vaccination.

A variety of structures for integrating scientific and technical expertise into policymaking has emerged internationally, reflecting distinctive cultures and traditions of decision-making. These can be formal or informal, internal or external, permanent or ad hoc, and operate at different branches and levels of government. The academic study of policy advisory systems in general remains largely atheoretical, focused on Western democracies, and based in qualitative case studies that can be difficult to translate to practice. As a subfield, the study of scientific advice similarly suffers from these deficits, with less attention to legislatures than the executive.

To further an understanding of scientific advisory systems internationally, we asked academics, science advisers, and policymakers across the globe to identify the most pressing research questions that will improve the practice of science advice to legislatures and broaden its theoretical and empirical foundations, using a three-stage research approach. Similar expert consultation exercises designed to identify the most important questions in science policy and ecology have been effective in informing government strategy.

Methods The study consists of three stages: research question collection, vetting, and prioritization. An online survey was first used to collect research questions from academics, science advisers, and policymakers worldwide. All submitted research questions were coded for categories of dynamics and actors in legislative science advice that were most frequently addressed, and assessments of coder inter-reliability were conducted. Cluster analyses identified the predominant combinations of categories. Next, during a workshop at the International Network for Government Science Advice (INGSA) Conference in November 2018 in Tokyo, Japan, participants scrutinized the question set. Finally, a subset of the original survey participants ranked which information they would most be interested in learning. Differences between experts in developed and developing countries, and roles in producing, providing, and using legislative science advice, were analyzed using Q methodology.

Study respondent characteristics From September to November 2018, 183 respondents in 53 nations submitted 254 questions. Approximately half of the respondents to our request for research questions were from nations categorized by the United Nations as developing (n = 90) and half from those considered developed (n = 93). While all had expertise in science advice to policy, almost three-quarters (74%) said they also had specific experience with legislatures in their roles. The roles of these experts in the science advisory system differed greatly. One quarter (25%) of respondents self-identified as conducting research on science advice for policy. The largest group—almost half (45%)—said that they provide scientific information to government. Only 11% said that they use scientific information within government (e.g. policymakers). One-fifth of respondents said that their roles were a combination of these categories or described them in other ways.

At the INGSA workshop in 2018, 36 people from 17 nations vetted the initial research question list. Subsequently, 90 participants in the original research question collection were asked to rank what information from the research questions they would be most interested in learning; 64 individuals from 31 countries responded. Thirty-three were from—or in one case studied—developing nations, and 30 from developed countries. Twenty percent said that they produce scientific information, 33% said that they provide it to government, or otherwise facilitate science advice processes in a practitioner role, and 8 percent said that they use it within government. The rest had served in multiple roles. All but one of the experts in science advice to policy said that they have experience specifically with legislatures.

Results The coded categories of our 254 submitted research questions reflect diverse themes: evidence use and development; characteristics and/or capacity of system components and actors; system design and implementation; and ethics. With the help of workshop participants, the research questions were distilled into a final list of 100 and a corresponding "short set" of 50, representative of the coded categories, that were ranked according to level of interest in the potential research findings by a subset of contributors (n=63).

Two broad findings emerge from the study. First, experts generally agree that the state of understanding of legislative science advice is deficient, especially for developing and lower-middle income nations. More than two-thirds of our second sample of international legislative science experts (68%; n = 63)—who ranked the list of legislative science advice research priorities—rated the state of the evidence as poor. Second, many fundamental questions about the function and design of legislative science advisory systems remain unanswered.

The top five ranked priorities for legislative science advice research are as follows:

(1) Whether legislative use of scientific evidence improves the implementation and outcome of social programs and policies (2) Under which conditions the use of scientific information changes the framing of policy debates (3) Under what conditions legislators and staff seek out scientific information or use what is presented to them (4) How legislators and their staff assess the credibility of scientific information (5) How different communication channels—hearings, face-to-face meetings, email, social media, etc.—affect informational trust and use

The types of research questions provided by the expert community lend an additional indication, in additional to the ranking, of their interests. Of the 254 submitted research questions, most addressed evidence use (63%) and communication (53%). Policymakers (70%), institutions (62%), and scientists (53%) were referenced the most frequently as system actors. Interestingly, study participants—only a minority of whom are policymakers—were more focused in their research interests on the institutional/policymaker side of the system, rather than that of scientists and information generation, or the brokers that span the two.

When asked directly in the initial survey collecting the research questions, slightly more than half of the experts (51%) said that some policy issue areas are more important to focus on than others. Of those that said some policy areas should be a priority for the field, a majority said that the environment (78%), health (64%), and natural resources (56%) should be the focus.

[Analyses underway] Finally, using Q Methodology, we assess differences between respondents in the ranking of research interests, which may be used to explore how an international research agenda for legislative science advice will need to accommodate national differences in legislative science advice systems, as well as differing perspectives between academics, science advisers, and policymakers.

Significance By characterizing experts’ priorities for legislative science advice research, we hope this collaborative project helps spur new initiatives not just in the United States and other developed Western nations, but spanning internationally, that will contribute to empirical, theoretical, and applied advances in the field. Furthermore, by analyzing differences in expert prioritization of research by national development status and roles within the science policy community, we characterize both areas of international consensus and divergence that are important to address to ensure a research agenda that is broadly inclusive and relevant to community needs.

Science Advice to the President and the Role and Impact of the President's Council of Advisors on Science and Technology (PCAST)
PRESENTER: Kirstin Matthews

ABSTRACT. Science advice within the White House dates back to World War II, but the current organization, led by the Office of Science and Technology (OSTP) was more recently developed. OSTP is charged with giving and overseeing science and technology advice to the president and coordinating activities within the executive branch. The director of OSTP is informally known as the president’s science advisor and has often also served as an assistant to the president, a title that indicates direct access to the president as a confidential advisor. OSTP was created during the Ford administration as a result of President Nixon dissolving the Office of Science and Technology (OST), a predecessor to OSTP, at the start of his second term. OSTP is currently led by Dr. Kelvin Droegemeier, former Vice President for Research at University of Oklahoma and vice chairman of the National Science Board, who was nominated by President Trump in August 2018 and confirmed by the Senate in January 2019.

Additional advice to the president has also come from the President’s Council of Advisor on Science and Technology (PCAST). PCAST is a federal advisory committee consisting of roughly 20 highly-respected scientists, engineers, and industry leaders appointed by the president. PCAST is led by the science advisor, and up to two other non-federal members from the science, technology, engineering, and mathematics (STEM) community. Created in 1990 under President George H.W. Bush, PCAST advises the president on a broad range of issues related to STEM including national and domestic security, workforce and the economy, energy and the environment, and public health and medicine. The council, with its high-profile membership and proximity to the president, has historically played a central role in shaping federal STEM policy and maintaining the overall health of the nation's STEM enterprise. While no PCAST members have been appointed so far in the Trump administration, PCAST’s charter has been renewed by executive order and OSTP has publicly indicated that it plans to continue its operations.

This paper will review the history of OSTP and PCAST, and the president’s science advisor who oversee both. A particular focus will be given to the membership, activities, and policy impact of PCAST during the past four presidential administrations: Presidents George H.W. Bush, Bill Clinton, George W. Bush, and Barack Obama. In addition, the paper will analyze the current state of the federal science advisory system and recent policymaking activities during the Trump administration. This analysis includes a documentation the demographics of PCAST members—including diversity based on gender, ethnicity, and areas of expertise—as well as the shifting focus and influence of OSTP and PCAST activities in each administration. The research is based on a comprehensive literature review of OSTP and PCAST reports, transcripts from PCAST meetings, news items, and other publicly available materials, as well as interviews with selected members of PCAST. This research is used to understand the role and impact of scientists and engineers working in the White House over the last forty years and to develop recommendations for best practices for PCAST and OSTP in the current and future administrations.

This material is based in part upon work supported by the National Science Foundation (Grant Numbers SMA-1735682 and SMA-1854055) and the Richard Lounsbery Foundation. Additional funding was provided by Baker Institute Civic Scientist Program donors, Benjamin and Winifer Cheng.

Measuring & Analyzing the Use of Science and Innovation in Policy: An Application to National Environmental Policy Act Decisions
PRESENTER: Cory Struthers

ABSTRACT. The use of science in policy decision-making continues to gain attention across disciplines (e.g., Head, 2015; Costa et al., 2018; Desmarais and Hird, 2014; Yannovitsky and Weber, n.d.), facilitated by the growing access to policy documents online and the increasing capabilities of programming software that expedite big data extraction and text mining. At the same time, citation analyses, where scholars programmatically extract citations from publications, analyze patterns in their components, and infer their meaning (such as whether the authors cited are representative of a field’s experts), is also expanding (e.g., Potthoff and Zimmermann, 2017; Chorus and Waltman, 2016; Ferreira et al., 2015). Yet most programming software designed to extract citations and the information within them is built to locate and parse peer-reviewed journal articles and not the technical reports, gray literature, and other sources often used by the policy community, posing major challenges to merging these two inquiries. In other words, we have yet to develop the automated methods to extract, analyze, and infer the scientific and innovative nature of the various sources used in documents that justify policy decisions.

We tackle this problem through two main tasks using Environmental Impact Statements (EISs) produced by the United States National Forest (USFS). EISs are an ideal set of documents in which to construct our methods and measures. The 1969 National Environmental Policy Act (NEPA) requires that all proposed projects in federal agencies be evaluated for their environmental impacts according to their level of environmental risk. EISs are required for the highest risk projects, meaning forest supervisors and planners along with their teams must justify their project decisions using the best available natural and social science (42 USC § 4332A). We focus on the USFS because it prepares more EISs than any other federal government agency (Bixler et al. 2016), and because its offices maintain detailed repositories of NEPA documents. The USFS is also a lead agency overseeing land and natural resources, including logging, recreational, mining, oil, and related activities (Kaufman 1967; Koontz 2007), meaning the scope of the science used is broad enough that the tools we build can be implemented across domains. We collect the universe of draft and final EIS documents from the USFS office websites from 2004-2017, providing more than 1,500 EIS and thousands of citation observations.

Our first task in analyzing the use of science in policy documents through citation analysis is developing programmatic tools that are capable of extracting citations across a wide range of citation types. Using a combination of automation and human-coding, we develop a model that can identify citations in academic journal articles, technical reports, newspapers, legal documents, and gray literature, and then parse these references into sub-components like author, title, and year. Because there is often no standard citation formatting across EISs (or other policy documents for that matter), we use Google’s search engine to identify the likelihood that differently formatted but similar citations across EISs are citing the same source. We use human-coding on a representative sample of observations to verify our approaches.

Our second task is developing a series of metrics to measure and analyze the science and innovation used in each EIS based on its citations. Our first measure is quantity, or the number of citations in a given EIS. The second is novelty, or number of citations in a given EIS that have not appeared previously in EISs prepared by the same office on the same topic. This is our proxy for innovation. The final measure is quality, assessed in two ways. One way is using the bibliometrics readily available through Google Scholar and Web of Science, such as the Incites Journal Impact Factor published by Clarivate Analytics (formerly ISI) and the h-index, both of which assign academic journals scores based on how widely cited their articles (Hirsch 2005) and are conventional approaches to measuring information quality (e.g., Andersen et al. 2006; Costa et al. 2016; Saha et al. 2003; Garfield 1999). However, since this approach is only applicable to academic journals and assesses quality based on how frequently the scientific community utilizes research, we develop alternative metrics that instead measure the value of literature according to the relevant policy community; that is, how widely particular sources of any type are cited across USFS EISs. In addition to quantifying the value of sources outside of academia, our “management impact” index allows us to consider whether forest supervisors and scientists value peer-reviewed research differently.

The main contributions of this work are threefold. First, we advance methods to collect and analyze citations, which are a key data source in which to infer the use of science and innovation. Second, our methods to collect and analyze citations are not limited to academic journals, but to the broader range of sources commonly used in policy documents. Accordingly, our project creates new opportunities to understand what “evidence-based” policy means in practice, and how and why it varies across policy decisions. For instance, research that follows this paper will involve explaining the variation in the quantity, novelty, and quality of USFS EISs, including the role of authors’ backgrounds, the availability of science through open-access journals, and public participation. Finally, although NEPA is widely considered one of the most important environmental laws in the US (Mandelker, 2010), it has attracted relatively little scholarly attention, and none of that attention has focused on the use of science (Bixler et al. 2016). Our project helps to fill this gap.

Government-funded research increasingly fuels innovation

ABSTRACT. Innovation increasingly relies on scientific knowledge. Research to generate that knowledge has historically been funded by both industry and government. Although industry and government research spending was relatively equal in 1980 in the United States, by 2010 their shares had shifted to 60% and 30%, respectively. Yet, despite this increase in industrial spending, firms appear to be pursuing—or at least publishing—less basic science. If corporations are doing less basic research, then where do they find the ideas to fuel their innovation? Here, we detail individual bibliometric linkages across tens of millions of documents and quantify the broad sweep and impact of U.S. federally supported research on patented innovation over most of the past century. We illustrate how patentees, both U.S. and non-U.S., and corporations in particular, increasingly depend upon federally supported research as a source of scientific knowledge. Though research has established the rise of “open innovation” communities, research consortia, and markets for ideas as sources of innovative ideas, the role of government funding for research has not been ignored. Recent research—at the micro level of individual science papers and patents—has quantified how 10% of grants awarded by the U.S. National Institutes of Health (NIH) generate patents. Macro-level and nonquantitative histories have also described the broad sweep of government impact. Here, we consider every U.S. patent assigned to U.S. inventors since 1926, and show that the proportion of patents relying on federally funded science has outstripped the overall increase in patenting, plateauing since a high of 30.0% in 2011. Despite this plateau, the absolute number of patents that rely on federally supported research has almost doubled recently, from 22,647 in 2008 to 45,220 in 2017. We count reliance when a patent is owned by the government, acknowledges government support, or directly cites a patent or science paper that acknowledges support. Corporations, identified as the owner of a patent from its “assignee” field, account for most of the increased reliance on government-supported research since the 1970s. Corporations have increased their own acknowledgment of support and their citation of government-supported science papers and patents [although alternatives cannot be entirely ruled out, see supplementary materials (SM) for evidence that the trends are robust to more versus less stringent definitions of reliance and also do not result from a change in reporting requirements or in response to requirements, an increase in the citable stock of federal research, or a mechanical process due to an increase in the number of citations per patent].

Prior to the 1970s, much of the government-supported patenting in the United States occurred inside the government. Universities, often supported by government research grants, began patenting more in the 1980s, motivated in part by the Bayh-Dole Act, which sought to spur commercialization through assignment of property rights to universities. Lone inventors constituted a large proportion of reliance upon government research in the early part of the last century, and their reliance has remained steady.

Although some portion of the overall increase results from a change in the composition of patented technologies, growing investment in health research, and an accompanying rise in patents supported by the Department of Health and Human Services (HHS) (which includes the NIH), all technologies have relied increasingly upon government support, albeit to varying degrees. Federal support for U.S. invention gets routed through a variety of agencies. In 2017, for example, the Department of Defense (including the armed services) supported 6.2% of the total of U.S. inventions, HHS 5.4%, Department of Energy 3.9%, National Science Foundation 2.9%, NASA 1.0%, and Department of Agriculture 0.50% (others 5.5% and unspecified 2.5%; see SM for illustrations of R&D investment and patenting by agency and by technology class).

Large, small, and micro firms all draw upon government research, as observed in patent-renewal data. Startups also depend heavily on government research, as 34.6% of the 121,765 patents assigned to venture-backed companies from 1976 to 2016 cited federally supported research; by comparison, for all corporate patents during the same time period, only 21.7% rely on federally supported research. The higher reliance of startups on government-supported research may be explained by resource constraints, which inhibit internal R&D expenditure in young ventures and encourage identification and external licensing of promising technologies. It also illustrates the intended consequence of the Bayh-Dole Act and an important path of commercialization for federal science and technology research.

Corporate patents that rely on federal research are consistently more important than those that do not, as measured by the number of prior art citations from subsequent patents (such citations establish the departure point of a new patent). Comparing population averages for all U.S. corporate patents granted in 2010, patents that rely on federal research receive 6.33 citations on average in the 5 years following the grant versus 4.42 for those that do not. Additional panels in 2000, 1990, and 1980 show similar but not always significant effects, possibly because of thinner data or changes in corporate patenting strategy. These population averages aggregate a great variety of different technologies, industry-specific patenting strategies, and number of inventors on the patent, however, which motivates a comparison of similar types of patents. Matching and pairing corporate patents that cite federally supported patents to similar corporate patents that did not (see SM), the former receive on average 3.39 more citations in the 5 years after issuance. The citation increase arises from both the firm's subsequent patents (1.53 additional self citations on average) and citations in other firms' patents as well (1.86 additional citations from other organizations on average), indicating that both the inventing firm and its competitors find these technological trajectories more fertile.

The same pattern holds for corporate patents that cite federally supported science publications relative to those that cite science publications that are not government supported, with a matched and paired mean difference of 3.57 citations, and a mean difference in self- and nonself-citations of 1.17 and 2.41, respectively. Prior research has estimated that an additional citation is associated with an increase in the value of the patent of up to 3%. Citation increases also benefit the larger economy as citations that occur from outside the inventing firm can be interpreted as knowledge spillovers and positive externalities from the inventing firm to the larger economy.

Consistent with the citation measure and again comparing matched patents, patents that rely on federal research are slightly more likely (<2%) to pay renewal fees (required payments that keep the patent in force) after 4 years. They are also more likely, on average, to introduce words that are new to the patent corpus, indicating foundational patents with greater novelty.

08:30-10:00 Session 6B: Chinese Return Mobility


Location: GLC 235
Returnee Academic Entrepreneurship in China
PRESENTER: Yanzhao Lai

ABSTRACT. 1.Research Objectives Over the past two decades, alongside the traditional teaching and research roles of universities, the so-called “third mission” of university, namely the direct contribution to the development of the wider social and economic community, has received great attention from both academic and policy community. However, as Siegel and Wright (2015) pointed out, given that many universities are attempting to increase their international profiles, an important omission in academic entrepreneurship studies is returnee academic entrepreneurship. Recently, the reverse flows of highly educated individuals from OECD countries to emerging economies such as China and India, have increased significantly (Qin and Estrin, 2015; Qin et al., 2017). Returnees, especially returnee academics are considered as important carriers of tacit knowledge and contributors of innovation and knowledge transfer in their home countries (Filatotchev et al., 2011). A developing research stream has analyzed the unique role of returnee academics in filling important innovation and economic gaps in emerging economies (Velema, 2012). Having an advanced degree from a foreign country are usually viewed as a sign of a high-quality education and advanced level research skills. Employers tend to perceive foreign degree holders as well trained and more research productive (Shin et al., 2014). Accordingly, growing number of returnee academics and university entrepreneurial activities lead to the research questions in this paper: Will the returnee academics be more likely to start new firms than domestic educated academics?

2.Theory and Hypotheses Scholars such as Acs et al. (2009) have introduced the knowledge spillover theory of entrepreneurship (KSTE) which identifies new knowledge as a source of entrepreneurial opportunities and considers entrepreneurship as a conduit of knowledge spillovers. Returnee academics, who studied or worked in OECD countries in the past, possess two major characteristics that differentiate them from domestic-educated academics. First, returnee academics may have specific human capital that relates to a spectrum of knowledge (Castanias and Helfat, 1992). They are simultaneously embedded in two distinctive knowledge contexts in the country they studied and their home country. Embeddedness in the country in which they studied gives them an opportunity to draw upon sources of advanced knowledge and new ideas. Moreover, returnees’ cultural background and language skills enable them to exploit non-local experiences and knowledge (Liu et al.,2010). Second, returnee academics may develop specific social capital through a network of social relationships in the host country (Cooper and Yin, 2005). They may be able to maintain social ties in the host country which enable them to continue updating technology when they return to their home country. We hypothesize: Hypothesis 1: The likelihood of an academic to involve into entrepreneurship increases if s/he has overseas experience (Ph.D education).

Doctoral and post-doctoral study abroad experiences are highly associated with future research activity and academic performance which leads to potential higher likelihood of an academic to involve into entrepreneurship. However, as Carstensen (2006) have pointed out that the time spending on studying and working has positive effect on the scientist’s academic performance. Returnee academics with oversea PhD normally spend more time (5-6 years) than postdoc (2-3 years) aboard. The more time they spend overseas, the returnees are more likely to possess more experience with advanced technological knowledge, which leads to returnee academics with a Ph.D. overseas with higher possibility to start new business than returnee academics with only Postdoc experience overseas. We hypothesize, Hypothesis 2: Returnee academics with a Ph.D. overseas are more likely to start new business than returnee academics with only Postdoc experience overseas.

The knowledge gap between countries may create entrepreneurial opportunities (Audretsch and Lehmann, 2005). Experiences in both developed countries and their home countries expose returnees to the technological and business practice gaps between these country contexts; exposure to these gaps is important because emerging markets generally follow developed countries in their economic and technology development (Batjargal, 2007). Returnees can contribute to the performance of technology ventures by identifying and capitalizing upon brokerage opportunities. We hypothesize, Hypothesis 3: The larger the gap in the level of economic development between the home country and the host country, the higher the likelihood that returnee academics get involved into entrepreneurship.

3. Data, Methodology and Preliminary results To address this issue, we bring to bear a unique dataset covers the entrepreneurial activities of 507 computer science faculties of 42 research-intensive universities in China from 2007 to 2017, including 138 returnee academics. To the best of our knowledge, this study is the first paper that investigates returnee academic entrepreneurship and provides initial clues on how aboard educational background affect academics’ entrepreneurial activities. Since we are working with a binary dependent variable (Entrepreneurship, the variable entrepreneurship is defined as an event whereby an academic becomes a shareholder or top manager of a company) and panel data, random-effect Logistic regression model have been suggested. To examine whether overseas experience will affect academic’s entrepreneurial decision, we have constructed two dummy variables to identify focal academic’s oversea experience: Returnee PhD and Returnee Postdoc, equals 1 if a scholar got their PhD overseas or was a post-doctoral fellow overseas, 0 otherwise. Our preliminary empirical findings have shown that academics’ overseas background has positive effect on academic entrepreneurship. Meanwhile, returnee academics with foreign Ph.D. degrees are more likely to start new business than the returnee academics with only Postdoc experience overseas. At last, the economic gap between the host and home countries do not have statistically strong effect on returnee academic’s entrepreneurial activities.

Reference 1. Acs, Zoltan J., et al. "The knowledge spillover theory of entrepreneurship." Small Business Economics 32.1 (2009): 15-30. 2. Audretsch, David B., and Erik E. Lehmann. "Does the knowledge spillover theory of entrepreneurship hold for regions?" Research policy 34.8 (2005): 1191-1202. 3. Batjargal, Bat. "Internet entrepreneurship: Social capital, human capital, and performance of Internet ventures in China." Research policy 36.5 (2007): 605-618. 4. Carstensen, Laura L. "The influence of a sense of time on human development." Science 312.5782 (2006): 1913-1915. 5. Castanias, Richard P., and Constance E. Helfat. "Managerial and windfall rents in the market for corporate control." Journal of Economic Behavior & Organization 18.2 (1992): 153-184. 6. Cooper, A. C., and X. Yin. "Entrepreneurial networks." The Blackwell encyclopaedia of entrepreneurship (2005): 98-100. 7. Filatotchev, Igor, et al. "Knowledge spillovers through human mobility across national borders: Evidence from Zhongguancun Science Park in China." Research Policy 40.3 (2011): 453-462. 8. Liu, Xiaohui, et al. "Returnee entrepreneurs, knowledge spillovers and innovation in high-tech firms in emerging economies." Journal of International Business Studies 41.7 (2010): 1183-1197. 9. Qin, Fei, and Saul Estrin. "Does Social Influence Span Time and Space? Evidence from I ndian Returnee Entrepreneurs." Strategic Entrepreneurship Journal 9.3 (2015): 226-242. 10. Qin, Fei, Mike Wright, and Jian Gao. "Are ‘sea turtles’ slower? Returnee entrepreneurs, venture resources and speed of entrepreneurial entry." Journal of Business Venturing 32.6 (2017): 694-706. 11. Shin, Jung Cheol, et al. "Research productivity of returnees from study abroad in Korea, Hong Kong, and Malaysia." Minerva 52.4 (2014): 467-487. 12. Siegel, Donald S., and Mike Wright. "Academic entrepreneurship: time for a rethink?" British Journal of Management 26.4 (2015): 582-595. 13. Velema, Thijs A. "The contingent nature of brain gain and brain circulation: their foreign context and the impact of return scientists on the scientific community in their country of origin." Scientometrics 93.3 (2012): 893-913.

From inbreeding to student laundering: How academic background is influencing the career attainment of Chinese overseas returnees

ABSTRACT. The paper aims to examine the hiring practices of Chinese universities and research institutions with the influx of its overseas returned scholars. The Chinese universities and research institutions used to retain a high inbreeding rate in adopting its own Ph.D students into tenure track positions. The labor structure of the Chinese academic labor market rapidly changed due to the reversed migration of its overseas students. Therefore, how did the hiring practices and corresponding institutional arrangement of the Chinese universities and research institutes tend to change. How did local networks and overseas experiences influence the career chances of the returnees? What are the implications for organization studies and labor market theories? There are the questions this paper will like to address. This paper used a sample of Chinese overseas returnees who secured their tenure track position in Chinese universities and research institutions through One thousand talent youth program in the year of 2013,2015,and 2017. Hypothesis are developed to examine the factors of undergraduate background, graduate background, local networks with former undergraduate program, overseas experience in influencing the career attainment outcome. It is found that the prestige of their domestic undergraduate school will significantly impact their career attainment outcomes. Close connections with their former undergraduate programs will not only offer higher odds of obtaining position in the same institution, but also provide faster promotion in their career. Chinese academic labor market is highly structured with trust that the hiring practices of student inbreeding has shifted to an organizational behavior of student laundering. Overseas experiences and domestic network embellished through undergraduate training will jointly influence the career attainment outcome of Chinese overseas returned scholars.

The career development of leading social scientists in China — A multidimensional comparison of returnees and locals

ABSTRACT. A surge of return migration in Chinese academics has aroused heated debates among returnee and local scholars. To date, most studies have focused on the performance differentials between returnee and local scholars in the fields of science and technology, while the returnee social scientists are the long-neglected group.

Chinese scientists and social scientists are using different criteria in assessing the success of academic career. Within the science community, a scholar’s career success is highly related to his/her publication record in international peer-reviewed journals. While in the fields of social sciences, gaining academic reputation nationally is important for a scholar to pursue career success. Therefore, as a social scientist, domestic scholarly publication are sometimes valued more in career development than international publication. Having transnational capital accumulated through overseas experience, returnee social scientists surely have advantages over local social scientists in international publication. However, further study is required to ascertain the post-return performance of returnee social scientists. Can they turn their transnational capital into local assets, which can be recognized by the domestic academic community? Or how will they adapt to the domestic recognition system? Do they share similar career patterns with local social scientists during the process of achieving career success? These questions are the main focus of this study.

This study targets at a group of 445 leading social scientists recruited by the Changjiang Scholars Program, which is one of the flagship talents programs in China. Based on the curriculum vitae data of these scholars, this study differentiates returnee social scientists from locals. By collecting additional data from multiple sources, this study compares the speed of achieving career success between two groups. This study plans to measure career success through multiple indicators, such as publishing high impact research in both international and domestic journals, getting highly selective national research fund, winning a career award, etc. Meanwhile, some often mentioned factors (e.g., administrative positions, social capital) influencing academic career will also be included in the analysis. This study expects to reveal the different career patterns between returnee and local social scientists. In light of the results, policy implications regarding the national talent policies will be discussed.

Returning scientists and the emergence of China’s science system
PRESENTER: Koen Jonkers

ABSTRACT. China has made investments in sending people abroad, in building domestic capacity, and in prioritizing advancements in targeted areas of science and technology. Given the scale and scope of these investments, and the unprecedented rise of a huge nation into the global science system, and the implications of this development for international affairs, this presentation presents the results of research into the growth of Chinese scientific personnel, their international engagement, the number of overseas Chinese scientists and returnees, and the contribution of these researchers to Chinese science.

China’s human resources development emerges at an unprecedented time, when the science system is global in scope and international in operation. This means that China cannot use the same model of participation that has characterized the rise of other nations over the past 30 years. Due to a lack of reliable statistics it is very difficult to estimate the exact number of overseas Chinese scientists and engineers and returnees. This presentation offers an estimate of the number of Chinese scientists in Europe and in the United States, as well as estimates of the number of returnees from these countries based on a bibliometric approach. We discuss the impact of mobility on the growth of China’s science.

08:30-10:00 Session 6C: Rethinking Technology Emergence
Location: GLC 236
Rethinking Technology Emergence: Measurement, Policy, Insights (Jan Youtie Panel Organizer, Submission #77); Paper Title: The Politics of Attention for Autonomous Vehicles: Technology Forecasts and Policy Images from U.S. Industry
PRESENTER: Gordon Kingsley

ABSTRACT. Background and Rationale: When developers of AV technology provide information to transportation agencies at the national and sub-national levels they are engaging with the units of government responsible for designing, building, and maintaining road systems and other related infrastructure assets. AV proponents project a future of transportation that shifts from individual owners and operators of vehicles to one in which mobility is a service with multiple forms of service providers (Skeete, 2018). Transportation agencies at the national and sub-national levels are not the end user of the technology. However, as the owner of the infrastructure assets on which mobility systems will operate they are interested in these forecasts as a means of planning a preparing for the changing nature of mobility demand.

In this light, technology forecasts from industry leaders are not simply statements of the comparative and competitive advantages of technologies. Embedded within forecasts are value propositions regarding optimal transition pathways. A key question that we explore is the degree to which industry forecasts incorporate the level of demand that optimal transition pathways are likely to place upon regime level actors responsible for transportation infrastructure. Geels (2004) finds that such disruptive innovations are likely to encounter strong inertial forces from public agencies with high levels of investment in the current prevailing technologies.

Researchers employing a multi-level perspectives (MLP) framework in technology forecasting have described such interactions as niche level innovators influencing, and in turn being influenced by, regime level actors and landscape level contexts (Skeete, 2018). However, recent studies have called for the further developing the MLP framework by incorporating a deeper understanding of the behavior of agents and actors as they attempt to shape information flows between the niche and regime levels (de Haan and Rotmans, 2018). We do so by exploring how AV industry leaders in the United States use technology forecasts to frame the politics of attention and agenda setting of policy makers in departments of transportation at the national and sub-national levels.

To help us explore the political processes into which industry actors’ project their technology forecasts for AV we turn to the politics of attention model (Jones and Baumgartner, 2005) of the introduction, and at times intrusion, of new information into the highly congested agenda setting processes of policy makers and public agencies. Doing so allows us to explore how industry actors may use technology forecasts to project desired policy images with regime level actors. First, industry forecasts can allocate attention to key issues (issue intrusion) in AV development. Second, forecasts can redefine and characterize problems and opportunities by encouraging policy makers to focus on key attributes of a technology (attribute intrusion). Third, forecasts can draw attention the preferred development paths of industry and include value propositions about how such pathways address regime level concerns (regime value alignment).

Methods: We explore the politics of attention embedded in an MLP framework by asking respondents to offer technology forecasts for AV that describe innovative developments at the niche level, opportunities and challenges associated with AV development stemming from the socio-technical regimes, and factors in the larger socio-technical landscape that may influence technology development paths. Drawing upon Geel and Schot (2007) we also asked the respondents to describe the most likely transition pathways.

Our approach is similar to Skeete’s (2018) study in that we interview AV industry leaders from a variety of firms representing a range of innovation niches. Where Skeete’s study focuses on industry leaders operating in the European Union our study explores industry leaders operating in the comparatively decentralized policy subsystems of the United States.

We examine 20 semi-structured interviews with AV industry leaders. We explore the variability of technology forecasts generated by AV industry actors occupying four distinct innovation niches: 1) automakers and hardware suppliers; 2) software developers, artificial intelligence, teleoperation and mobility services; 3) industry consultants, legal services, and think tanks; and 4) industry users in long-haul freight, delivery and shipping services, and agricultural equipment. We then examine the degree to which industry actors from similar innovation niches share attributes of technology forecasts from the perspective of the politics of attention factors. This allows us examine within niche and between niche similarities and differences in industry forecasts.

Anticipated Results: While our analysis is ongoing, the preliminary evidence suggests the following trends in industry forecasts for AV technology: a) There is a high degree of convergence across niches that the AV technology under development will be able to operate on the existing transportation grid with no changes required from departments of transportation. This is a form of issue intrusion designed to discourage policy-makers, particularly at the state and local levels of government, from developing regulatory requirements. While the niches project different transition pathways, all agree that competing and conflicting regulations across the jurisdictions will hinder the development of AV technology. b) There is little agreement across the niches regarding the timeline for AV deployment. Individual actors amongst the automakers and the mobility service providers project the most aggressive timelines for AV deployment. However, there is disagreement regarding the timeline even within a specific niche of actors. Projections range from commercial deployment of AV within 5 years to 40 years. c) There is a range of attribute intrusions in industry forecasts that have a direct bearing on the transportation grid. Automakers and mobility service providers highlight the challenges associated with curbside management as customers shift from the need to park their vehicle to accessing vehicles. Companies specializing in signals technology, teleoperation, and mobility services all noted challenges and opportunities associated with data management and data sharing across jurisdictions. Companies specializing in mobility and delivery services projected a range of new vehicle types that leading to a demand for new classes of infrastructure. Automakers, signals technology companies, and mobility service providers all noted the capabilities of AV to operate with greater precision allowing closer following distances, higher speeds of operation, and narrower lane needs. All of these attributes have significant impacts on the transportation grid. d) At present, none of the niches demonstrated a strong level of regime value alignment with regards to the departments of transportation that provide the infrastructure environments in which their commercial ventures will operate. There is a low level of comprehensive in industry forecasts regarding the range of services that departments of transportation provide at the state and local levels. The posture of industry actors regarding issue intrusion drives their positions related to regime value alignment. There is very little evidence linking the range of attribute intrusion to regime value alignment in industry forecasts and their preferred innovation pathways.


Geels, F.W. (2004). From sectoral systems of innovation to socio-technical systems: Insights about dynamics and change from sociology and institutional theory. Research Policy, 33, 897-920. Geels, F.W. and Schot, J.W. (2007). Typology of sociotechnical transition pathways. Research Policy, 36, 399-417. de Haan, F.J. and Rotmans, J. (2018). A proposed theoretical framework for actors in transformative change. Technology Forecasting & Social Change, 128, 275-286. Jones, B.D. and Baumgartner, F.R. (2005). The Politics of Attention: How Government Prioritizes Problems. University of Chicago Press. Skeete, J.P. (2018). Level 5 autonomy: The new face of disruption in road transport. Technology Forecasting & Social Change, 134, 22-34.

Emerging Technologies: Responsible Development and Challenges of Societal Alignment
PRESENTER: Philip Shapira

ABSTRACT. This presentation is submitted as part of the panel organized by Jan Youtie (Georgia Institute of Technology) on Rethinking Technology Emergence: Measurement, Policy, Insights

Alongside increased attention in recent years to defining, prioritizing, and measuring technological emergence, there is a renewed focus on the responsible development of emerging technologies. In particular, frameworks for responsible research and innovation (RRI) have been put forward. There are variations in RRI emphasis, resources, activities, embedding, and labelling across countries and by technological domains. Broadly, however, RRI frameworks seek to anticipate positive and negative impacts of emerging technologies, reflect on societal and ethical dimensions, and engage with diverse stakeholders and actors in the agendas and governance of research and innovation. A key aim is to promote scientific and innovation strategies and practices that better address societal challenges, foster inclusive and sustainable development, and mitigate (if not avoid) the societal and other risks associated with new technologies. As an overarching goal, RRI thus seeks to align societal needs with the products and processes of research and innovation. Examples of RRI (or RRI-like) initiatives are found in the synthetic biology and in other areas of engineering and scientific research in the UK, the major 7-year European Union Horizon 2020 Research and Innovation Programme, and in the US national nanotechnology initiative, through its centers for nanotechnology in society, as well as in multiple other countries around the globe. As RRI initiatives and experiences have grown, it is timely to probe the question of how well RRI is contributing to the alignment between society and emerging technologies.

The paper considers conceptualizations and framings of RRI, drawing on recent debates about the possibilities – and tensions – of effectively addressing the societal alignment of emerging technologies. It is recognized that there are multiple challenges inherent in attempting to foster the responsible development of emerging technologies. These can be framed in terms of two broad problems. First, the longstanding dilemma of technological control. As highlighted by Collingridge, at the early stages of technological development, when there is relatively high flexibility in technological design, there is relatively low knowledge of potential impacts. It is only as the technology is deployed and embedded in society that its impacts, including negative effects, are more fully realized. However, by the time these can be identified, change has become more expensive and difficult. Foresight and technology assessment are among the mechanisms that have been applied to attempt to address this dilemma. The second complementary problem is that of societal alignment, which is about how diverse stakeholder perspectives and the needs of diverse publics can be incorporated into the governance and configuration of emergent technologies. This is not so much about projecting the potential implications of new technologies, which has typically been seen as a technical exercise, but more about including society and consideration of societal values and needs in the early stages of designs – and changing technological designs vis à vis these values and needs. RRI seeks to address both of these problems, through early anticipation, reflection and engagement. We discuss this debate, then draw on experiences of a social sciences research group focused on RRI and engaged with synthetic biology researchers, industry, policymakers, and other stakeholders in the UK. Through reflective analysis, including review of examples and projects over a five-year period, we examine the operationalization of RRI within the context of UK synthetic biology development. Taking a perspective from the social sciences, we highlight how the challenges of the social alignment of this emerging technology have been both framed and addressed by the synthetic biology community. We also explore the opportunities and limitations of current methods for engaging publics; for eliciting values from different groups of actors; and for achieving a key RRI dimension, that of “responsiveness” (i.e. changes in scientific and organizational practices as a result of anticipation, reflection and engagement). Insights are developed both for the further development of RRI practices in synthetic biology and more broadly for other emerging technology domains.

Rethinking Technology Emergence: Measurement, Policy, Insights

ABSTRACT. Sesion Title: Rethinking Technology Emergence: Measurement, Policy, Insights Alan Porter&Jan Youtie

New considerations in defining new emerging technologies

Technology and innovation have been significant since understanding the knowledge as an additional factor of production. From the perspective of evolutionary economics, this understanding force policymakers to forecast or foresight technological improvements and potential innovations in rapidly changing environment. This rapid shift of technological developments has created emerging technology phenomenon which was tried to be conceptualized and modeled by researchers in innovation policy domain. Although there were no consensus on the definition and model, limited number of publications were seen in the literature. In this study, I aim to raise some questions for conceptualizing emergence phenomenon and identifying new emerging technologies with an applicable way. Firstly, for conceptualizing emergence, I gathered some criteria by considering term “emergence” with the help of 150-year discussions in philosophy of science and complexity domains. Because conceptualizing emergence is a future-oriented approach, I try to emphasize their explanatory power by using existing literature and the findings of current studies. Secondly, for identifying new emerging technologies with an applicable way, I try to categorize the conceptual dimensions as applicable or not-applicable. Based on these findings, I propose a model for visualizing conceptually the criteria and their possible effects on conceptualizing and identifying new emerging technologies. Finally, I want to discuss this model with participants for validating its internal and external validity.

Dispatches from the Indicators of Technology Emergence Project

ABSTRACT. Background

Measuring and understanding science and its translation to technological innovation greatly concerns technology management and ST&I policy (Wagner and Popper, 2003). Such knowledge is vital to inform R&D prioritization. The focus of the Technology Emergence Project is on developing indicators of frontier R&D topics in the text of titles and abstracts of scholarly publications, their activity patterns, and the “players” (researchers or inventors, R&D organizations, and countries) engaging those topics.

"Emergence" has roots in evolutionary theory, addressing phenomena with properties unlike those found previously (Goldstein, 1999; Corning, 2002). Systems science also pursues emergent phenomena in complex systems (Mitchell, 2009). Our focus is on distinguishing and measuring topics within target technological domains that evidence accelerating R&D activity. A key theme to our work is the connection between tracking historical progression of an S&T area and anticipating future prospects. The National Research Council (2005) suggested the need for “tech emergence indicators” to provide early warning, but this is difficult to do. Expert opinion methods are cumbersome, expensive, and fraught with “within-the-box” thinking. Trend-based forecasting fares best with incrementally progressing technologies, epitomized by Moore’s Law for semi-conductor gains (Roper et al., 2011), but are not well-suited to highly novel advances. “Hot papers” (e.g., the Web of Science indicator of papers heavily cited right after publication) can flag novel Science and Technology (S&T) breakthroughs, but don’t systematically track emergence.

Rotolo, Hicks, and Martin (2015) review literatures addressing technological emergence, noting variations in approaches, concepts, and definitions. They show marked growth in scholarly attention to emerging technologies in scientific literature and popular coverage, particularly since 2003. They note various approaches to measurement, narrowing to the detection and analysis of emergence in S&T domains. They mention several scientometric approaches that utilize abstract records drawn from global ST&I databases (e.g., Boyack et al., 2014). However, they express concern about the lack of connection between such empirical formulations and conceptual models of technological emergence. They offer their conceptualization of the key five attributes of an emerging technology to take into account in deriving indicators.


Since 2011, several Project members have been involved in the development of a technical emergence indicator: “e-score.” The roots trace back to the US Intelligence Advanced Research Projects Activity (IARPA) Foresight and Understanding from Scientific Exposition (FUSE) Program that funded emergence indicator development. Newman and Porter participated in conceptualizing emergence with the SRI group on the FUSE program (Alexander et al. 2012). A key issue that arose is measurability. Conceptual foundations for empirical indicators are clearly desirable. However, many concepts do not translate well to “the data.”

The resulting e-score operationalizes components embedded in these conceptual definitions of technological emergence: novelty, growth, community, and persistence (Carley et al., 2018). The e-score indicator is applied to text from titles and abstracts of Web of Science publications which have been extracted using Natural Language Processing and multistage data cleaning and consolidation. The project then applies the e-score to multiple broad-based S&T domains including nanotechnology, synthetic biology, autonomous vehicles, and dye sensitized solar cells to determine the extent of consistency in the e-scores.

Outcomes and Impacts

This presentation will discuss outcomes of three aspects of the project. 1. Testing: There is little written about how to test and validate assessments of emerging technologies, except allowing for the passage of time. This work puts forth a validation test process involving perturbing measures of the components that comprise the technological emergence indicator. The results suggest that the indicator is not very sensitive to changes in assumptions about novelty, growth and community. However, it is very sensitive to changes in assumptions about term persistence, treatment of unigrams and multigrams (which for some domains vary by country), and handling of “standard” stopword terminology (Liu and Porter, under review). 2. Emergence as an explanatory variable: This study empirically examines the association between the extent of emerging technological ideas in a scientific publication and its future scientific impact measured by number of citations. Metadata of scientific publications in three scientific domains are analyzed: Nano-Enabled Drug Delivery, Synthetic Biology, and Autonomous Vehicles. By employing the e-score -- the extent to which the publication contains emerging technological ideas in each domain is measured. Then, the size and statistical significance of the relationship between the publication-level technological emergence score and the normalized number of citations accruing to the publication is estimated. The analysis shows that the degree to which a paper contains technologically emerging ideas is positively and strongly associated with its future citation impact in each of the three domains. An additional analysis demonstrates that this relationship holds for citations from other publications, both in the same field as, and in different fields from, the scientific domain of the focal publication. A series of tests for validation further support the argument that the greater the extent to which scientific knowledge (a paper) contains emerging ideas, the bigger its scientific impact (Kwon et al., under review). 3. Contest: One approach for disseminating the results of the project, as well as obtaining new ideas for identifying emerging topics in scientific literature, is through a prize contest. Prize contests have been widely used in recent years in various government programs (Kay, 2012). This project puts forth a contest to challenge participants to devise a repeatable procedure to identify emerging R&D topics within a designated S&T domain (e.g., synthetic biology). The data resource to be mined is a Web of Science extracted publication dataset provided to participants, on the designated S&T domain. The key criterion is based on who best predicts topics that are notably active in the following two years of research? More than 20 scholars have expressed interest in the contest to date.


Alexander, J., Chase, J., Newman, N.C., Porter, A.L., Roessner, D. (2012). Emergence as a Conceptual Framework for Understanding Scientific and Technological Progress, PICMET, Vancouver, 2012.

Boyack, K.W., Klavans, R., Small, H., Ungar, L. (2014). Characterizing the emergence of two nanotechnology topics using a contemporaneous global micro-model of science, Journal of Engineering Technology Management 32, 147–159.

Carley, S.F., Newman, N.C., Porter, A.L., Garner, J. (2018). An indicator of technical emergence, Scientometrics, 115 (1), 35-49.

Corning, P.A. (2002), The re-emergence of ‘emergence’: A venerable concept in search of a theory, Complexity 7 (6), 18-30.

Goldstein, J. (1999), Emergence as a construct: History and issues, Emergence, 1 (1), 49-72.

Kay, L. (2012). Opportunities and challenges in the use of innovation prizes as a government policy instrument. Minerva, 50(2), 191-196.

Kwon S., Liu, X., Porter, A.L., Youtie, J. (under review). Scientific Research with Emerging Technological Idea Has a Greater Scientific Impact.

Liu, X., Porter, A.L. (under review). A 3-dimensional Analysis for Evaluating Technology Emergence Indicator.

Mitchell, M. (2009). Complexity: A guided tour. Oxford University Press

National Research Council. (2005). Measuring Research and Development Expenditures in the U.S. Economy. Washington, DC: National Academies Press.

Roper, A.T., Cunningham, S.W., Porter, A.L., Mason, T.W., Rossini, F.A., Banks, J. (2011), Forecasting and Management of Technology, 2d edition, New York: Wiley.

Rotolo, D., Hicks, D., Martin, B.R. (2015), What is an emerging technology? Research Policy, 44, 1827-1843.

Wagner, C. S., Popper, S. W. (2003). Identifying critical technologies in the United States: A review of the federal effort. Journal of Forecasting, 22(2‐3), 113-128.

08:30-10:00 Session 6D: Patents & Regulation
Location: GLC 225
Institutional ownership of academic patent in Japan: empirical analysis of its impact on academic research

ABSTRACT. Since the enactment of the Act on the Promotion of Technology Transfer from Universities to Private Industry (the “TLO Act”) in 1998, Japan has implemented various policy measures to promote university–industry collaboration (UIC). Following the incorporation of national universities in 2004, the number of joint research projects substantially increased. In addition, the institutional ownership of academic patents started, while national universities before incorporation could not claim patent ownership (professor ownership). Accordingly, a surge of academic patent applications are found after 2004. Through joint research with companies, university faculty can gain a deeper understanding of R&D activities within industries that are related to their research interests and develop research agendas with concrete goals for innovation, such as designing a new product or a new manufacturing process. In turn, faculty exposure to industry activities can potentially increase the likelihood of universities and other public institutions that conduct R&D activities, leading to industrial innovations. However, research activities in universities should not be limited only to the areas that leads to direct applications at industry. While universities, public research organizations, and corporations play individual roles in the system of national innovation, universities are generally expected to pursue research in basic areas in which companies are ill-equipped to address. The role of universities is to open up academic frontiers and produce research results with long-lasting and large spillover effects over a wide range of disciplines. Stronger university incentive to engage in UICs may affect research, giving it a more practical focus and possibly leading to the neglect of basic research, which is the natural domain of university research. Some research suggests that although the number of patent applications by universities increased in the US following the enactment of the Bayh–Dole Act in 1980, the quality of the patents declined, because excessive focus on commercializing research results led to a neglect of basic research, which universities should certainly focus on. Drawing on a novel dataset linking research article and patent in Japan, this study provides a quantitative analysis of how UICs have changed since UIC policies were introduced in the late 1990s. Specifically, we linked the data of research papers (Scopus by Elsevier) and patent data (IIP patent database) at author/inventor level for entire population of researchers at Japanese research organizations to see how academic discipline and technology field interlinked at individual (academic) researcher level. A disambiguation of patent inventor work has been conducted for all academic inventors in Japan identified in IIP Patent database, covering all patent application, published by Japan Patent Office. Based on rare name derived from telephone dictionary, training datasets are constructed. Rare names supposed to be relevant information to construct an unbiased training set. We fit a number of discrimination models and clustering methods, which are standard in the field of machine learning, to our rare name based training set. Then the list of academic inventors are linked with Web of Science (WoS) research article authors by approximate match of name and affiliate information. Furthermore, we have constructed the linkage information of NPL (non-patent literature) cited by academic patents, with WoS research articles. The dataset allows us to look into both embodied (author inventor) and disembodied (citation) knowledge flow from academic publication, and we could systematically evaluate the impact of institutional ownership change of academic patents in 2004 on academic outputs (publication and patents). Existing literature of the academic patenting on publication shows some mixed results In general, complementary relation between patents and papers are found in life science field in the United States (Azoulay et. al, 2009), and in German samples (Crarnitzki et. al, 2007). However, the difficulty involved in these empirical approach is to control for endogeneity in patenting and publication, in a sense that a talented scholar could both better than her colleague. With this regards, Japanese policy experience of incorporation of national university in 2004 (and institutional ownership of academic patents) allows us to use such policy shock as an exogeneous instrument for identifying the causality of academic patents to publication. Another feature of this study is that we could evaluate the impact of scientific research, not only by citation (paper and patents), but also by patant to patent citation, based on the information of author inventor linkage. A broader picture involving three-way citation routes, paper to paper, paper to patent (NPL citation of patent) and patent to patent (PL citation of patent) leads to the understand consequences of academic patenting more precisely, not just by (quality weighted) publication performance, but also the changes of research scope of academic researchers.

Assessing how the Colocation of Science and Technology Benefits Innovation
PRESENTER: Stephen Carley

ABSTRACT. The interplay between science and technology motivates continuous discussion. The scale at which the former impacts the latter prompts the current inquiry. Scholars like Francis Narin have demonstrated a strong and increasing connection between basic science and the development of technology in the United States, but to what degree is this linkage spatially mediated? We seek to assess how the advancement of technology depends on local science. Tip O’Niell is famous for the phrase “all politics is local.” Can the same be said about the location of science with respect to the technology that relies on it? That is, what role does proximity between science and technology play in terms of how the former influences the latter? Can technology successfully develop independent of homegrown science, or does technological advancement depend on science at the local level? Using patents as a proxy for technology and publications as a proxy for science we measure the distance between inventors and the location of the scientific authors they cite and display the same on a heatmap at the zip code level—i.e. we pull inventor zip codes from fielded patent data and the zip codes of scientific authors from Web of Science publication data. A general proxy for proximity can be whether a given inventor and the scientist he or she cites are collocated in the same metropolitan statistical area. To assess whether answers to the questions we pose depend on technology space results are filtered at IPC and disciplinary levels. With benefit of a time slider we show changes over time in our heatmap. It is anticipated that the colocation of science with industry will significantly benefit the latter, but perhaps not for all technologies. Demonstrating the degree to which specific technology spaces are relatively more dependent on localized science can only serve to better inform policy.

The effect of patent pledge policy on patent pledges and regional innovation

ABSTRACT. Background and significance Pledging patents as loan collateral is an emerging financing innovation to solve the ‘finance gap’ imperfections, presenting an upward trend in many countries, such as the US, the UK and China. This financing tool is considered important to mitigate financial constraints of small and medium-sized firms (SMEs), and thus benefits their survival and development. Broadening the access to finance for SMEs has the potential to boost knowledge-driven growth. In the US, there were only a negligible number of pledged patents in 1980, while this number skyrocket to nearly 36,000 per year recently. In China, from 1996 to 2006, only 682 patents were pledged, while this number grew to 4,177 and the total amount of patent-backed loans has reached 72 billion in 2017. This practice recently aroused increasing attention. The previous studies, mostly based on cases in the US, suggested that loans backed by intangibles including patents reflected a thoughtful and up-front screening and selection mechanism developed by lenders for mitigating risks. This mechanism is often used to explain patent or intangible pledges. The quality and redeployability of collateral, collateral’s owners and the market-related factors in terms of the collateral resale are related to the chance that the patent is pledged. However, a rational mechanism lender developed to select collateral and borrowers in the countries with a strong legal system and developed market conditions seems insufficient to account for the booming pledged patents in the developing environment of China. It is not surprising that patent pledges have become prominent in advanced economies. The reason is that these countries often have a relatively mature environment for patent pledges, e.g., increasing liquid market for intangibles, stronger legal protection for Intellectual Property Rights (IPR) and creditor rights and advanced financial system. However, the surge of patent pledges in emerging economies, such as China, seems counterintuitive due to the lack of required legal and market environment. Because of this, the growth of China’s patent pledges is considered a unique phenomenon. Data The sample consists of invention patents whose first applicants are companies granted from 2002 to 2015. The patent data is derived from the State Intellectual Property Office of China (SIPO). We keep invention patents whose first applicants are companies granted between 2002 and 2015. Invalid patents due to the delayed payment of maintenance fees are excluded since these patents cannot be pledged according to the regulation. There are 340,369 sampled patents, yielding 1,059,024 patent-year observations. Among the sampled patents, 4,372 patents have been pledged during the sample period. There are 80,659 unique companies that are the first applicants of the sampled patents and 1,928 companies had pledged patents. We collect data on companies from diverse sources. We identify the year of cities’ policy adoption and specific measures based on local governments’ official documents. Using a set of keywords, we collect these documents by searching governments’ websites, online news, and China National Knowledge Infrastructure (CNKI) database. Other control variables at the regional level are from China Statistical Yearbook, China City Statistical Yearbook and so forth. Method We use different in differences strategies, instrumental variables and matching strategies to explore the relationship between cities’ policy adoption and the probability that patents are pledged using difference-in-differences strategies, propensity score matching, and instrumental variables estimation. We dig into the specific mechanisms of the effect of policy adoption, considering the timing of policy introduction before or after provincial governments, implementation of four specific measures, the relationship between governments and lenders, and local technology market and IPR protection. Basic findings 1. From 2002 to 2015, 78.76% of the firms that pledged patents in China are SMEs, and 47.57% of them are national high-tech enterprises. This implies that SMEs and high-tech firms are the main beneficiaries of patent-backed loans. Patents owned by smaller and younger firms are more likely to be pledged. 2. The market failure regarding patent pledges can be (at least partly) addressed by governments’ policies. After policy adoption, the probability that patents are pledged increases by 47.9% to 124.2%, relative to that of patents whose owners in a city that has not adopted such a policy. In the cities that provided government funding, the probability that patents are pledged is 2.396 times as much as that of patents whose owners in the cities that did not provide financial support. 3. The effect of policy adoption on the probability that patents are pledged is found for patents whose owners in the city that adopted policies before or after their superior governments. 4. Policy adoption increases the probability that patents are pledged to banks by 25.9% and the probability that patents are pledged to non-bank pledgees by 62.5%. 5. In the cities with a more active trading market for technology, the effect of the policy is intensified. 6. Compared to non-bank pledgees, banks are more likely to provide patent-backed to older, larger and non-state-owned firms. 7. Patent pledge policy increases regional patent applications by 18% to 35.2%.

08:30-10:00 Session 6E: Measuring Innovation in Africa

Global South

Location: GLC 222
Towards participatory innovation measures and indicators: exploring the value of participatory visual methods
PRESENTER: Il-Haam Petersen

ABSTRACT. Introduction

Much of the research that informs key innovation measures and indicators is based on highly industrialised economies, and tends to focus on innovation in specific sectors, in formal businesses, and on technology-based innovation. In developing country contexts, the informal sector forms a significant part of the economy, and innovation tends to take more non-technological and incremental forms. This raises concern about the suitability of current standardised measures and indicators, which are crucial to assess progress and ensure that the proposed policy objectives are met. How do our measures and indicators take into account and reflect the lived realities and forms of innovation of informal businesses? These dimensions are crucial for understanding learning and innovation in economies with relatively large informal sectors.

However, we have very little experience in researching and measuring innovation in these settings. The available evidence suggests that the traditional methods of business innovation surveys may not be suitable to elicit the kind of data and indicators on innovation that we need.

This paper therefore turns attention to the research tools we use to gather empirical evidence for developing measures and indicators.

We argue that participatory visual methods, which are traditionally used for community development and in public health research, hold promise for linking with such hard-to-reach social groups. Such arts-based methods are thus becoming popular as participatory research tools for producing in-depth, rich data on underexplored topics. Participatory visual methods are promising for researching learning and innovation because they: - Are based on community-led approaches that recognise community-based actors’ rights in the research and values and legitimises indigenous/local knowledge. - Provide tools for identifying ‘community assets’ rather than focussing on ‘community deficits’ (Delgado, 2015: 129). - Enable a collective process that may catalyse social change, and participants are encouraged to co-produce knowledge with researchers rather than act as passive subjects. - Provide tools for researching social issues that may be difficult to talk about and social situations that may be difficult to access, thus potentially providing new insights and ways of thinking that are not possible through traditional research methods. - Provide tools that may be empowering. Participants are treated as research partners, and are encouraged to tell the personal stories they choose to tell and that are shared through platforms that they choose. Learning to use a camera, tablet or video recorder can also be empowering. - Use arts-based methods, and thus participants who do not speak a dominant language or are illiterate can be included as co-researchers. - Enable the production of research outputs that can be easily translated, thus addressing a major challenge in research and engagement.

In this paper, we reflect on our experimentation with the use of two types of participatory visual methods that are gaining popularity in research, digital storytelling and photovoice. Both methods involve the use of easy-to-learn audio and visual recording equipment, and have been shown to be useful for researching economic activities in informal settings such as townships in South Africa.


In September and December 2018, we conducted two workshops, one digital storytelling and one photovoice workshop, with informal businesses in the services sector and community-based actors in one of the largest township areas in Cape Town, South Africa. The digital storytelling workshop was held over five days and focused on understanding the process of innovation. The guiding topic was: ‘tell me a true story of a time when you have done something differently in your business and what happened”. The photovoice workshop focused on exploring knowledge flows between formal and informal actors, and key mechanisms and strategies for learning. The workshops, which included seven to 10 participants, are similar to focus groups but involve a more lengthy and intensive process, allowing for in-depth and participatory discussion. Through intense engagement over several days, participants, guided by the facilitators, scripted and recorded their own stories on how they learn and innovate. The outputs of the workshops are thus co-produced. For the digital storytelling workshop, each participant produced a script and short (three to five minute) video recording of his/her story. The photovoice process involved participants taking photographs in their community that related to the topic. The participants, guided by the researchers, produced a set of individual photo-stories and a collective story including a series of photographs with captions.

We also conducted a set of semi-structured in-depth interviews with the participants, and intermediaries and formal knowledge producers identified through the workshops as playing an important role in learning in the local setting.

Proposed analysis and significance of the research

The stories and workshop discussions will be analysed to identify forms of innovation and learning, specific strategies and mechanisms used for learning, key sources of information and other resources, types of resources required, drivers of and obstacles to innovation and so on.

We will show how these qualitative methods provide useful data and ways to measure the tacit forms of learning that are difficult to research.

The workshop discussions and outputs are being used by the research team to inform the development of a conceptual framework, methodology and instruments for a pilot study on measuring innovation in the informal sector in South Africa. We find that digital storytelling and photovoice workshops are useful as complementary methods.

In this way, we aim to contribute to building an understanding of how participatory visual methods can facilitate science, technology and innovation measurement, and the development of indicators that are grounded in the lived realities of actors that have been traditionally marginalised from the formal national system of innovation (NSI), from the formal economy and from decision-making processes.

Special Session Presentation: Linking Work Organisation with Capacity Development and Innovativeness in Microenterprises in Nigeria
PRESENTER: Oluseye Jegede

ABSTRACT. Special Session Title: Towards a Measurement Programme for Innovation to serve the developmental needs of Sub-sahara Africa

Special Session Presentation: Linking Work Organisation with Capacity Development and Innovativeness in Microenterprises in Nigeria

Mainstream measurement of innovation is informed by the Frascati manual (OECD 2015) and the Oslo Manual (OECD, 2018). The Frascati manual considers innovation as an input informed by expenditures on research and development (R&D) as well as personnel involved in R&D activities such as scientists and engineers with PhD degrees, quantity and quality of researchers, number of publications, and number of patents, amongst others while the Oslo manual considers innovation as a process and as an outcome. The innovation process involves capturing activities such as training, acquisition of patents and other technical know-how, acquisition of machinery, equipment, hardware or software, acquisition of buildings, feasibility studies, pilot plant testing, amongst others that may eventually lead to a tangible outcome. While an innovation outcome signifies new or significantly improved products (goods and services) and new or significantly improved productions processes/methods. Because of the wide acceptability of these manuals, there has been bias in the literature which has led to the assumption that very small enterprises, which constitute the major players in Africa’s economy, are non-innovative. Hence, have been excluded from national innovation surveys which have also had grave consequences on policy attention these microenterprises receive from government. This presentation will build on recent previous studies emerging from developing countries, which provides evidence that micro enterprises innovate despite their numerous challenges. This presentation will provide evidence from three sub-sectors (agro-processing, clothing and ICTs) from Nigeria to establish this point. Using original data covering 120 interviews was used in the study. The study argues that the mainstream literature fails to capture important dynamics and practices that are central to innovation in Africa by excluding very small firms from innovation national surveys. Surveys were carried out on both employers and employees in each enterprise. The study sampled approximately equal numbers of formal enterprises and informal enterprises during the survey. The data collected was analysed both quantitatively and qualitatively. And the results presented in Tables. The result of showed a strong association between innovation and work organisation. A new knowledge advanced in the presentation is the specific role work organisation/enterprises routine play in the build-up of competence for innovation in these micro enterprises. The study explored alternative approaches to capturing innovation other than the conventional metrics. The soft component of innovation was explored. This involve “Learning by Doing”, Learning by Using”, Learning by Interacting”, Learning by Searching” “Learning by Producing” and “Learning by Imitating”. This soft component of innovation form part of the daily routine in the work space in most Microenterprises. The paper outline how daily work activity/business routine help enterprises accumulate skills and knowledge. The study found out that majority of the work schedule in the enterprises involved: learning new things, solving unprecedented problems, applying employee’s ideas, sharing work-related information with colleagues, suppliers, customers. The study also gathered that majority of the enterprises recorded that their daily job routine involves rotating tasks amongst employees thereby fostering knowledge and information sharing. It was also largely reported by the employees that they acquired most of their skills through daily work organisation. About 71% of the enterprises implemented process innovation this was captured as new processes or technologies. About 68% of the enterprises implemented product innovations captured as new products or services. The study found out that most of the innovations were modifications of what technologies (product and processes) from the formal sector (used by larger firms). Another importance discovery in this study was that the fact that the prevalence of innovations that represent adaptations or modifications of technologies was many folds more than the prevalence of innovation that were a result of adoption of technologies. This is to say that the enterprises apply their own knowledge and skills in the innovations. They adapt existing technologies to soothe their local needs. Another interesting finding was that the source of the knowledge and skills put to use wasn’t really from formal institutions as only about 35% of the employees had secondary school education.

Acknowledgement: The seed funding provided by the Africalics Network is hereby acknowledged. The opinions presented in this paper represent those of the authors not necessarily that of the funder.

Innovation for inclusive development (IID): Learning from informal settings in the Global South

ABSTRACT. Session title - Towards a measurement programme for innovation to serve the developmental needs of sub-Saharan Africa Session chair - Glenda Kruss (Center for Science, Technology and Innovation Indicators, CeSTII, South Africa)

ABSTRACT The informal economy accounts for a large share of the economic activities in the Global South. The livelihoods of millions of individuals depend on these activities, yet knowledge and understanding about the underlying processes that sustain this large share of the economy is very meagre leading to an invisible, forgotten or un/non-observed phenomena (Feige, 1990; Fourie, 2018; OECD, 2002). When moving further, questioning about the nature and dynamics of innovation processes in the context of the informal economies, the answers get extremely reduced. The intersection of these two fields of knowledge, the informal economy and innovation studies, leads to a substantially underexplored area (Kraemer-Mbula & Wunsch-Vincent, 2016).

Complexity encompasses the quest for understanding innovation in the informal economy, and their relationship with inclusive and sustainable development. The informal economy shapes a very heterogeneous landscape. Several authors have attempted to set the different types of activities that shape the informal economy (Portes, 1995; Tokman, 1991). Informal activities are of vital importance, not only because of their role on the provision of livelihoods, in which more than 61% of the global employed population (2 billion people) get their livelihoods in the informal sector, but, relatedly, because the fulfilment of SDGs is strongly dependent on tackling this phenomenon (International Labour Organization, 2018).

The informal economy in Sub-Saharan Africa is the second largest in the world, after Latin America and the Caribbean. Within these continents, large variations coexist: within the African region it ranges from 20-25% of GDP in Mauritius, South Africa, and Namibia, to 50-65% in Tanzania and Nigeria (International Monetary Fund, 2017). Differences in Latin America and Caribbean (LAC) countries were enormous: from 70% of GDP in Paraguay or Nicaragua to 20% in other countries (Vuletin, 2008). Uruguay is one of those LAC countries with lower share of informal economy. Yet, within the country, disparities are large between regions and between sectors. In the north of the country, the share of social security uncovered workers is larger than 40%, while in the South is less than 25% (MTSS, 2016). In spite of the fundamental relevance of these informal activities on the life of billions of people, the understanding of innovation dynamics in the informal settings is quite scarce. How does innovation operate in these contexts? How does the interaction between the formal and the informal function, and how does it affect innovation processes? How different innovation processes in the informal economy are compared to what we know from the formalized economy? What are innovation diffusion mechanisms, and what are the intervening actors? What are the incentives for innovation? (de Beer, Fu, & Wunsch-Vincent, 2013) are some of the questions that drive this research project.

The analytical framework utilized to find that space of intersection between the informal economy field, and innovation for development studies is the Local Production and Innovation Systems (LIPS) framework (Cassiolato, Martins Lastres, Szapiro, & Matos, 2017), which involves the following dimensions: (i) a characterization of the specific LIPS being studied; (ii) the socio-economic features of the territory of the LIPS; (iii) LIPS’ production and innovation capacity building processes; (iv) the national and international context of the LIPS; and, (v) the wider policy environment affecting the LIPS (Lastres & Cassiolato, 2005).

Advancing the understanding of innovation processes in the context of informal settings require to take into consideration: (i) measuring the social contexts/outcomes of innovation, (ii) not all transfers in the informal sector are mediated by a market (transfers of goods/ services may happen for social or other reasons); (iv) actors other than firms, including entire communities, need to be studied (Charmes, Gault, & Wunsch-Vincent, 2016)

The LIPS approach guides the two case studies (Yin, 2003), the one in South Africa and the one in the Northeast region of Uruguay. In South Africa, the case study locates in the Msunduzi Local Municipality, in the KwaZulu-Natal Province. On a regional scale, the municipality is situated at the junction of an industrial and agro-industrial corridor. It is a region with around 35,500 people, mostly between 18-64 years, with a large share of them being economically inactive.

The Northeast region of Uruguay is the one with lower per capita GDP, and lower overall development indicators (IDH and economic indicators, among others). It is also the region with higher productive concentration around agriculture, (Rodríguez Miranda, Galaso, Goinheix, & Martínez, 2017). This case study draws on a previous work about innovation capabilities in micro- and small-enterprises in the Northeast region of the country (Bortagaray, 2017).

This work attempts to contribute to the learning and understanding of some of the linkages between inclusion, innovation and development in the context of the ‘underexplored (invisible, understudied) informal economy’.


Bortagaray, I. (2017). Análisis de capacidades y oportunidades de innovación de las MIPYMES de la región Noreste. Retrieved from Tacuarembó: Cassiolato, J. E., Martins Lastres, H. M., Szapiro, M., & Matos, M. (2017). Local production and innovation systems in Brazil: A balance of 20 years. Retrieved from Rio de Janeiro: Charmes, J., Gault, F., & Wunsch-Vincent, S. (2016). Formulating an Agenda for the Measurement of Innovation in the Informal Economy. In E. Kraemer-Mbula & S. Wunsch-Vincent (Eds.), The Informal Economy in Developing Nations: Hidden Engine of Innovation? : Cambridge University Press. de Beer, J., Fu, K., & Wunsch-Vincent, S. (2013). The Informal Economy, Innovation and Intellectual Property - Concepts, Metrics and Policy Considerations. Retrieved from Feige, E. (1990). Defining and estimating underground and informal economies: The new institutional economics approach. World Development, 18(7), 989-1002. Fourie, F. C. v. N. (Ed.) (2018). The South African Informal Sector: Creating Jobs, Reducing Poverty. Cape Town: HSRC. International Labour Organization. (2018). Women and men in the informal economy: a statistical pricture. Retrieved from Geneva: International Monetary Fund. (2017). The Informal Economy in Sub-Saharan Africa. In International Monetary Fund (Ed.), Regional Economic Outlook: Sub-Saharan Africa. Restarting the Growth Engine. Washington D.C.: IMF. Kraemer-Mbula, E., & Wunsch-Vincent, S. (Eds.). (2016). The Informal Economy in Developing Nations: Hidden Engine of Innovation? : Cambridge University Press. Lastres, H. M. M., & Cassiolato, J. E. (2005). Innovation Systems and Local Productive Arrangements: new strategies to promote the generation, acquisition and diffusion of knowledge. Innovation: Management, Policy & Practice, 7(2), 172-187. MTSS. (2016). Estudios sobre Trabajo y Seguridad Social. Retrieved from Montevideo: OECD. (2002). Measuring the non-observed economy: A Handbook. Retrieved from Paris: Portes, A. (1995). En torno a la informalidad: Ensayos sobre teoría y medición de la economía no regulada. México D.F.: FLACSO. Rodríguez Miranda, A., Galaso, P., Goinheix, S., & Martínez, C. (2017). Especializaciones productivas y desarrollo económico regional en Uruguay. Retrieved from Montevideo: Tokman, V. (1991). El Enfoque PREALC. In V. Tokman (Ed.), El Sector Informal en América Latina, dos décadas de análisis. . México, D.F.: Consejo Nacional para la cultura y las artes. Vuletin, G. J. (2008). Measuring the Informal Economy in Latin America and the Caribbean. Retrieved from Yin, R. K. (2003). Case studies research: designs and methods (Second ed. Vol. 5). Thousand Oaks: Sage Publications.

Towards a measurement programme of innovation in the informal sector in Africa
PRESENTER: Nazeem Mustapha

ABSTRACT. Session title - Towards a measurement programme for innovation to serve the developmental needs of sub-Saharan Africa Session chair - Glenda Kruss (Centre for Science, Technology and Innovation Indicators ( CeSTII), Human Sciences Research Council, South Africa)

Article title - Towards a measurement programme of innovation in the informal sector in sub-Saharan Africa *Mustapha, N., *Jegede, O., *Petersen, I., †Bortagaray, I. and *Kruss, G.

* Human Sciences Research Council, South Africa. † University of the Republic, Uruguay

ABSTRACT presented at the 2019 Atlanta Conference on Science and Innovation Policy Background In this paper, we argue that the traditional measurement programmes are not suited to developing economies on the African continent, where the majority of countries have economies that are largely informal, combined with underdeveloped measurement structures. If the science, technology and innovation (STI) measurement programmes that African policy leaders install in their countries follow those designed for developed nations, then the policies in these countries will, modens podens, tend to follow those of developed nations too, and vice versa. By a measurement programme we mean a process, which includes a measurement framework that sets out a conceptualization of innovation, and develops and standardises definitions, instruments and methods for collecting and analysing data. Traditionally, such measurement programmes are considered at the nexus of policy makers, academics and measurement professionals. This paper puts forth an argument for a measurement programme of innovation in the context of local economic development as an imperative for Africa. Rationale Measuring innovation is important for any country, to develop indicators that inform the design of industrial policy, and monitoring and evaluation to enable effective implementation (Charmes, Gault & Wunsch-Vincent, 2016). Measuring innovation in formal enterprises is a complex and difficult task - one that has evolved over decades through the iterations of the Frascati and Oslo Manuals. Over the years, a well-established measurement programme for research and development and innovation has emerged, promulgated by the OECD mainly. Many countries in Africa and the global South have adopted and adapted the measurement frameworks and instruments derived from these programmes, and indeed, started collecting data within these frameworks. To what extent are these standardised frameworks and instruments aligned to policy frameworks and development needs at the regional and national levels? The measurement challenges arising in informal contexts are even more daunting, and there is little international precedent. A foundation has been laid by experimental African attempts, including sectoral studies such as that done in Senegal (Konte & Ndong, 2012); in the Nigeria’s Information and Communications Industry (Jegede et al., 2017; Jegede and Jegede, 2018); a set of case studies and surveys of innovation in the informal economy (Kraemer-Mbula & Wunch-Vincent, 2016). None of these, however, has focused on a measurement programme designed to serve continental needs. Is there space for the inclusion of informal sector actors in the measurement programme? Conceptual foundations Lundvall and his colleagues at Aalborg University IKE research group provided a broadened version of the National Innovation System beyond science and technology organisations and firms/industries to include other features such as the macroeconomic and regulatory context, education and training systems, infrastructure, market conditions, networks, inter alia (Lundvall, 1988; 2007; Lundvall, et al., 2002). While other scholars further delineated the concept of innovation system into smaller units such as the regional innovation system (Cooke et al., 1997; Cooke, 2001; Asheim & Gertler 2005), sectoral innovation system (Breschi & Malerba, 1997; Malerba, 2002) and technology innovation systems (Nelson & Nelson, 2002). However, none of these models responds directly to the challenges of the informal sector in Africa. Some approaches have moved closer to answering these needs, see for instance the works by Breschi & Lissoni (2001) on knowledge spillovers and local innovation systems, Mytelka & Farinelli (2000) on local clusters, innovation systems and sustained competitiveness, Asheim (1996)’s work on industrial districts as ‘learning regions’ as well as the very recent work by Markusen (2017) entitled “Sticky places in slippery space: a typology of industrial districts”. Whilst the literature provides different metrics and indicators for capturing innovation, none of these are based on data or insights from Africa. Moreover, the approaches used all lean towards the formal sector. The proposed measurement programme draws on the local innovation and production systems (LIPS) approach advanced by Cassiolato, Lastres and their colleagues at RedeSist, based on their experience in Brazil (Cassiolato et al., 2003; 2017). The LIPS methodology brings together innovation systems and development thinking (Cassiolato et al., 2014). As Cassiolato and colleagues argue, innovation processes are shaped by social, economic and institutional contexts, which necessitates an analysis of spatial dynamics and the local level. The LIPS approach provides useful tools for identifying and analysing the different components of the production value chain, linkages between them and how wider social, economic and institutional contexts influence these components. Taking into account specificities of the informal sector in countries, the measurement programme described in this paper is based on a conceptual framework developed through grounded theory that builds on and extends existing approaches to measuring innovation, including the LIPS methodology and the Oslo manual framework. Our input methodological techniques include ethnographies, participatory appraisals, in-depth interviews, key informant interviews combined with a quantitative survey. The paper will attempt to draw on preliminary data collected from participatory workshops with informal business people to illustrate the suitability of this approach. Institutionalisation of measurement in the African context Measurement requires the institutionalisation of structures for a programme of measurement to be sustainable. In the African context, such institutional structures are minimally existent or weak. This paper proposes a methodological approach towards developing an innovation measurement programme for the informal sector that compensates for the lack of institutionalised measurement structures in Africa. In order to further this end, we propose the use of a standardised quantitative tool based on and extended from the Community Innovation Survey series adapted to informal settings. Furthermore, in contrast to other studies, our methodological approach is not dependent on the availability of lists to use as sampling frames, which may or may not be available from informal institutions, within a specific trade or sector. Instead it adopts a methodology for listing informal businesses using a local (spatial) approach, adapted from the health sciences. We present some preliminary illustrative results from one local region in South Africa from this approach. Some implications for policy, practice and theory The current tools for measuring innovation are not suitable for informing on the majority of economic activity on the continent. The measurement programme adopted from developed country programmes themselves are unsuitable as an effective policy . In order to have a robust representation of the state of innovation in African countries, there is need for measurement to go beyond formal businesses alone and extend towards informal businesses and households, the public sector as well as civil society. For the determination of measurement needs to be inclusive, we suggest actors involved in innovation in the informal sector need to be part of the innovation policy agenda setting. There is a need for a better understanding of the policy formulation process in the context of innovation.

08:30-10:00 Session 6F: National Level Policy
Location: GLC 324
Policy learning as a key drive of science, technology and innovation policy in China——How the Concept of National Innovation System was introduced and developed in China

ABSTRACT. In the late 1990s, a few years after the national innovation system theory became popular in the world, it was accepted by China at the policy level. Since then, building a national innovation system with Chinese characteristics has become an important goal of China's scientific and technological system reform and innovation development. In China's future development, strengthening the national innovation system is still an important goal of China's scientific and technological development. The purpose of this paper is to study the process of the introduction, acceptance and development of the concept of national innovation system in China's policy circle through historical research, policy text analysis and prospect analysis.

1. The concept of national innovation system was introduced into China's policy circle With China's reform and opening up, China began to reform its science and technology system in 1985 to promote the integration of science and technology with economic development. The introduction of the national innovation system concept in the middle period of 1990s has opened a new vision of the reform of science and technology system and greatly promoted the reform of science and technology system. Soon after the concept of national innovation system became popular in the world, Chinese scholars began to introduce and study it,but the scholars' research did not lead to acceptance by policy community . Based on literature analysis and expert interview, this paper Identifies the two facts is crucial for the concept of the national innovation system be accepted by policy community : (1) the International Development Research Centre (IDRC) made an evaluation study on 10 years reform of Chinese S&T system by the invitation of the Ministry of Science and Technology in 1995, the evaluation report recommends to the policy community the concept of national innovation system; (2) at the same time, the idea of the knowledge economy promoted by OECD were popular in most sectors of Chinese society. Its report discusses the role of national innovation system in the development of knowledge economy. These two events created nice thought environment for the concept of the national innovation system to be accepted in Chinese policy circles

2. The concept of National Innovation System was accepted and developed at the policy level of China In December 1997, the Chinese Academy of Sciences (CAS) submitted to the central government a research report entitled "Embracing the era of knowledge economy and building a national innovation system", which was approved. In 1998, the Knowledge Innovation Project of the Chinese Academy of Sciences was implemented, which initiated the reform of China's science and technology system and the development of science and technology under the guidance of the concept framework of National Innovation System. In 2005, the National Medium and Long-term Science and Technology Development Plan Outline (2006-2020) put forward that "the comprehensive promoting the construction of national innovation system with Chinese characteristics be as an important task of the outline. Later, two important policy document--Opinions on Deepening the Reform of Science and Technology System and Accelerating the Construction of National Innovation System(2012) and the Outline of National Innovation Driven Development Strategy(2016) issued by the Central Committee of the Communist Party of China (CCPC) and the State Council, both put the construction of national innovation system as an important position. By analysis the content and background of the several important policy documents, it can be found that the Chinese learned and introduced the concept national innovation system from the international Academia and policy community, but also made their own creation: come up with the idea of knowledge innovation and its system, put forward that China's national innovation consists of five subsystems: technology innovation, knowledge innovation, national defense science and technology innovation, regional innovation, science and technology intermediary service system

3. Challenges for China's national innovation system construction in the future As Lundvall (2009) points out, China’s national innovation system is a fragmented or dual innovation systems and need to be developed. In the future, several factors must considered for construction national innovation system: national development strategy, new changes in global competition, the trends of new scientific and technological revolution. A new understanding and conceptual framework on innovation and innovation systems is needed. The future development of the innovation system concept needs to incorporate sustainable development goals, environmental governance and inclusive innovation to support broader innovation policies

What 50 Years of R&I Policy History Tells us about How to Make Future Policy: Evidence from a study of policy changes in Sweden
PRESENTER: Erik Arnold

ABSTRACT. Introduction This paper is based on a study of the drivers of R&I policy changes in Sweden over the last half century or so. The study uses a mixture of document and literature review and interviews with key people involved in policy changes. Our choice of changes to study was pragmatic: we chose those that appear to have been consequential, where there was sufficient documentation to allow us to study them and (except in the oldest example) where people involved in the change are still available for interview. We presented some of the findings emerging from the work at Atlanta 2017, and propose now to finish the story at Atlanta 2019. An obvious limitation of the work is that it focuses on a single country – and all countries are in important respects unique. Nonetheless, many of the change processes we saw have echoes of events elsewhere, so we cautiously claim that an intelligent reading of nationally specific events will provide some general lessons.

Why does this matter? The ‘societal turn’ in R&I policy since the start of this century is more than a fashion. It involves the extension of policy from a focus on industrialisation and growth towards addressing ‘societal challenges’ such as climate change, the effects of ageing population, the plastics pollution crisis and so on, some of which can be seen as a direct consequence of the success of previous policies. It is moving us into a ‘third generation’ of R&I policy governance and changes relations between science and society. Everyone, from practising R&I policymakers through mainstream R&I researchers to students of socio-technical transitions understands that we must learn how to do R&I policy in a new way. This paper will not specify what that ‘new way’ looks like. But it will describe the drivers that were needed to effect change in the past and suggest that we will need to enrol the forces for change identified in this study in order to make good policy in the future.

A model of policy change Based on a review of some of the key ideas from political science, we mapped how the mainstream literature explains policy change. We used this to develop a checklist for describing the various policy changes studied. We found some of the factors identified are important, but there are also others that matter which we had not found in the literature.

The cases studied We looked at nine cases, starting in the early 1960s with the creation of the world’s first innovation agency (in the modern sense) and ending with a programme of platforms managing R&D to promote innovation and socio-technical transitions that is still ongoing. • The emergence of the Swedish National Board for Technological Development (STU) in 1968, under the influence of the OECD and a productivity crisis in Swedish industry • The reform of STU ten years later to escape from the lock-ins of earlier practice and become an effective change agent • A process of more or less tribal warfare across three decades between the ‘basic research’ and ‘innovation’ factions, eventually leading to a major reform of R&I funding in 2000 • The creation of the Wage-Earner Fund foundations and a new institutional model for funding research • Another new, large-scale form of R&I policy in the 1980s: the national microelectronics and IT programmes • Reforming the structure and relevance of Swedish university research in the Competence Centres by transferring a design from NSF’s Engineering Research Centres • Addressing regional R&I by twisting the US SBIR programme into a new shape • Trying and failing to change university behaviour via a competition to increase their institutional funding (Strategic Research Areas) • Responding to government demands for policy change by using public-private partnerships and fostering transitional change in the Strategic Innovation Programme

Some conclusions In relation to the model in Figure 1, it is clear that ‘events’ outside the ambit of R&I policy are often important in triggering change, be they in the form of problems arising in the economy or new political forces. Continuous work by policymakers and analysts often supports change, too, so that ideas are available about what to do. Curiously, events seem more often to be important than the occurrence of problems in society. For us as evaluators, a disappointment was how rarely evaluations affected big changes; evaluation is more to do with business as usual than with radical change. Again, for those of us who mostly work with the technocrats who run R&I policy and funding systems it was a surprise to see how important the political level was in triggering change. The bigger the change, the more necessary it was to get the power of the political level behind it. As a result, we would almost always identify one – often several – policy entrepreneurs at the political, policy or even key stakeholder level. Sometimes the policy entrepreneur was in effect an advocacy coalition. And on mature reflection, it seemed that the idea of ‘windows of opportunity’ is almost tautological: change only happens if there is one. The policy cycle, on the other hand, turned out to have almost no relevance to major change. Ideas for policy come and go. Many get put in the garbage can; fewer come out again. Three of our cases involved policy transfer: the rest invented their policies from scratch, though often (and, over time, increasingly) informed by what was being done in other countries. However, there were interesting drivers of change that we had not spotted in the literature. • The most important is the need to set up new organisations or institutions in order to implement the most radical changes. Sometimes it is not enough for old organisations to invent new policy instruments • Lock-ins from previous governance and policy practices – which can impede the work of new organisations or provide the impetus to create new ones • Directors-general of agencies are very important, both as policy entrepreneurs and in the generation of new policy initiatives through turf wars among ministries and (especially) agencies • Other lock-ins arise from power-struggles among different factions, which pursue their own interests (rather than promoting a big idea, as advocacy coalitions do)

Next-generation R&I institutions This evidence leads us to suspect that – just as new institutions and organisations were needed from the 1960s in order to connect science better to economic needs and innovation ¬– we will again need institutional changes to address third-generation, ‘societal’ policies. These demand longer-term engagement and wider societal participation as well as the ability to experiment and learn, changing goals along the way. It is not at all clear that our current R&I funders are up to the job. The political level needs to be strongly involved in order to effect the big changes needed. The power of events will need to be combined with that of the policy stream and policy entrepreneurship to effect the big changes we suspect are needed in order to make third-generation policy effective.

The role of politics in transformative science, technology and innovation policies in emerging economies: the case of Colombia

ABSTRACT. Recent trends in science, technology and innovation (STI) policy have focused on addressing societal and environmental challenges. In this framework, ‘transformative change’ emerges as a STI policy frame aimed at fostering sociotechnical change in order to prompt sustainability transitions (Smith, Voß, & Grin, 2010; Schot & Steinmueller, 2018; Kuhlmann & Rip, 2018). This approach was proposed in Colombia by the national governmental STI policy agency –Colciencias– as depicted in ‘El Libro Verde 2030’. This policy document defines the transformative approach for Colombian STI policy to address the Sustainable Development Goals, and outlines transformative policy principles, objectives, strategies, financial sources and evaluation orientations. The future implementation of this policy approach in Colombia does not seem clear, however, mostly for political reasons: it was launched during the final months of the last government 2014-2018, and the current administration 2018-2022 is not clear about this frame in its governmental program.

Considering this context, the chapter addresses the following research question: what is the role of politics as a key variable to enable or block transformative STI policies to be implemented in developing counties like Colombia? The aim of the chapter is to critically asses the intrinsic features and viability of transformative STI policy implementation in Colombia, from the point of view of policy and governance studies. We argue that current literature on ‘transformative STI policy’ does not account for the political variables that strongly influence agenda-setting and implementation process in the context of highly unequal and contested democracies. This reflection aims at contrasting, on the one hand traditional STI policy approaches strongly biased towards economic competitiveness and its overestimated contribution to wellbeing (Godin & Vinck, 2017) and, on the other, the largely overlooked potential transformative innovation policies have with regard to social inclusion, peace and human development (Cozzens, Gatchair, Kim, Ordóñez, & Supnithadnaporn, 2008; Kuhlmann & Ordóñez-Matamoros, Introduction: governance of innovation in emerging countries: understanding failures and exploring options, 2017; Ordóñez-Matamoros, Centeno, Arond, Jaime, & Arias, 2018). This is a qualitative research and the methodology followed is a case study, which implies reviewing and analyzing policy documents, as well as conducting interviews among different types of stakeholders.


Cozzens, S., Gatchair, S., Kim, K.-S., Ordóñez, G., & Supnithadnaporn, A. (2008). Knowledge and Development. In E. Hackett, O. Amsterdamska, M. Lynch, & J. Wajcman, The Handbook of Science and Technology Studies (pp. 787-812). Cambridge: MIT Press.

Godin, B., & Vinck, D. (2017). Critical Studies of Innovation. Alternative Approaches to the Pro-Innovation Bias. Edward Elgar Publishing.

Kuhlmann, S., & Ordóñez-Matamoros, G. (2017). Introduction: governance of innovation in emerging countries: understanding failures and exploring options. En S. Kuhlman, & G. Ordóñez-Matamoros, Research Handbook on Innovation Governance for Emerging Economies. Towards Better Models (págs. 1-34). Edward Elgar Publishing Ltd.

Kuhlmann, S., & Rip, A. (2018). Next-Generation Innovation Policy and Grand Challenges. Science and Public Policy, 45(4), 448-454.

Ordóñez-Matamoros, G., Centeno, J. P., Arond, E., Jaime, A., & Arias, K. (2018). La paz y los retos de la política de ciencia, tecnología e innovación en Colombia. (C. Soto, Ed.) Seguimiento y análisis de políticas públicas en Colombia. Anuario 2017, 137-168. Retrieved from

Schot, J., & Steinmueller, E. (2018). Three frames for innovation policy: R&D, systems of innovation and transformative change. Research Policy, 47(9), 1554–1567.

Smith, A., Voß, J.-P., & Grin, J. (2010). Innovation studies and sustainability transitions: The allure of the multi-level perspective and its challenges. Research Policy, 39, 435–448.

08:30-10:00 Session 6G: Biomedicine & Policy
Location: GLC 158
R&D incentives and innovation spillovers: evidence from the European orphan drug regulation
PRESENTER: Philippe Gorry

ABSTRACT. This paper, explores the impact of a regional innovation policy for incentivizing drug innovation, on different R&D phases and outcomes, within a global market. According to the OECD(1), 45% of global pharmaceutical sales come from the United States, 30% from Europe, and 9% from Japan while the US retains the highest share of R&D expenditure (58%) compare to EU (28%) as reported by the Association of the British Pharmaceutical Industry (2). In Europe, the UK has the highest share if exchange rate effects are excluded. Specifically, we exploit the European Commission’s Orphan Drug (OD) legislation, which was implemented in 2000, two decades after the US Orphan Drug act (1983). Both legislations are enhancing rare diseases innovation through drug regulatory and financial incentives, mainly through a market exclusivity period independent of the patent monopoly. Our aim is to evaluate the impact of the EU legislation and identify potential spillovers in i) different phases of the R&D process, ii) different drug categories, and also to iii) show that a regional regulation, such as the European, has an impact on the global market. To evaluate and estimate the effects of this policy implementation, we employ a difference-in-differences approach, by utilizing data from the Pharmaproject© drug pipeline database, Questel-Orbit© patent intelligence database and the Wharton Research Database (WRDS) business intelligence research platform. Our sample consists of 61 pharmaceutical firms among the top R&D intensive firms for the period between 1997 and 2003 according. The variables explored for each firms are: R&D budget, Sales revenue, number of preclinical R&D projects, number of Phase 1, phase 2, and phase 3 clinical trials, number of registered and launched drugs, number of patent application by priority date or publication date, number of Orphan Drug status, number of Rare Diseases (RD) or cancer therapeutic indications, number of generic drugs, and merger/acquisition experience. Our results show that the OD regulation leads to a marginal increase of the number of Phase I and Phase II clinical trials for the firms’ global portfolio of drugs. This increase is at the level of 1.2 clinical trials on average for each firm in each of these phases. On the other hand, the regulation does not affect earlier and later stages of the drug discovery and development process, such as patent activity, preclinical studies and Phase III clinical trials. The control variable of firm’s gross sales affect patents and Phase II clinical trials positively, but with low magnitude. In addition, after controlling for generics drugs production, we find a significant positive increase at the preclinical stage and all the clinical trials phases. Moreover, the acquisition experience does not determine innovation at any stage. In addition, we find significant evidence that the orphan drug regulation leads pharmaceutical companies to increase their innovation output on specific drug categories such as drugs for rare diseases and cancer drugs. Particularly, this effect causes an average increase of 2.39 and 4.5 drugs for rare diseases and cancer drugs, respectively. The gross sales control variable and the acquisition experience do not influence cancer and rare disease innovation, while generics drugs production does. Overall, our findings suggest that firms respond positively to the legislation, by shifting their innovation activity towards cancer and rare diseases drugs. At the same time, their total drug portfolio is also positively affected, but marginally and in specific R&D phases, revealing a short-term strategy from their side.

Resolving the healthcare needs in developing countries: Role of inclusive innovations

ABSTRACT. Healthcare systems all over the world rely on drugs, vaccines and medical devices to provide effective and inclusive healthcare (WHO, 2012). However, the development and access to some healthcare technologies such as medical devices and biological have remained an ongoing healthcare and technological challenge in developing countries. Medical devices include everything from highly sophisticated computerised medical equipment down to simple wooden tongue depressors (WHO, 2010). They are critical for diagnosis, effective use of medicines, patient care in operating theatres, at the bedside, and even before a patient is admitted to the hospital, or after being discharged. In developing countries, the majority of research has focused on development and access to pharmaceuticals and vaccines even though medical devices constitute a key component in health care technologies (Kale, 2011). WHO (2010, 2012) highlights that developing countries depend on the imports from the advanced countries to satisfy their healthcare needs, creating challenges of access to affordable and appropriate medical devices. Cheng (2007) revealing the ‘mismatch’ between supply and demand shows that in most cases imported medical devices are mostly unsuitable for local conditions and endanger lives of patients, health workers and communities. Referring to the situation in Africa as ‘medical device graveyards’, Miesen (2013) comments,

“The premature and low birth weight babies lie cordoned-off from the rest in a narrow space where 20 incubators are arranged like Tetris pieces; most were donated by NGOs and bilateral agencies like USAID. Many lay open, and the silence is interrupted only by cries of newborns; no sound emanates from the machines. They are not on. 13 of the 20 incubators are broken. The instructions for one (are) in Dutch. Ugandans typically speak Luganda, Kiswahili and English. Mulago’s experience is not unique. Across Sub-Saharan Africa, “medical device graveyards” litter the empty closets and spare corners of the hospital”

A World Bank review of the Bank’s investment in medical devices from 1997 to 2001 provides clear evidence of this mismatch. The review found cases where, “about 30% of sophisticated equipment remained unused, while those in operation have 25% to 35% equipment downtime because of weak capacity to maintain the equipment” (World Bank, 2003). Similarly, a recent WHO (2010) report shows that more than 50% of devices remain unused in developing countries due to structural and cost factors, indicating a further widening of the mismatch. As a result, access to appropriate and affordable medical devices has remained an ongoing challenge for most developing countries.’

Similarly, biologicals have emerged as a key challenge for developing countries. Biologicals are therapeutic drugs comprising of large complex molecules and traditionally have been developed to address the most challenging Non-Communicable Diseases (NCDs) such as cancer, autoimmune diseases, diabetes, growth hormone deficiency and arthritis. Biologicals drugs have emerged as significant therapy in treatment for cancer and autoimmune disease, but only 8% of patients can access these therapies due to the high costs of these drugs (Rao, 2016). In the last few years, some blockbuster biological drugs have gone off-patent creating potential for developing biosimilars and providing scope for affordable therapies. However, in developing countries, the absence of local production capability and dependence of import continues to make it harder to deliver affordable therapies. The promise of biosimilars as a significant opportunity for reducing healthcare costs and create an affordable treatment for non-communicable diseases (NCDs) is being heralded in an era of growing ageing populations and increasing healthcare costs. However, for many firms from developing countries the complexity of manufacturing biologicals and lack of financial resources has emerged as entry barriers and created a need for new sets of regulatory frameworks, institutional arrangements and partnerships (Kale and Niosi, 2017).

In the last decade, some research has focused on the issues of diffusion and access to medical devices, and biosimilars in developing countries and much has been written about the development of local production capabilities in developing countries, import from MNCs and emerging country firms and role of global institutions (WHO, 2012; Kale, 2011, Nadvi, 1999; Loureiro et al., 2008; Kale and Niosi, 2017). Yet, the healthcare technology needs of low-income populations and inequality of access to imported devices and biological drugs appear largely intractable challenge and need more attention. This raises the question: How healthcare technology needs of the low-income population will be met? Evidence increasingly suggests that innovations based on traditional research and development (R&D) investment and existing institutional arrangements over the years have excluded the healthcare needs of low-income populations (Chataway et al., 2014). There is a gap between skills and knowledge required to understand the healthcare needs of these excluded populations and the intersection of private and public sector boundaries towards meeting them. This calls for a new framework of engagement and some researchers argue that the emerging ideas and models around inclusive innovations and social technologies incorporate the needs, interests and knowledge of low-income populations with resources and capabilities of private sector, thereby providing opportunities to create appropriate solutions for intractable challenges of accessibility, availability, affordability and appropriateness (Kale et al., 2014). While much progress has been made on recognising the new models of innovation and institutional arrangements, discussions failed to provide overall coherent theoretical and policy insights that can aid in resolving the healthcare needs of low-income populations. This paper bridges this gap by using case studies inclusive innovations and social technologies from India, Kenya and Tanzania. It explores the role of inclusive innovation and social technologies in generating and delivering new ‘physical technologies’ and innovation processes needed by low-income users. Specifically, this research critically examines the potential for new ‘social technologies’ (innovative institutional and organisational forms and divisions of labour) and inclusive innovations, to provide a way forward to improve the development and delivery of physical technologies in the medical device and biological drug needs.

Disease burden, research efforts and scientific visibility: Exploring factors diverting research attention from global health needs

ABSTRACT. 1. Introduction: factors influencing health research agenda setting

In the context of health research, priority setting has systematically excluded the needs of those in low- and middle- income nations and has instead focused primarily on diseases that predominantly impact those within high-income countries (Global Forum for Health Research, 1999; Sarewitz & Pielke, 2007; Ràfols & Yegros, 2018). This finding reflects the broader observation of a pervasive misalignment between priority setting in research and societal needs in a diversity of contexts (Ciarli & Ràfols, 2018; Sarewitz & Pielke, 2007). In investigating this misalignment, Ciarli and Ràfols (2018) describe many socio-political factors that likely contribute to the setting of research agendas. These include the observation that problem framings are likely set by the scientific community (and not a diversity of relevant actors), the presence of pervasive legacies of global inequality in the position of power and legitimacy to determine funding agendas, and also widespread path dependencies resulting in the legacy of decision making determining potential future trajectories of research.

The global health research agenda is undoubtedly influenced by a diversity of factors, and within the current analysis, we adopt the conceptual framework described by Ciarli and Ràfols (2018) which highlights that research outputs and agendas are the outcomes of interactions among a diversity of actors with a diversity of value commitments. Research evaluation systems, publication pressures, and norms and incentives in research institutions and communities are likely among the factors influencing the setting of health research agendas. Many current research evaluation systems value scientific visibility, which is often proxied through either the prestige of the journals that the authors publish in, or the number of citations their publications receive (de Rijcke et al., 2015). In some cases, evaluation systems may incentivize researchers in low- and middle-income countries to conduct research that is relevant in high-income countries rather than local national needs (Vessuri, Guédon, & Cetto, 2014). For example, Raitzer and Norton (2009) highlight that in the context of agricultural research, there exist pressures within the scientific community to publish on ‘trendy’ or prestige-laden topics, since these will likely contribute positively to researchers’ career developments. Within this study, we aim to investigate the potential influence between these pressures inherent within the academic community, and how they relate to the distribution of global disease burden.

First, we will build on and expand the results by Ràfols and Yegros (2018) on disease burden vs. number of publications; and second, we explore publication visibility in terms of the citations and the journals that publications are published in and their relationships to disease burden.

2. Previous studies on disease burden vs. publication counts

To investigate the degree to which there exists a misalignment between disease burden and publication output, we build on Ràfols and Yegros (2018). Publications are assigned to diseases by using PubMed and Web of Science data. Subsequently, the distributions of these publications are compared to the disease burden disability adjusted life years (DALYs) estimates provided by the World Health Organization (WHO) for the year 2012. Ràfols and Yegros (2018) examined this misalignment both globally and nationally in the context of Spain’s (as an example of a high income country) disease burden and research output. Below, we will focus on the findings of this investigation at the global level.

This prior research uncovered dramatic misalignments between disease burden and research output at the global level. Consider, for example, that in 2012, malignant neoplasms (cancer) contributed to under 10% of total global health burden (as measured by DALYs) while during the period of 2009-2013, over 22% of global disease publications investigated malignant neoplasms (Ràfols & Yegros, 2018). In contrast, while infectious and parasitic diseases contributed to nearly 20% of global disease burden, they received less than 10% of research attention as measured by the proportion of publications.

Looking at this misalignment at the level of individual diseases often reveals even more dramatic disparities. For example, while skin diseases, breast cancer, and diabetes mellitus each contributed to less than 3% of total global disease burden, these diseases each received a much higher proportion of publications throughout the analysed time period. In contrast, birth trauma, diarrhoeal diseases, Iron-deficient anaemia, and other diseases that predominantly impact low- and middle-income countries receive a considerably lower proportion of publications than their proportional global disease burden.

3. Exploring research visibility (citation impact) across diseases and countries

It is within this context of investigating the potential influences on how health research agendas are set that this study takes place. This paper is novel in that it seeks to investigate the relationship between research visibility (as measured through citation scores and journal impact factors) and disease burden. We hypothesize that publications investigating diseases that are disproportionately prevalent within low-income countries may be less likely to receive citations. Furthermore, we hypothesize that this effect is more pronounced for publications that are produced by low-income countries. As a result, publications that are the most relevant for low-income countries (as measured by disease burden) are likely to receive relatively less attention (as proxied by citations and lower journal impact factor scores) than publications that focus on diseases that are more relevant for high-income countries. Therefore, we hypothesize that there is either a negative or non-existent relationship between citation impact measures and national disease burden for publications originating from low-income countries. This relationship, or lack thereof, indicates the potential inadequacy of using citation-based evaluation methods when evaluating and rewarding health research.

4. Expected results and future work

With this study, we will improve upon previous work that has sought to investigate the relationship between disease burden and research output. This will include analyses at a diversity of levels of aggregation, from the global to the national. Furthermore, we will disentangle effects as they exist at the level of the different disease types that we identified, and at the level of individual diseases. These analyses will benefit from recent work we have conducted to update the correspondence table between medical descriptors in bibliographic databases and the International Classification of Diseases (ICD10) to more accurately link WHO data on diseases burden and publications.

In addition to building on prior work, we will conduct analyses of the relationship between disease burden and research visibility as proxied by both citation score and journal impact factor. The study of this relationship represents a highly novel contribution to the field. This analysis will also be conducted at several levels of aggregation, including fine grained looks at the effects within individual diseases and countries. We aim to investigate the effects of several factors that may influence this relationship, in particular the effects of international collaboration and funding, which are known to effect visibility (Confraria et al., 2017).

Medical research versus disease burden in Africa
PRESENTER: Hugo Confraria

ABSTRACT. Background Africa is a continent facing severe, urgent, and often unique health challenges. The region has made overall progress during the last decades in reducing mortality and prolonging life but its burden of disease per population continues to be two times higher than that of higher income countries. At the same time, most African countries have difficulties in supporting medical research, and the pharmaceutical industry may be reluctant to sponsor research in lower income countries because the prospects of profit are limited, even if effective treatments are developed. Nevertheless, it has been well recognized that health research conducted in low-income countries is of great importance. Medical research done and applied in lower income contexts can provide enormous contributions to discovering previously unknown diseases, which have substantial social and economic impact on the world. Furthermore, medical research based in low-income countries can help researchers have a clear understanding of the local constraints and barriers to and facilitators of the implementation of research in practice. Lack of funding for research is the major barrier to the development of clinical research capacity in Africa and the majority of clinical research is based on funding from external donors. Since local governments are taking a peripheral role, there is a risk that research conducted in African countries will follow topics determined by foreign funders for their purposes and consequently fail to respond to specific local health needs. In our research, we will look specifically at the scientific output of African researchers and also at the funding institutions acknowledged in their publications. This will allow us to evaluate whether international development funders and pharmaceutical companies are supporting research that matches the research needs of African regions. This approach and research question may be interesting for two main reasons. First, due to the tremendous health challenges the continent faces, improved Africa-relevant health research can have an important role in changing the professional practice of health care providers and a significant impact on health outcomes. Second, the improvement in research capabilities in the continent in the health sciences may demonstrate that persistent support and funding from development partners such as the Wellcome Trust, NIH, European Union or Bill and Melinda Gates Foundation, might pay off. This study uses DALYs (Disability-Adjusted Life Years) in each disease field and African region as a proxy for societal needs in health, which is compared with scientific research in each corresponding disease field and region. We focus on four African regions (Eastern Africa, Northern Africa, Southern Africa or West & Central Africa) as defined by the UN classification, and our central research questions are: 1) Is the amount of research produced on various diseases by African researchers related to their countries’ burden of disease? 2) What kind of health research is being funded by international organizations? 3) What are the main drivers of medical research in Africa?

Data and Methods Data Sources. To identify medical priorities, we used DALYs from World Health Organization to measure the burden of disease. One DALY can be thought of as one lost year of healthy life, and the measured disease burden is the gap between a population’s health status and that of a normative reference population. To present the funded (and unfunded) research efforts, we rely on scientific research publications indexed in WoS. From the ‘funding organization’ field in each article, we identify whether the research was financially supported, and the name of the funder(s). Publication records were assigned to a specific disease field by searches in abstracts and titles. We built a set of keywords that are strongly associated with a specific disease (or group of diseases) based on the ICD-9 codes and previous research. After building our queries, two external peer reviewers reviewed the keywords for each one of our 28 disease categories. In total, we have 59,486 documents that were associated with at least one specific disease (28) and one African region (Eastern Africa, Northern Africa, Southern Africa or West & Central Africa). Metrics. With the hypothesis that health needs in earlier years should drive the research agenda in later years, we compare the number of articles published between 2011-2015 with the disease burden in 2010. First, we count DALYs in each disease per region/period and number of publications in each disease per region/period. Then, since different diseases have different propensities to affect people and be researched, we also compute specialization indexes to assess the specialization of each disease in a given region. Econometric approach. In this study, our primary research question is to understand whether disease burden specialization is associated with medical research specialization between different African regions across different diseases. To address this, in our multivariate regression analysis (OLS), we use scientific specialization as our dependent variable, and disease burden specialization as our main independent variable. Since most African countries are highly dependent on international research collaboration, in our model we control for level of international collaboration. We also control for previous scientific specialization due to the path-dependent nature of scientific production.

Preliminary findings Surprisingly, what we find is that in sub-Saharan Africa most diseases with a high disease burden are also the ones with relatively more research efforts. We find that the region with the highest positive association between disease burden and research efforts is Eastern Africa. Northern Africa is the region where these two dimensions are less aligned. Contrary to what some literature suggests, our results indicate that the regions with higher dependence on public non-African and philanthropic funding institutions are the ones where the alignment between disease burden and research specialization is higher. These findings are interesting for two main reasons. One, it has been argued that there are substantial misalignments, at the global level and at local levels, between research efforts and WHO estimates of health burden for a given disease. What we find is that while this may be true at the global level (high-income countries perform most of their medical research on diseases that are not the ones with a higher global disease burden), Sub-Saharan African researchers are performing research that is relevant for their regional health needs. Second, some authors also argue that researchers from high-income countries secure most of the funding for global health research projects in low-income regions, and often dictate the research agenda which leads to inappropriate projects unrelated to local research needs, and derive conclusions that do not have any direct local benefit. The results of this article seem to contradict this idea. What we find is that most international research funders (public non-African and philanthropic) support research on HIV/AIDS, tuberculosis and parasitic and vector diseases, which are diseases that generate a big share of disease burden in sub-Saharan Africa. What our regression results seem to show is that, on average, high levels of dependence on international donors are not necessarily associated with less alignment between local health needs and local medical research efforts. Our study has limitations and the results must be interpreted with caution since first publications in WoS (or DALYs) are imperfect estimates of research efforts (health needs) in a specific disease and country. Second, we do not know to what extent the research that is funded is actually used to contribute to health action.

10:30-12:00 Session 7A: Entrepreneurs & Firms
Location: GLC 233
Emergent small business narratives in three high-technology industries

ABSTRACT. Background

Small firms operating in high-technology are diverse in terms of size, age, capabilities, management composition, science and technology focus, and industry. Yet, for market facing activities, they often must develop messaging around their products, services, and culture. Messages can be thought of as pre-packaged storylines that convey key facts and symbolic representations about who the firm is, what it offers, and how it produces and interacts with customers: Stories package “factual information about [a firm’s] stock of tangible and intangible capital into a simpler, more coherent and meaningful whole” (Martens et al., 2007).

Various streams of research examine the broad question of narrative development and dissemination and outcomes. For example, prior work in entrepreneurship focuses not only on market facing storylines but also on building symbolic actions to further resource acquisition and realize enhanced performance outcomes. Storylines may also be important for established small firms due to the inherent uncertainty of new technologies and scientific discoveries.

Storylines and plots matter not just for individual firms but for industries. New technologies may initially lack standards and/or clears applications for use, therefore limiting opportunities to attract attention to the industry. In response, technology developers craft storylines and plots to create niches where they can co-interpret opportunities and marshal resources in networked settings. When packaged into plots of expected patterns and conclusions, a sectoral “dominant logic” of plots may appear in different industries over time.


The analytical approach isolates narrative development on small firm websites by examining topical change that naturally occurs on select ‘about us’ pages. ‘About us’ pages provide relevant context about a firm’s main storyline, e.g., information about the firm’s history, industry, technology orientation, R&D capabilities, culture and mission, partners, and management structure. But not all firms will cover the same topics and in the same order. Some topics may appear more often than others, and this gives rise to heterogeneity in narratives, both across firms and industries. We draw our sample frame of inventive firm websites by identifying patenting firms in the three industries using USPTO’s PatentsView data platform.

We operationalize our dependent variable as a series of changes in topics over time, as measured in distinct paragraphs. We use a common topic modeling algorithm, latent Dirichlet allocation, which treats observed words as belonging to latent topics in observed documents, in this case website ‘about us’ pages. The progression of top-most likely topics in each paragraph can then be thought of as a state change from one topic to another. We model this phenomenon using multi-state continuous Markov modeling, which estimates transition intensities between paragraph-topic state r to paragraph-topic state s based on time t and individual or time-varying covariates z(t):

q_rs(t, z(t)) = lim_delta(t)->0 P(S(t + delta(t)) = s|S(t) = r)/ t

Individual firm covariates include firm size (number of employees), firm age, total number of patents granted by USPTO, and a measure of disruptiveness of a firm’s overall patent portfolio.

Preliminary and anticipated results

We present two sets of findings. First, the topical composition of ‘about us’ pages by industry differs in expected ways: In synthetic biology, there are two optimally identified topics, one that focuses broadly on innovation for improved health and another on a highly educated scientific workforce. In nanotechnology, the topics describe customized services, clinical treatment, manufacturing materials for electronics, printing, and human capital. Renewable energy topics focus on customized solutions, solar power cells, historical milestones and logistics, investor relations, and human capital.

Second, anticipated findings from the Markov models are expected to reveal how narratives – i.e., topical change – differs depending on individual firm covariates. Firms that are smaller and younger may choose narrative constructions that point toward human capital topics, in part because these firms often have fewer well developed capabilities and/or a shorter history to draw upon. In the green goods and nanotechnology industries, higher levels of patent grants are expected to positively predict state changes to topics associated with product areas and/or manufacturing processes, thus reflecting the capability to scale-up production. This may not be true in synthetic biology, where (a) the scientific invention closely resembles the end product, and (b) manufacturing may not play as critical of a role.

Higher disruption indices of a firm’s patent portfolio may be more likely to predict the key defining topics of an emerging industry, but not a well-established one. So, for newer industries such as synthetic biology, disruption may coincide with the dominant logic of ‘innovation for improved health’ and ‘human capital’. Disruptive firms in older industries such as renewable energy, in contrast, may show lower propensities to transition to dominant topics in their narrative construction.


There are two significant contributions of this work, the first of which draws on descriptions of the topics that appear on small firm websites, by industry. We explore which firm factors (e.g., size, patenting activity) are most likely to explain changes in the narrative. Because narrative generation is a subjective determination made by the firm, we are able to determine when and why certain firms may choose a narrative that differs from its peers. In addition to its research merit, the findings should generate interesting implications for science-based small firms, who often struggle to put forth compelling storylines that can, in turn, attract key resources.

Our second contribution is methodological. Computational social science methods have not been used widely for studying firm narratives as they result to science-based product innovation. Using freely accessible patent and website data, we are able to scale our collection and automate our method for large-scale narrative analysis.

Understanding the role of Local Institutional Entrepreneurs in inclusive innovation initiatives in rural communities in Colombia

ABSTRACT. Background and rationale Offering new opportunities to those living in poverty and inequality conditions is a common challenge in the world (Papaioannou, 2014). Like in Colombia, solutions to those issues have been focused on the implementation of policies to foster the industrial sector and cash transfer programs. These policies have shown poor performance; for instance, in the last 57 years, the level of productivity in Latin America and Sub-Sahara Africa has been negative on average (Cavallo & Powell, 2018), and indexes of inequality (i.e. Gini index) remain high in those regions. In this vein, and despite these policies, this challenge remains a global concern.

In this frame, a new set of strategies named ‘Inclusive Innovation’ is emerging with the purpose of working directly with marginalised communities to solve their needs and concerns by using innovation (Harsh et al. 2017). It is the case of some policy programmes in Colombia, such as ‘Ideas para el Cambio’ and ‘A Ciencia Cierta’ which have been implemented by the Administrative Department of Science, Technology and Innovation (COLCIENCIAS) since 2012. These programmes are bringing the opportunity to marginalised people of enjoying the benefits of the development of science, technology and innovation (STI) projects by considering their needs and their voices. In these initiatives, Local Institutional Entrepreneurs (LIEs) are actors that break the status quo and bring about new rules of the game (Pacheco et al. 2010), contributing to overcome the challenges that inequality and poverty represent.

Although those actors have been studied in environments such as enterprises (Munir & Phillips, 2005), processes of regional integration (Fligstein, 1997) or the health care systems, little is known about their roles in emerging economies’ territories (Battilana, Leca, & Boxenbaum, 2009) or about their specific function in conciliating interests between actors from the government, academia and local communities. Also, limited knowledge is available about the way they open up path-transformative opportunities (Sotarauta & Pulkkinen, 2011; Westley et al. 2011) for those who are living in areas in conditions of marginality. In this vein, the proposed paper strives to set down a starting point to understand to what extent are LIEs in regions crucial to foster the success of inclusive innovation initiatives in local communities in Colombia. To doing so, and as part of a broader research project, the paper will discuss a hypothesized set of answers around this question.

Theoretical Approach To address the above mentioned question, this research uses three theoretical lenses. Firstly, it considers institutional entrepreneurship literature, mainly based on the work by DiMaggio (1988), Leca & Naccache (2006) and Battilana et al. (2009). Understanding the role of Institutional Entrepreneurs in the process of institutional change opens up the opportunity to reflect on path dependence, which is the second set of lenses in this study. In this vein, authors such as Levi (1997), Sydow et al. (2005), Martin & Sunley (2006), Martin (2010) and Dawley et al. (2010) are considered to study the criticisms they have made about this theory, as well as to explore the way in which this theory could be linked to the LIE. Finally, the notion of Inclusive Innovation steers the study of the previous theories in those realms where communities are looking in innovation the way to tackle poverty and in inequality. In this purpose, authors like Cozzents et al. (2007), Cozzens & Sutz (2014), Heeks, et al. (2014), Papaioannou (2014), Swaans et al. (2014), Pansera & Owen (2018), Harsh et al. (2017), Kuhlmann & Ordóñez-Matamoros (2017) are consulted.

Methods Based on these lenses and secondary information, this research follows an abductive approach (Awuzie & McDermott, 2017) supported in Case Study Research (Yin, 2018) and Process Tracing methods (Beach & Brun Pedersen, 2013), to the aim of hypothesising, reflecting and discussing the causal mechanisms that could give a preliminary answer to the research question. To do so, three in-depth case studies will be carried on. The cases will be selected based on five criteria. First, cases where the State has tried to foster local initiatives allowing communities to identify their problems and solutions to improve them. Second, cases where the intervention process has finished. Third, projects linked with productive-development solutions carried out by the communities. Fourth, each case belongs to one of three possible socioeconomic contexts (high, medium and low socioeconomic performance) defined in this research to contrast the implications of contextual variables in the cases. Finally, safety conditions and geographical proximity for doing fieldwork in Colombia are considered.

Preliminary results As a provisional result, this study has identified a new category of change agents in Institutional Entrepreneurship theory, called as ‘Local Institutional Entrepreneurs’. The analysis of them, their action repertories and skills, and the way they build bridges between diverse actors in inclusive innovation initiatives provides a fine-grained understanding of the role of the agency as an endogenous source of development in the Colombian regions.

Along with the previous result, the analytical and speculative reflexion that provides this paper will set down the starting points in an abductive approach that is followed to answer the research question mentioned. In this vein, it will be identified likely and plausible causal mechanisms that explains how the LIEs are conciliating (or not) interest from different actors such as National authorities, Science, Technology and Innovation’s experts, and local communities, and how they are opening up (or not) opportunities to foster path-transformative development in local communities based on inclusive innovation initiatives supported by COLCIENCIAS.

Significance Understanding LIEs in ‘Inclusive innovation’ projects supported by the State is a centrepiece in the puzzle of the transformative policies required in marginalised regions. The LIEs’ skills, strategies and the pathways built by them to develop a shared vision and introduce divergent changes ((Battilana et al. 2009) in the communities’ practices and institutions will contribute to explain how to tackle poverty and inequality as a joint effort of multiple actors in emerging economies. Along with this implication, the study of these actors is crucial to identify concrete insights about how cooperation emerges and is created between actors thanks to the conciliation efforts done by LIEs, and also how this cooperation led by LIEs makes possible to improve the coordination between programs launched by national entities such as COLCIENCIAS and local communities.

Cooperative R&D and the performance of start-up firms: Does founders’ human capital matter?

ABSTRACT. Principal topic In this study, we examine whether founders’ human capital moderates the relationship between cooperative research and development (R&D) and the performance of start-up firms. The main aim of this study is to clarify whether founders’ prior knowledge plays an important role in determining firms’ absorptive capacity and thus enjoying the benefits of cooperative R&D. This study contributes to the literature on the effectiveness of cooperative R&D strategy of start-up firms and provides a clue for policy makers by investigating the unexplored question.

Backgrounds It is not easy for start-up firms to be successful in innovation, because of a lack of resource. Therefore, to share risks and costs and achieve innovation outcomes, cooperative R&D may be an effective strategy for R&D-oriented start-up firms. Meanwhile, it is well recognized that absorptive capacity—the ability of a firm to recognize the value of new external information, assimilate it, and apply it to commercial ends—is critical to innovative capabilities (Cohen and Levinthal, 1989, 1990; Zahra and George, 2002). Although absorptive capacity is a function of the level of prior related knowledge, including basic skills and a shared language and scientific or technological developments in a given field (Cohen and Levinthal, 1990), start-up firms without a business history typically lack prior related knowledge (Shane, 2000). This suggests that start-up firms face difficulties in acquiring external knowledge.

It is widely recognized that new technology-based firms (NTBFs) established by individuals with greater human capital should outperform other NTBFs because of their unique capabilities. In practice, a number of studies showed evidence that founders’ human capital is a valuable resource of the start-up firms and plays a critical role in its performance. In addition, founders’ human capital, composed of their previous knowledge and skills accumulated prior to establishing the start-up, plays a critical role in determining the absorptive capacity of start-up firms (Debrulle et al., 2014; Kato, 2017), because it may compensate for a lack of prior related knowledge at the firm level in start-up firms dependent heavily on the knowledge sources provided by founders (Colombo and Grilli, 2005; Van der Sluis et al., 2008). However, there has been limited evidence as to whether and how founders’ human capital plays a key role in achieving superior performance via cooperative R&D because of their superior absorptive capacity.

Data and method Using a data set of R&D-oriented start-up firms based on original questionnaire surveys conducted in Japan from 2008 to 2010, this study explores whether and when cooperative R&D is an effective strategy for the post-entry performance of start-up firms. In particular, we shed light on whether the effect of cooperative R&D on firm growth (in terms of employment and sales) depends on founders’ prior knowledge.

In this paper, we estimate a treatment-effect model, taking into account that there may be systematic differences between firms that engage in cooperative R&D and others. The probability of engaging in cooperative R&D is estimated to obtain propensity scores in the first stage of model. Then, the effect of cooperative R&D on lagged growth is estimated based on the matched sample.

Results and implications The logistic regressions for the model determining the probability of engaging in cooperative R&D show that firms managed by founders with high levels of human capital, measured as educational backgrounds and industry-specific work experience, are more likely to engage in cooperative R&D.

Then, the results in the second stage indicate that cooperative R&D is on average not effective in achieving high growth for start-up firms. In addition, the subsample results for firms with and without cooperative R&D show that firms that engaged in cooperative R&D achieve superior growth performance when their founders have experience in innovation prior to start-up. On the contrary, the findings indicate that firms whose founders have no experience in innovation do not achieve superior performance via cooperative R&D.

It suggests that the founders’ prior knowledge plays a crucial role in determining firms’ absorptive capacity and thus enjoying the benefits of cooperative R&D. From an economic policy perspective, the findings of this study implies that organizational absorptive capacity should be enhanced before promoting research partnership between organizations as a means of public support for innovative start-ups, especially when founders do not possess technical knowledge at start-up.

Ecosystems of Entrepreneurship: Configurations and Critical Dimensions

ABSTRACT. This paper characterizes heterogeneous configurations of entrepreneurial ecosystems and important dimensions thereof. These dimensions refer to characteristics that are common across different ecosystem setups and others that vary from locality to locality. Such performance creates a kind of hierarchy whereby factors deemed most important across many ecosystems are ranked higher than factors with more specific relevance to one or few ecosystems. We adapt classic models of entrepreneurial ecosystems and empirically investigate the extent to which different localities within a large region (State of São Paulo) of an emerging economy (Brazil) follow a uniform ecosystem configuration and why, or why not. Our primary target is to suggest an empirically validated model of entrepreneurial ecosystem behavior that can be consulted by policy decision makers in guiding future infrastructural investments. Knowledge-intensive entrepreneurship (KIE) stands for a group of new, innovation-driven firms that not only achieve high levels of competitiveness, but that also positively affect aggregate capabilities in the systems in which they operate (Malerba & McKelvey, 2018). Nonetheless, these firms still lie at the margin of innovation systems’ approaches (Ács et al., 2014; Autio et al., 2014; Zahra et al., 2014). An important recent development has been the growing attention to ecosystems of entrepreneurship (EE), offering a more effective framework to address the interplay between individual and contextual dimensions (Audretsch & Belitski, 2017; Borissenko & Boschma, 2016; Stam & Spigel, 2016; Fischer & Nijkamp, 2018) This is encouraging given that KIE is highly connected to the systems in which it is embedded, presenting higher levels of systemic interactions than other cohorts of new firms (Malerba & McKelvey, 2018). These connections are fundamentally local, relying on the neighboring availability of knowledge, institutions, resources and demand (Stam, 2009; Isaksen & Trippl, 2017; Fischer & Nijkamp, 2018). As a consequence, we observe a spiky geography of such entrepreneurial ventures and significant regional asymmetries in the context of both developed and developing countries (Ács & Armington, 2004; Florida, 2005; Pan & Yang, 2018; Fischer et al., 2018). In our view the extant literature offers inadequate guidance by focusing too much on anecdotal evidence from outstanding outliers, such as Silicon Valley and Route 128 (Nicotra et al., 2018; Stam & Spigel, 2016). The resulting policy implications tend to neglect that KIE is deeply embedded in local contexts and that these contexts are highly heterogeneous across regions and countries (Radosevic & Yoruk, 2013; Ács et al., 2017; Brown & Mason, 2017; Boschma & Martin, 2010). This is true even among high-profile successful cases, which do not necessarily share similar evolutionary patterns (Saxenian, 1994). Hence, the challenge of transferring successful ‘recipes’ from one place to another remains in the case of entrepreneurial ecosystems considering differences in areas such as culture, traditions, capabilities and networks (Malecki, 1997a). Understanding the distinct characteristics, trajectories and interplay of relevant drivers of entrepreneurial ecosystems becomes a strategic matter in this debate (Etzkowitz, 2019; Kantis, 2018; Feldman, 2001). As cities face demands to become more involved in promoting knowledge-intensive businesses (Tiffin & Jimenez, 2006), the need to identify indicators that influence the behavior of ecosystems assume a critical position (Auerswald & Dani, 2017; Spigel, 2017). In this article we build on the notion that ‘generic’ and comprehensive configurations of EE present only a limited understanding of their intrinsic dynamics. We argue instead that a hierarchy of ‘ingredients’ exist and that these elements can be combined in different ways as to generate intense entrepreneurial activity. Accordingly, the goals of this research are to: (i) identify the fundamental dimensions that cut across distinct ecosystems of entrepreneurship; and (ii) assess the different forms these overarching conditions combine with other drivers to promote KIE activity. We gather insights from the extant literature on business location and the dynamics of entrepreneurial ecosystems to develop a workable model for our research inquiries. Empirical data come from the State of São Paulo, Brazil. KIE is approximated using information from PIPE projects, a program funded by the State Research Foundation (FAPESP) that supports innovation-driven entrepreneurship in small companies (similar to the SBIR Program in the United States). The full dataset includes 1856 grants allocated across 135 cities in the State during the period 1998-2017. The methodological approach is developed around classic analyses of entrepreneurial ecosystems such as Isenberg (2010) and the empirical analysis uses Qualitative Comparative Analysis (QCA) with fuzzy sets. This offers a novel way to address the configurational dynamics of EE, from the bottom up, complementing insights from qualitative case studies and traditional econometric approaches . Findings strongly suggest that a common core of features for thriving EE, namely: local presence of strong research universities, availability of entrepreneurial habitats (incubators and science parks), and the knowledge intensity of labor in the city. Infrastructure, human capital, income levels and credit availability make up a secondary layer of important attributes. At a higher level of detail, four variants of ecosystem configurations could be identified – albeit at a relatively marginal rate of differentiation. Our analysis also contributes to the literature by offering perspectives on the geography of innovation and entrepreneurship in the context of an advanced region in an emerging economy. It thus helps filling a gap in terms of scant evidence on EE outside the United States and Europe (Hayter et al., 2018). It also allows bringing forward issues associated with the differences between laggard and developed economies when it comes to the location dynamics of entrepreneurship (Crescenzi & Rodríguez-Pose, 2017).

10:30-12:00 Session 7B: Gathering Input for Policy
Location: GLC 235
Succeeding and failing in delivering scientific evidence: a close look of two cases of environmental conflicts in Chile.

ABSTRACT. Chile has one of the most diverse environments in the world: from the driest desert of Atacama, to the eternal snow of Antarctica. However, our rich biodiversity is often threatened by all sorts of human activities. In the north, copper and more recently lithium mining activities have for decades, polluted its soil, air and water, creating toxic environments for the inhabitants of cities like Antofagasta. In the south, urbanization and growing population has pushed the limits of the cities like Valdivia farther into native woods and wetlands that regulate water flows and hold unique ecosystems. These developments have resulted in socio-environmental conflicts, affecting the wellbeing of people and the environment While similar in some aspects, the two cases of Valdivia and Antofagasta had unfolded differently. In Valdivia, decision makers developed a close relationship with local scientists and citizens’ organizations, and together they pushed a national legislation to protect urban wetlands, which is currently being discussed in the national Congress. This is a mayor accomplishment for any local issue outside the capitol. In contrast, in Antofagasta, scientist and local activist have for years warned about the noxious effects of mining activities, yet there has not been any response from the authorities. Recent scientific findings related to soil contamination by a group of local scientists are being ignored, despite of the efforts of scientists to convey their concerns over the health issues that can affect the local population. This proyect aims to compare these two cases in order to identify, characterize and understand the mechanisms that facilitated the inclusion scientific evidence into policies of environmental management. The comparison will allow for a detailed analysis in three key issues: (1) highlight best practices, (2) identify barriers to interdisciplinary work between scientists and policy makers; and (3) better comprehend how local contexts can influence the interplay between science and decision-making. This latter point is very important when best practices are sought to be applied in different cities or countries. This comparative case study will be conducted through document analysis (press reports, video of Senate discussions, academic papers) and a series of interviews with relevant actors such as scientists, city’s office staff, members of the congress, journalists and others, in order to characterize the different strategies developed by groups of actors for each of the cases, the types and mechanism by which scientific evidence and other forms of knowledge and information are communicated and shared, and the mechanism for conflict resolution, coordination and cooperation that lead to policy engagement and change. This study can have an important applicability in the global south, where scientists are scarcer, and the interaction between them and policy makers is still a work in progress, with few institutions facilitating it.

The Connection Between Battery Storage and Climate Change: Testing for the Politicization of Energy Storage Research in the Media

ABSTRACT. Background

There is a disconnect between scientific research itself and how it is reported in the media, which can have serious effects on the policy agenda setting process. As a result, this project is motivated by the need to understand how the public’s access to science may be constrained. Two assumptions are central. First, public inaccessibility to science is a function of its inaccurate portrayal in the media. Second, science can be easily politicized, principally when research findings are questioned and challenged in the public forum. The focus here is electrical energy storage, which has the potential to solve a host of problems facing the electronics utility industry such as blackout preparedness, grid-wide load shifting, and the intermittency of wind and solar power generation. Yet, energy storage technology is tied to climate change in terms of advancing clean energy production systems, the widespread use of electric vehicles, and subsequent greenhouse gas reductions. Climate change is the now-classic example of how scientific consensus can be undermined through media-based effects.

We employ the tools of science communication to analyze how the topic of energy storage is being discussed in the news media, which is instrumental for setting the science and technology policy agenda. Simultaneously, politicization of science and technology limits the sharing of otherwise credible scientific information and decreases trust in scientific evidence. To our knowledge, no research has examined news content related specifically to energy storage, and this paper thus focuses on energy storage-related content in the New York Times and The Guardian. Given the politicization of climate change, and given the connections between climate change and energy storage technology, it is hypothesized that energy storage scientists and energy storage technology are discussed in articles if political framing content is also present.


A comprehensive content analysis is conducted of all battery storage-related articles published through mid-2018 in the New York Times and The Guardian; a total of 527 articles. An iterative process of handcoding these articles identifies content regarding governmental agencies, political groups, individual scientists, references to academic sources, references to organizations (e.g., universities, national labs, companies), mentions of consumer products (e.g., electric cars, electronics, home energy storage appliances), and the specific type of energy storage technology mentioned. The following are the identified dependent variables: the presence of political content (political parties, political ideologies, governmental branches/committees/agencies, the European Union, foreign governments, multinational panels/agencies/organizations/treaties), the presence of scientist-related content (when any scientist, researcher, or analyst was mentioned and identified as a scientist, researcher, or analyst, including those from the natural sciences, the social sciences, or a related field, and specialists in the private sector), and the presence of technology-related content (mention of lithium ion technology, hydrogen fuel cell technology, nickel-metal-hydride technology, pumped hydroelectric energy storage technology, compressed air energy storage technology, capacitor or supercapacitor technology, flywheel technology, concentrated solar technology, flow battery technology, vehicle to grid technology, or other types of technology).

Tags are grouped according to co-occurrences by the Clauset-Newman-Moore clustering algorithm. For articles from the New York Times, the following groups have been identified: car and battery; policy and politics; infrastructure and energy actors; innovation and energy. For articles from The Guardian, the identified groups include: overview (i.e. causes and implications of battery storage technology); policies; politics and autos; Australian energy policy; urban and corporate social responsibility. Each of the dependent variables is regressed on these groups of tags to test the aforementioned hypothesis.


Descriptive statistics confirm that politics dominates media content on energy storage. For the New York Times, the dominant topics discussed are centered on Car & Battery and Policy & Politics, 42.1 percent and 54.6 percent, respectively. In The Guardian, the dominant topic is the Overview set of tags, followed by the Policies group (42.1 percent), the Politics & Autos group (37.7 percent), the group focusing on Australian Energy Policy (16.8 percent), and the group that covers Urban & CSR (5.1 percent).

Logistic regression of the dependent variables on these groups of tags shows that, for the New York Times articles, the presence of Car & Battery topics decreases the probability of political content being present by approximately 47 percent. The presence of Policy & Politics topics as well as Energy Actors & Innovation topics, however, increases the probability of political content being present by 3.8 and 3.9 times. In terms of predicting the presence of scientists in New York Times articles on energy storage, only the Policy & Politics group is statistically significant, the presence of which makes it 2.4 times more likely that scientists will be mentioned. Finally, the presence of technology is significantly increased 1.6 times and 2 times, respectively, with the presence of the Car & Battery and the Infrastructure & Energy Actors topics. All other groups do not significantly predict the presence of technology-related content. The presence of content discussing scientists, thus, is strongly predicted by political content, while technology content is predicted by Car & Battery-related content as well as Infrastructure & Energy Actors-related content. We can conclude that energy storage content in the New York Times is discussed in political terms.

With regard to The Guardian, only the Overview group and the Australian Energy Policy group significantly predict an increased likelihood that political content will be present, 2.1 and 5.2 times, respectively. The group of tags/topics conveyed in Policies strongly and significantly predicts the presence of scientists being mentioned in an article. None of the various topic groups predict the presence of technology being mentioned. There are, thus, significant differences between news outlets in terms of how energy storage technology is presented.


This paper curates and codes a dataset of energy storage-related content in the mainstream media. This content can be connected to climate change research, which has been politicized over the last two decades. While there is an abundance of political content in both The Guardian and New York Times, our statistical analysis confirms that energy storage content is not politicized; however, it is frequently connected to the political realm with regard to its related communications in mainstream news outlets. Given the centrality of topics such as greenhouse gases and climate change in newspaper content focusing on energy storage, the foundation is may be set in the United States (i.e. based on the New York Times-related analysis) for energy storage to be just as politicized as other energy-related topics that have been ideologically connected to climate change.

A Schizophrenic Public? Understanding the Shifts of Public Opinion on Nuclear Energy Policy through the Deliberative Polling Process

ABSTRACT. “Too cheap to meter.” Once touted as the source of unlimited energy, nuclear power is now shunned as risky and dangerous. In the aftermath of the Fukushima Daiichi nuclear accident, the world has seen several countries announcing a complete phase-out of nuclear power (notably Germany and Switzerland) and others slowing down or giving up the construction of new nuclear power plants (NPPs). A reversal of nuclear energy policy has not been more dramatic than in South Korea, however. It was one of the thirty-one countries holding on the plan to construct more NPPs despite the Fukushima accident, which came as no surprise given the long history of nuclear energy as the symbol of the nation’s independent technological capacity as well as its potentialities as dual-use technology in the geopolitical dynamics of the Korean peninsula. With the new government formed after the first presidential impeachment amid the widely televised candlelight demonstrations in the winter of 2016, South Koreans witnessed the shutdown of its first NPP (Gori-1) in June 2017, which was heralded by anti-nuclear environmentalists as the beginning of South Korea’s journey towards a nuclear-free country. A large segment of the expert community comprised of nuclear scientists, engineers and industrialists (nicknamed as the nuclear mafia by anti-nuclear groups) voiced their concerns with the new government’s decision to phase out nuclear energy, criticizing it as a populist policy catering to uninformed citizens. In particular, the new administration’s plan to stop two NPPs under construction announced during the Gori-1 shutdown ceremony provoked the most intense responses by citizens, environmental groups, and nuclear expert communities, leading to a decision to derive public consensus out of deliberative polling on the issue. The three-month deliberative polling process ensued between August and October 2017, which ended up in an apparently contradictory result. A majority of the citizens participating in the process agreed to resume the construction of two new NPPs, which was the clear victory for the nuclear expert community. At the same time, a majority of the citizens endorsed the new government’s nuclear phase-out policy, which was the clear victory for the anti-nuclear community aiming to end the nation’s fifty-year commitment to nuclear power and technology. Which side then won? Were the public schizophrenic? Is this, after all, the manifestation of the populism of citizens making uninformed decisions or the manifestation of the elitism of nuclear experts monopolizing the privilege of technical knowledge? This study is an attempt to understand the context and process leading to the apparently contradictory conclusion reached out by general citizens deliberating on a highly technical and politically charged issue by drawing on a few theoretical concepts developed at the interface of politics and technology including the deficit model, sociotechnical imaginary, and deliberative democracy. The first section of the study presents a historiographical analysis of South Korea’s nuclear policy in an effort to situate it in the historical context of South Korea’s S&T capacity building. Relying on the sociotechnical imaginary concept (Jasanoff & Kim 2015), it reveals the unique characteristics of nuclear policy discourse, which is distinctively marked by the discourses of nation-building and economic take-off (notably with South Korea’s first government-funded research institute being the Korea Atomic Energy Research Institute established in 1958). The following section looks into the 2017 process of the deliberative polling commissioned to the Institute for Social Development and Policy Research at Seoul National University. Running for 89 days with deliberative discussions of 471 citizens selected representatively, the process culminated in four opinion polls showing notable changes in initial opinions, which were the very cause of cheers and disappointments of pro- and ant-nuclear camps. The third section attempts to assess the design and outcome of this deliberative polling process in order to derive generalizable lessons and insights for the policymaking on a technically complex issue in light of lay expertise and deliberative democracy (Fishkin 1991 & 2009). In particular, a multivariate analysis of the four opinion polls utilizing the deficit model is run to test the interactive effects of scientific literacy and political interest on the opinion changes through deliberation as a way to assess the degree of susceptibility to populist vs. elitist arguments on the future directions of nuclear energy policy. The final section concludes with a short summary and implications of the study result for the politics of sociotechnical policymaking.

10:30-12:00 Session 7C: Team Science
Location: GLC 236
Enhancing the Effectiveness of Team Science: A Study of Evaluating Current State of China's Scientific Research Teams

ABSTRACT. background and rationale The benefits of cooperation and team science has been increasingly recognized by contemporary science community during the past decades. Due to the increasing complexity and cost of modern research which far exceed the capability of individual elite scientist, collaborative working relationships among researchers became an integral part of scientific discovery. Scientists working in teams interact and integrate across disciplinary, professional, and institutional boundaries. In most fields, research teams have grown significantly and scientific collaboration has become more strikingly prevalent.

Understanding formation and evolution of research teams and evaluating their performance have become more and more important for scientific community, policy makers, and other stakeholders. The benefits of cooperation and team science can be summariezed as follows: Increasing research impact, novelty, productivity, and reach offering greater options in obtaining research funding; Finding innovative approaches to problem solving; Teams and groups produce more highly cited publications and patents than do individuals (Wuchty, Jones, and Uzzi, 2007); Teams and groups across disciplines were more likely to put novel combinations of prior work together, and to develop work that assimilated novel ideas into high-impact publications (Uzzi et al., 2013); highlighting benefits for research productivity and dissemination (Hall et al., 2012; Stipelman et al., 2014).

There are a large body of literature on important properties of the internal structure of research teams and groups from different perspectives, such as organizational, cognitive, and social, but little attention has been paid to current state of China’s scientific research teams. However, it is unclear that which factors at the PI or team level (e.g., leadership, team membership, geographic dispersion) influence the effectiveness of research teams, how current tenure and promotion policies acknowledge and provide incentives to academic researchers who engage in team science, what we know about the consequences of supporting research team. Current State and future development of China’s research team

methods This paper employed evidence-based evaluation methods and selected 1352 innovative research team as the research sample to investigate and capture the current state of China’s scientific research teams. Its questionnaire was designed for the academic leaders and key members from basic situation, team composition, operation mechanism, output performance and policy support role.

Collect funding data from official websites of NSFC, MOST, MOE and related funding agencies. Gain basic information from the websites of research teams, including demographic characteristics of PIs and members. Questionnaire Investigation and Field Investigation Conducted questionnaires and surveys from 2016 to 2017 Sent out 1353 questionnaires, 171 valid questionnaires among them Field interviews with 143 research teams,from 29 universities and research institute, 9 cities.

results or anticipated results four important prerequisites for shaping an excellent team are illustrated as follows: formation on the basis of long-term cooperation, rational structure of academic echelon with an outstanding academic leader, strong research ability and infrastructure, excellent culture of scientific research and effective operating mechanism.

1. Current team supporting programs are scattered, repetitive, lacking top-level design. 2. The national funding for excellent research teams is insufficient and sustained, which affects their long and stable development. 3. The evaluation system is unreasonable, which is not conducive to the formation and evolution of research teams. 4. Technical and experimental staff or talent in most of team are scarce, and it is difficult to recruit excellent postdoctoral fellow. 5. Team’s institutional and cultural construction is relatively lagging behind, and its operation mechanism is not smooth. 6. Excessive talent planning, undermining the normal scientific research and teamwork

Significance The policy recommendation to further improve and perfect China’s scientific research innovation team is proposed basing on the findings, including speeding up the implementation of central government’s science and technology plan, guiding team formation and operation, strengthening team support fund management, improving rules and regulations and operation mechanism, and it provides an important reference for the country to improve its support policies and plans for research and innovation teams.

Strengthen the management and overall coordination of Programs supporting teams from different channels of NSFC, MOST and MOE. Clarify the mission of their respective team plans, and avoid cross-duplication. In principle, these above programs do not fund the same team during the same cycle. Put forward the guidance on the construction and development of scientific research team at the national level, to publicize the team spirit, to emphasize the sense of the team, to let the concept of the team take root, and at the same time to set strict requirements on major issues such as team formation conditions, formation pattern, feasibility and development goals, in order to avoid the random construction of the team and excessive competition of resources. Improve funding management of programs supporting and improve the efficiency of team funds. Increase the funding intensity of the team plan and extend the funding cycle. Adopt a relatively stable funding model, funding cycle of up to 10-12 years or so, phased (around 3 years) to conduct an overall performance evaluation of the team, the evaluation for the outstanding person to provide continuous funding. Develop and evaluate broad principles and more specific criteria for allocating credit for team-based work to assist promotion and tenure committees in reviewing candidates. Guide professional development for leaders of science teams and larger groups.

cumulative empirical knowledge to assist scientists, administrators, funding agencies, and policy makers in improving the effectiveness of team science. Taxpayer investments in team science yield valuable returns. Scientists and leaders of teams and groups also need information on how to effectively manage these projects. Understanding formation and evolution of research teams and evaluating their performance have become more and more important for scientific community, policy makers, and other stakeholders.

Enabling combinatoric innovation

ABSTRACT. Background The process of innovating is a canonical example of science, technology, and society (STS) in which the environmental context is as important as the idea to the success of the new product. The innovation process involves several key dimensions: identifying a societal need, exploring and exploiting phenomena (explanations of nature) and technologies (embodiment of a use of phenomena for a purpose) in the ideation phase, an environment that supports the development of demonstrable ideas, and finally the market uptake that eventually determines the success of the product. As dynamics underlying these key dimensions have been shifting, so has the dominate style of innovation in particular areas. A continued focus on research by global society has enriched our understanding of nature yielding detailed explanations across a wide set of domains developed in increasingly deep specializations. Education and the pursuit of phenomena have also been shifting from a select elite towards a more democratized population which furthers the call for public accessibility of research results. This democratizing move is partially enabled through technologies, such as the internet, but also through changing societal norms of who could and should be educated. Manufacturing processes have been decreasing costs of technologies whilst levels of global income have been rising resulting in technologies that are increasingly available to a wider swath of the population. The simultaneous accumulation of phenomena and technologies coupled with increased accessibility and affordability has partially enabled the rise of new, combinatorial styles of innovation. New and emerging structural features have encouraged different modes of individual participation in the innovation process. As assessments of individuals through impact factors and increasing communication capabilities allowing global connectivity have risen, we have seen the role of the institution as the locus of excellence for intellect and innovation diminish. Hence, we now see multi-institutional teams arising that are capitalizing on multiple sets of expertise to develop novel innovations. Strong, motivating collective goals (Manhattan project, moon-shot, etc.) have been replaced by low-affect collective goals (human genome project, counter-terrorism, etc.) and a proliferation of individual goals, which dramatically shifts the context of the societal need. This change is correlated with decreased trust in societal governing bodies, higher needs for individual and collective accountability and return on investment (ROI), and lower tolerances for failure. Funding that enables demonstrable ideas has not only fractured into smaller amounts within this new reality, but it is also bound by metrics designed to ensure success. The context described above has led to changes in the process of innovation. Low resistance approaches that explore and exploit the now-available large store of available phenomena and technologies have led to a preponderate source of novel innovations. Combinatorial innovation processes, as opposed to discovery processes, can either use assimilation or compilation to relate disparate domains in a novel product. The emphasis of the ideation phase of the assimilative innovation process is exploration of phenomena through extrapolation, association, analogy, and inspiration to yield novel innovations; examples of the outcome of this process include bio-storage techniques (encoding of data into DNA), Geckel Glue (inspired by dry adhesion of gecko feet with the wet adhesion of mussels), and mechanical or biological logic gates in analogy to 0/1 computational logic (achieved through mechanical device flexure or protein production). Innovation through compilation exploits existing technology through assemblage, bricolage, and incorporation to develop surprisingly useful amalgamations; examples of this process include autonomous wheelchairs, the integrated circuit, and camera phones. Both of these combinatorial innovation styles minimize needed funding by capitalizing on existing phenomena and technologies. They thus increase potential ROI whilst conforming to a failure intolerant environment: multiple ideas can be pursued simultaneously or in rapid sequence, increasing the overall rate of success. The novelty of the innovation process is located in the combination of phenomena or technologies from disparate domains. Understanding how to develop, encourage, and nurture cohesive yet geographically and disciplinarily distributed networks is key to combinatoric innovation styles. The ability to nurture a combinatoric capacity in these networks – similar to the absorptive capacity found in institutions – is critical. The capability to sort the ever-increasing amount of both data and information, to identify relevant items in that global corpus, and then be able to convert those combinations into elements with social value is necessary to both assimilative innovation and innovation using compilation.

Methods To better understand how an institution fits into this new universe of research-defined social networks, this team explored Sandia National Laboratories’ combinatoric capacity. The team conducted semi-structured interviews with research staff at the labs. The sample consisted of ~30 research staff members who the team believed were employing combinatoric innovation styles. The interview protocol was designed to determine key differentiating attributes between combinatoric innovation styles by focusing on the subject (phenomena or technology in particular domains), the risk of failure, and the pace of innovation in the area. The protocol also addressed key aspects of the combinatoric capacity of the researchers. The protocol was designed to elicit information that would allow the team to describe the ideation phase, including how the researchers selected phenomena and technologies and the diversity of and degree of domain incorporation. Inquiries during the interview about the personal attributes needed to successfully stay abreast of and assimilate the external landscape for novel innovations would provide general suggestions about personality types most likely to succeed in this environment. Finally, the protocol was structured to provide information on how the researchers developed social networks to pursue the development of the ideas, products, and customers.

Anticipated Results The results of these semi-structured interviews are expected to confirm that the assimilation and compilation innovation styles are being employed by a subset of Sandia’s staff and that this subset is for an identifiable group of domains. The combinatoric capacity of these individuals and, to a limited extent, the absorptive capacity of the institution will be analyzed as the researchers describe their personal characteristics and the research processes they used when employing assimilation and compilation innovation styles. And lastly, the veracity of key perceived differentiating attributes between combinatoric innovation styles will be verified.

Significance Sandia National Laboratories is a respected federally funded research and development center (FFRDC) with the charter to maintain technical superiority and leadership in critical national security domains. It has a 60+ year history of scientific and technological research achievements. Many aspects of the combinatoric capacity challenge the traditional policies and operations at Sandia, and hence elevate the risk that Sandia may not be able to maintain leadership in domains where assimilation and compilation innovation are dominant. By building a deeper understanding of how individuals at Sandia are successfully employing these combinatoric styles and if and how the institution is responding to these new innovation styles, we hope to contribute to a broader discussion on the future of research institutions in the light of these new, emerging innovation styles.

The Importance of Dynamic Scientific Networks in Science Policy of Africa

ABSTRACT. Scientists are embedded in scientific networks that influence and are influenced by their collaborations with other scientists and the recognition they receive within their scientific network. According to Pouris and Ho (2014), single-author articles seem to be disappearing in Africa due to the effect of foreign funding sources. However, rising awareness has been seen on the collaborative research in the scientific network which has led to a sharp increase in the number of collaborations. In a networked environment, the position of each scientist plays an important role in diffusing knowledge. An author is more likely to know the relevant work of his co-authors and therefore cite their papers, resulting in a system in which relationships to other authors can lead to more citations. The network information of scientists helps us to broaden our understanding of how a scientist’s network is relative to performance measures. Some attempts have been made to incorporate network measures in publication and citation productivity (Otte and Rousseau, 2002; Yan and Ding, 2009), but this relationship has not yet been well-understood. Social network analysis is a topic that has become increasingly important over recent years. It is a connection of individuals that are attached to each other via links within the network. Identifying the important, influential scientists is a very essential task. In particular, we face high-time complexity in finding the influential scientist in a large network. Time complexity is of great importance in dynamic networks, and it is very useful to employ different centrality measures that have been developed over the years for such analyses. The co-authorship links offer an interesting case study to investigate the structure of scientific collaborations. The connection between scientists directly influences the measures of research impact. A basic issue faced in these collaborations is identifying how these scientists collaborate within the scientific networks. The size of co-authorship networks is growing, making their structures even more complex. Various measures are required to detect these structures and understand the collaboration patterns. The study of the complex network structures reveals how knowledge flows among scientists within the scientific networks. It also gives an insight into the knowledge sharing in networks that result in publishing articles. The sharing of ideas and experiences in networks have gained a central point of attention to detect patterns in such collaborations. Our study contributes to the state of knowledge on social network analysis use in scientific collaborations by focusing on the publication data from African scientists. Scientific co-authorship has become an interesting topic in recent years in Africa, and researchers try to shed light on the dynamics and motives of collaboration in the African research system. Pouris and Ho (2014) pointed out that collaboration patterns in this continent have been substantially higher in comparison to the rest of the world. They also found that during the five-year period of 2007-2011, the share of international collaborations grew by 66%. In this analysis, we aim to observe the co-authorship network of authors that published their articles between 2000-2015. In order to conduct our analysis, the data has been derived from the Leiden database of the Web of Science in the health field. We test the relationship between network measures and publication counts, as well as citation counts as an established measure of impact and investigate the performance of authors based on productivity and citation measures. This analysis thereby contributes to the current debate on co-authorship collaborations of scientists employing some network measures derived from social network analysis. Moreover, we investigate how international collaborations affect research performance in Africa. To perform this analysis, we build dynamic networks over five years to represent collaboration relationships and applied a variety of measures to investigate the impact of network measures on the research output of scientists. We use centrality measures and clustering coefficient measure from Social Network Analysis (SNA) for examining their effect on the performance of scientists in a Health discipline. We observe that scholars tend to collaborate more since the year 2000. With respect to the number of authors, the trend shows a steady increase in the average number of authors up to 2015. International collaborations (i.e. outside of the African continent) increase to as high as nine countries on average in 2015, which magnifies the role of international collaboration on African publications. These international collaborations are more likely to happen in small countries because the scientific community is small as they search for collaborators outside of the country (Narin et al.,1991). Large countries such as Algeria, South Africa, Libya, Egypt, and Sudan have fewer international collaborations compared with Somalia and Guinea, for example. Moreover, it is relevant for African countries to collaborate with countries outside of the continent that speak the same language such as English, French, and Arabic. Our results indicate that the network measures of scholars are a strong determinant of the collaboration of scholars, suggesting that all centrality measures play an important role in the productivity of their research. We found that if an African scholar has more connections or higher-quality connections, the likelihood of having more publications or receiving more citations or publishing in high impact journals is increased. Occupying central positions in the network enables authors to extend their relationships with co-authors as well as with the co-authors of co-authors, providing opportunities to exchange knowledge that could lead to higher performance and higher efficiency of research. Regarding the effect of closeness and betweenness centrality, we find that the intermediary position of scholars in the network or being at the shortest distance within the network from all other authors affects the research impact of authors. In order to improve this impact and benefit from research, authors can expand their collaborations; these collaborations have become more important over recent years. In fact, betweenness centrality and closeness centrality illustrate how globally central a scholar is and what level of strategic importance they can offer. We can determine that it is more beneficial if scholars have high closeness centrality (have short paths to other authors within the network) and a high degree of centrality if they wish to have direct, strong connections able to exchange knowledge that will lead to higher performance and efficiency. Moreover, with respect to the clustering coefficient, scholars who maintain relationships out of their clustered network perform better than authors who collaborate with scholars within the same cluster. Our results raise an important question for policy implications: while the co-authorship collaborations lead to higher performance, individuals still choose some level of co-authorship although it may not be optimal. This is due to the fact that individuals must address some constraints in collaboration and compromise with the ideas in the group. They also must manage the costs associated with the collaboration. There is a need for policies to consider uncertainty and organizational constraints that are involved in inducing scholars to participate in co-authored projects, particularly in Africa.

Aligning scientific impact and societal relevance: The roles of academic engagement and interdisciplinary research

ABSTRACT. Context Scientific findings from publicly-funded research are increasingly expected to demonstrate both scientific impact and societal relevance. Scientific impact is associated with achieving recognition within the community of scientists; while societal relevance is related to the capacity to respond to the needs of non-academic audiences. Despite the advocacy of policy discourses, the pursuit and achievement of this dual mission face important challenges.

Relevant literature There is a long-lasting discussion in the management and sociology of science literatures about the difficulty to reconcile scientific impact and societal relevance from scientific research (Amara et al., 2018; Bartunek and Rynes, 2014; D’Este et al., 2018a; Watermeyer, 2014). Some scholars argue that academics and practitioners hold irreconcilable views about what constitutes academic quality and relevant research, since the communities of academics and practitioners have different, often conflicting, methods, research agendas and priorities (Priem & Rosenstein, 2000; Kieser & Leiner, 2009). In contrast, other scholars, although acknowledging that academics and practitioners belong to different and potentially conflicting institutional logics, argue that differences are not only negotiable in the context of a particular research, but they are also susceptible to lead to greater relevance and better science (Hodgkinson et al., 2001; Gulati, 2007).

Literature gap This conflicting relationship between scientific impact and societal relevance has often been referred to by the expression ‘research-practice gap’. Our research focuses on two features of the knowledge production process that have been suggested in the literature as potentially conducive to closing this gap: (i) academic engagement and (ii) interdisciplinary research. First, we examine whether academic engagement in productive interactions with non-academic actors contribute to attenuate the potential tensions between scientific and societal goals, by shaping scientists’ cognition, skills and attitudes. Second, we investigate whether scientists who exhibit a stronger involvement in interdisciplinary research approaches are particularly capable to achieve greater performance in both scientific impact and societal relevance.

Research questions Our research questions deal with the roles of academic engagement and interdisciplinary research in aligning scientific impact and societal relevance. We expect that engagement via joint research is likely to be conducive to benefits associated to both scientific impact and societal relevance, since this type of interactions provide the mechanisms to arbitrate conflicting interests and respond to goals that meet the expectations of academics and practitioners alike. We also argue that interdisciplinary-oriented scientists are likely to benefit from enhanced scientific performance in terms of both scientific originality and potential applicability.

Data and methodology Our primary data derived from a large-scale survey of 57,406 scientists in the Spanish public research system. The population covers all fields of science including engineering and physical sciences (STEM), biology and medicine (BIOMED) and social sciences and humanities (SSH). We obtained a total of 11,992 valid responses. In addition to the survey data, information was collected from two secondary sources. First, altmetric data provides information on publication mentions in social media platforms. We have collected mentions to scientific articles from three social media platforms which try to cover non-academic audiences - i.e. blogs, news and policy briefs - as a proxy to capture societal relevance. Second, bibliometric data from WoS, which included the number of publications published by each scientist as well as the number of citations received by each paper in order to capture scientific impact. We run an econometric analysis in which we consider two dependent variables: societal relevance and scientific impact from research results, and two main independent variables: academic engagement and interdisciplinarity. We also include a range of control variables that cover individual and organisational characteristics that may influence scientists’ research performance, both in terms of scientific impact and/or societal relevance.

Results Our findings suggest that the involvement in joint research with non-academic actors and in interdisciplinary research teams contribute positively to the scientific researchers’ capacity to jointly reach societal relevance and scientific impact from public science.

In the case of academic engagement, we argue that this positive association is particularly linked to the participation of scientists in joint research activities with non-academic communities: what we call, co-production research modes. Co-production modes favor the mobilisation of bidirectional flows of knowledge in research activities. These bi-directional flows of knowledge and expertise are likely to enhance a mutual awareness of gains from research results from both scientists and practitioners and, eventually, facilitate the alignment of research incentives and goals among these two communities. In the case of interdisciplinary research, we argue that the positive association is a result of the access to a diverse range of knowledge sources and research perspectives. A greater diversity of cognitive perspectives favors the capacity of scientists to integrate pluralistic perspectives about what is worth investigating. And also, greater cognitive diversity helps embracing a more reflexive approach on how research goals can be achieved in most effective and socially responsible ways, making scientific findings better informed and closer aligned with societal demands.

Our results also support the presence of highly heterogeneous profiles among the population of scientists. Whereas some scientists achieve impact within scientific communities, others achieve greater visibility among non-academic audiences, while still others produce research results which reach both the communities of scientists and practitioners.

Contribution to scholarship and practice Our study provides consistent empirical evidence to an on-going discussion in the management and sociology of science literatures, about the capacity of scientists in publicly funded research organisations, to deliver results that are both scientifically outstanding and socially relevant. Reaching both the communities of scientists and non-academic users in scientific research is far from straightforward. Our paper investigates factors associated with knowledge production processes that contribute to reconcile these two missions. The results of this study have important policy implications, since they inform on modes of research that might be particularly conducive to integrate distinct research logics, and to overcome the challenges of pursing research goals to reach the communities of scientists and practitioners.

Bibliography Amara, N., Olmos-Peñuela, J. and Fernández-de-Lucio, I. (2019). Overcoming the “lost before translation” problem: An exploratory study. Research Policy, 48(1), 22-36. Bartunek, J.M. and Rynes, S.L. (2014): “Academics and practitioners are alike and unlike: the paradoxes of academic-practitioner relationships”, Journal of Management, 40(5): 1181-1201. D’Este, P., Ramos-Vielba, I., Woolley, R. and Amara, N. (2018a): “How do researchers generate scientific and societal impacts? Toward an analytical and operational framework”, Science and Public Policy, 45(6), 752-763. Gulati, R. (2007): “Tent poles, tribalism, and boundary spanning: the rigor-relevance debate in management research”, Academy of Management Journal, 50 (4): 775-782. Hodgkinson, G.P., Herriot, P. and Anderson, N. (2001): “Re-aligning the stakeholders in management research: lessons from industrial, work and organizational psychology”, British Journal of Management, 12: S41-S48. Priem, R. L. and Rosenstein, J. (2000). Is organization theory obvious to practitioners? A test of one established theory. Organization Science, 11(5), 509-524. Watermeyer, R. (2014): “Issues in the articulation of ‘impact’: the responses of UK academics to ‘impact’ as a new measure of research assessment”, Studies in Higher Education, 39: 359-377.

10:30-12:00 Session 7D: Human Capital Indicator Challenges
Location: GLC 225
Core Indicators of U.S. S&E Higher Education

ABSTRACT. Talk for the session: Science and Engineering Indicators: Human Capital Challenges for S&E

The U.S. higher education system consists of diverse academic institutions—including research and doctoral-granting universities, minority-serving institutions, community colleges, and others—that train students in S&E across degree levels and fields. Public institutions award 70% of S&E degrees. Public institutions award a majority of S&E degrees. In 2015, a small number of institutions with very high research activity —public and private— awarded 72% of doctorates, 42% of master’s degrees, and 37% of bachelor’s degrees in S&E fields.

This presentation, based on the S&E Indicators report to be released in the summer of 2019, will provide a portrait of S&E higher education in the United States, including trends over time and comparisons with other nations. The report is divided into four main sections. The first section provides an overview of the U.S. higher education system, including minority-serving institutions, community colleges, and for-profit institutions, as well as distance and online education. This section also provides information on sources of aid for undergraduate and graduate S&E education, with a focus on the federal government’s role. The second section looks at trends over time in S&E degree awards at the undergraduate and graduate levels, highlighting patterns by field. The third section focuses on the demographic attributes of S&E degree recipients, including sex and race and ethnicity. It examines trends by degree level and field. The final section focuses on international S&E higher education. This section provides data on students on temporary visas who study or earn degrees in the United States. It also benchmarks the United States with other nations in terms of S&E degrees awarded.

STEM K-12 Education Indicators
PRESENTER: Karen White

ABSTRACT. Session Title: Science and Engineering Indicators: Human Capital Challenges for S&E Session organizer: Beethika Khan

Abstract: Elementary and secondary education in mathematics and science is the foundation for student entry into postsecondary science, technology, engineering, and mathematics (STEM) majors and STEM-related occupations. In the elementary and secondary grades, U.S. students discover STEM and cultivate the mathematics and science knowledge that enable them to succeed in pursuing STEM-related careers—both career options that require postsecondary degrees as well as skilled technical career options that may not require a degree. This presentation will provide a portrait of K−12 science, technology, engineering, and mathematics (STEM) education in the United States, based on the S&E Indicators report scheduled for release in the summer of 2019. It will include comparisons of U.S. student performance with that of other nations, by discussing indicators of pre-college mathematics and science learning and how that learning affects postsecondary and career outcomes. Data sources are from the National Center for Education Statistics (NCES) of the U.S. Department of Education and other public sources. This report focuses on overall patterns in STEM education and notes variations in STEM access and performance by students’ socioeconomic status, race or ethnicity, and sex. There are two main sections in this report. The first presents indicators of U.S. students’ performance in STEM subjects in elementary and secondary school. It begins with an analysis of elementary school students’ growth in mathematics and science knowledge over time. Next, it presents national trends in mathematics, science, and technology and engineering literacy (TEL) assessment data for eighth graders. This section then places U.S. student performance in an international context. The second section focuses on transitions from high school to postsecondary education or directly into the workforce. It presents national data on Advanced Placement (AP) coursetaking, immediate college enrollment after high school, and students’ choice to major in a STEM subject in college. This section also examines the transition to the skilled technical workforce (STW) for students who enter the workforce immediately following high school.

Science and Engineering Labor Force

ABSTRACT. Session Title: Science and Engineering Indicators: Human Capital Challenges for S&E Session organizer: Beethika Khan

Abstract: Scientists, engineers, and technologically-skilled workers make important contributions to improve a nation’s living standards, economic growth, and global competitiveness. These workers fuel a nation’s innovative capacity through their research, development, and other technologically advanced work activities. This presentation, based on the S&E Indicators report to be released in the summer of 2019, will provide an analytical overview of the U.S. S&E labor force. A variety of data sources are used throughout, demographic and workforce surveys conducted by the National Center for Science and Engineering Statistics (NCSES) are a major source of data for this report. Topics covered in this report include the definition, size, and growth of the S&E workforce as well as career pathways from education to employment. Employment sector and industry, salary, and unemployment rates provide information on where S&E workers are in the economy and their labor market conditions. A particular focus will be a new analysis of the role of the skilled technical workforce in the US, defined as those technical workers whose jobs do not require a bachelor’s degree. Internationally comparable data, although limited, provide strong evidence of a widespread, though uneven, growth in the S&E workforce of the world’s developed nations. The emphasis on developing S&E expertise and technical capabilities has been a global phenomenon.

Science and Engineering Indicators Reimagined: S&E Indicators for the Digital Age
PRESENTER: Beethika Khan

ABSTRACT. Session Title: Science and Engineering Indicators: Human Capital Challenges for S&E Organizer: Beethika Khan

Abstract For many decades Science and Engineering Indicators (Indicators) has served as the National Science Board’s (NSB’s) flagship report on indicators of the state of the U.S. science and engineering enterprise. It is produced by NSF’s National Center for Science and Engineering Statistics (NCSES) on behalf of the Board and transmitted to Congress and the President on or before January 15 of even-numbered years in fulfilment of statute.

The voluminous report covers U.S. and international data on all levels of STEM education, the STEM workforce, R&D investment, knowledge and technology industries (e.g., high-tech manufacturing), innovation, public attitudes towards and knowledge of science and technology, and S&E-related state indicators. The report uses a combination of graphics and explanatory text, organized in thematic chapters, to convey information.

Indicators has had important role informing science policy over the years; it was widely cited in 2018 for providing objective and compelling evidence of the rise of S&E activity in China and Asia. However, the 2018 edition spanned more than a thousand total pages of text and figures; its sheer volume limiting it ability to communicate effectively.

NCSES and NSB are redesigning Indicators for the 2020 cycle. This presentation by the Indicators program director will explain the rationale for this decision and the changes that NCSES envisions for presenting data users with high quality, policy relevant statistical data and analysis.

10:30-12:00 Session 7E: Universities & Innovation II

Triple Helix

Location: GLC 222


Science commercialization in the context of US research universities suggests that strategic human resources management (SHRM) theory needs to be revisited. The primary reason being that the organizational and human resources strategies of many US research universities have increasingly diverged; the secondary reason being that this won’t be changing anytime soon. Using qualitative and quantitative information from a purposive sample of 21 university-industry research centers established by the US National Science Foundation, we present evidence that what the SHRM literature refers to as “technical” HRM practice is perhaps more important to organizational performance than what the SHRM literature refers to as “strategic” HRM practice. Related, the cases suggest SHRM theory should expand to include constructs from the study of leadership styles and network management approaches.

Science commercialization isn’t new to US research universities. Despite the ivory tower characterization, research universities in the US have long played a critical role in the transfer and application of new knowledge and technologies to/in the marketplace, starting with agriculture (Mowery & Rosenberg 1989). What’s new is the adoption by US research universities of formal organizational strategies for science commercialization. The en masse establishment in the 1980s of technology transfer offices on (or near) American campuses (Geroski 2000) notwithstanding, more recently US research universities are involving industry sponsors and advisers in the development of new curricula and in the establishment of new academic programs and research centers (Jarvis 2013). Perhaps the best-advertised example to date is Arizona State University (see Crow & Dabars 2015).

But as far as we can tell organizational strategies for science commercialization haven’t been accompanied by human resources strategies for science commercialization. Despite US research universities’ willingness to involve firms in facets of academic business that historically have had little to no direct industry involvement (Beckman et al. 1997, Jarvis 2013), academic faculty are still predominantly incented to perform open science over proprietary science. Though industry contracts, patents, and invention disclosures count for something in performance reviews of academic faculty (Azoulay et al. 2006), publishing in the open literature and the grants that enable them to do so have been (O’Meara 2005) and remain (Boardman 2016) the type of academic production that US research universities value first and foremost. Which makes US research universities unique units of observation for organization and management studies, specifically for SHRM theory. They’re unique because, unlike conventional firms and government agencies, the core or “A” workers (Huselid & Becker 2005) – in US research universities, academic faculty – are governed predominantly by professional norms and expectations for behavior, not organizational ones. It was De Solla Price (1963) and then Polanyi (1967) who first implied as much, characterizing academic scientists and engineers as governed first and foremost by one another and next by the national interest, with no mention of the organizational strategies of higher education institutions. And for good reason, because, at least then, the organizational strategies of US research universities were one-and-the-same with the norms and expectations of the broader profession – both emphasizing open science as premier. The resilience in US research universities of open science human resources strategies in the face of emergent organizational strategies for science commercialization presents a new contingency for SHRM theory to address (e.g., Sydow et al. 2012, Feller et al. 2013, Paradeise & Thoenig 2013) – because such resilience is counter to SHRM’s key proposition that the alignment of organizational and human resources strategies is a key predictor of organizational performance (Ferner et al. 2004, Morris et al. 2006, Cascio & Boudreau 2012). A number of US research universities with open science human resources strategies are no less quite successful at science commercialization – in terms of spin-offs (Lockett et al. 2005), patenting (Seigel et al. 2003), inventions (Czarnitzki 2012), and so on – and especially the institutions that are world leaders in open science (Mowery et al. 1999). We posit collaboration managers in US research universities who are successful at science commercialization to use what the SHRM literature refers to as “technical” HR practices to do so. Such as performance reviews and the basic personnel authorities performance reviews entail (Huselid et al. 1997). Because what the SHRM literature refers to as “strategic” HR practices are more or less constant across US research universities, including but not limited to team-based job designs, path-goal leadership, flexible work environment, employee empowerment, and pay-for-performance (Cyr & Schneider 1996, Laursen 2002, Cano and Cano 2006, Souitaris 1999, Itoh 1994). Therefore we expect technical HRM practices for science commercialization, especially personnel authorities at the lab- and project-levels, to vary across US research universities. For instance, science commercialization activities and outputs like technology transfer and patents may count for more in some academic faculty performance reviews (e.g., for promotion and tenure) than in others (if at all). Moreover, some collaboration managers may hire non-academic faculty first and foremost to perform and produce science commercialization activities and outputs and thusly review these faculty (but not academic faculty) per their respective science commercialization behaviors and productivity.

Which makes the appropriate level of observation for this study the project and/or lab, not the university. To investigate our expectation that collaboration managers are developing technical HR practices to elicit science commercialization behaviors and outputs, we draw on 21 case studies of university-industry research centers, all established by the US National Science Foundation. We employ a deductive-inductive approach (Maxwell 2005, Miles & Huberman 1994) to analyze qualitative and quantitative information and match patterns across the cases (Ritchie et al. 2003) to detect observations that refute our expectations as well as those that meet our expectations.

Effectiveness of technology transfer in public research institutions in South Africa: A review of national survey data using the revised contingent effectiveness model, and implications for future measurement
PRESENTER: Nazeem Mustapha

ABSTRACT. ABSTRACT presented at the 2019 Atlanta Conference on Science and Innovation Policy


How effective is technology transfer at public research institutions in developing countries? We argue that a narrow definition of economic impact alone to define effectiveness is insufficient when assessing performance. Rather, policy makers should use multi-pronged evaluation criteria to assess, through measurement, the multiple outcomes of technology transfer investments and their impact pathways. In order to suggest some answers to these questions, we use the case of South Africa, where we examine data from the first national baseline survey of intellectual property and technology transfer at publicly funded research institutions (PRIs) in South Africa through the lens of the revised contingent effectiveness model of technology transfer. The survey covered the period 2008 (when government first passed intellectual property rights legislation specifically aimed at PRIs) until 2014.

Background and Context

In April 2017, the Minister of Science and Technology of South Africa, Naledi Pandor, launched results of a national study into intellectual property (IP) and technology transfer at publicly funded research institutions. Pandor reflected on a selection of key indicators showcased in the research report—technologies managed, licenses executed, revenue generated, start-ups formed, as well as the skills levels of staff employed within technology transfer offices—concluding with modest praise and encouragement to all concerned:

"What these results show we are making progress, incremental but progress nonetheless, which is exciting. We are beginning to see enhancements to economic impact from public investment in R&D—and that’s what we all really want for our country."

These highlighted indicators present an assessment of the state of IP and technology transfer at technology transfer office in South Africa within a “taken-for-granted” conceptual model of what that process is and what its assessment criteria are. However, there are many reasons to question this notion of a well-performing public IP protection and technology transfer system. Even more so in a developing country where the needs of researchers and inventors at public institutions must, by necessity, be balanced with those of the majority of the populace.

The opportunity of future surveys of IP and technology transfer in publicly financed research institutions presents fresh possibilities for refining the specific criteria against which policy makers measure the effectiveness of the system, and the assessment frameworks they use to discuss the effectiveness of publically generated IP and technology transfer. On that front, Alessandrini and others have argued that measuring the effectiveness of technology transfer in a country such as South Africa must take into account the country’s “socio-economic and institutional specificities”, though they stop short of providing a clear path in this regard. Nyatlo has developed a framework for effective technology transfer offices. This paper contributes to this discussion. There are two key and interrelated tasks. The first is to define and develop a concept of effectiveness that is rooted in evidence. The second is to ascertain the socio-economic and institutional specificities in which the practice is carried out so that the definition can be context-specific.

This article aims to contribute to these two tasks, by reflecting on aspects of the five decades of US experience of protecting IP and transferring technology generated by its public universities and laboratories. In what ways can South Africa benefit from the US experience, when it comes to effectiveness of technology transfer? A first learning opportunity between the US and South African situation is rooted in policy. South Africa’s IPR Act adopts similar principles to the exploitation of publicly financed IP as its equivalent Bayh-Dole Act or Patent and Trademark Law Amendments Act (Pub. L. 96-517, December 12, 1980). The Bayh-Dole Act was developed in response to economic challenges in the US—a period of high inflation and unemployment (stagflation). Perhaps analogously, South Africa’s persistent poor economic performance in recent years, including its high unemployment rate, has prompted calls by policy-makers at the highest-level for innovation-led growth and development.

Revised contingent effectiveness model

The contingent effectiveness model of technology transfer (‘the model’) was developed in 2000 by Barry Bozeman, then based at the Georgia Institute of Technology’s School of Public Policy. Bozeman proposed this model as a way of organising a proliferating body of literature on assessing performance of technology transfer arising from the activities of researchers and practitioners within universities and state research facilities in the United States. The model recognises, as Bozeman writes, “that parties to technology transfer have multiple goals and effectiveness criteria,” and among the model’s explicit purposes therefore was the elaboration of a more integrated understanding of effectiveness.

"Unfortunately, many studies of technology transfer never make clear what is meant by effectiveness and seem simply to assume that we all hold some unspecified unitary concept of effectiveness [...] This assumption is wrong, as we have shown with both statistical (sic) and case study (sic) evidence."

On the input side of the original model, as it were, Bozeman proposed five determinants of effectiveness, namely: characteristics of the transfer agent; characteristics of the transfer media; characteristics of the transfer object; demand environment; and characteristics of the transfer recipient. On the output side, it proposed six criteria against which effectiveness is typically but not uniformly assessed in the literature. These criteria are opportunity cost; scientific and technical human capital; political; economic development; market impact; and “out-the-door”. The revised model adds a single additional criterion: public value. As Bozeman et al write:

"The addition of the Public Value criterion arises from the recognition that transfer agents, particularly public sector transfer agents but others as well, are housed within agencies and organizations that are themselves in pursuit of broad public-interest goals. Thus, their endeavors are motivated, influenced, and directed by ever-changing constellations of public values."

They go onto argue that since the Public Value criterion to some extent counterbalances the emphasis placed on economic impacts. It also brings into the narrative elements such as sustainability, health and safety, equity and inequality, and the improvement of quality of life through addressing societal needs and grand challenges.

The paper goes on to address the performance of the South African national public IP and technology transfer system using this framework from both the economic dimension and by comparison with the National Development Plan and the Sustainable Development Goals.


There are key issues that arise from a mapping of indicators to the revised contingent effectiveness model. We present preliminary results from our analysis which brings out this complexity from data from the baseline survey and other data sources.

There are multiple and competing discourses about the particular role of these institutions, and their contributions within different economic and social spheres. Concepts such as the “entrepreneurial university”, “neo-liberal university”, “decolonised university”, or the “university’s third mission” have emerged, reflecting some uncertainty within academia about where a university’s institutional purpose begins and ends. As institutions seen to be pivotal to South Africa’s development trajectory, universities and science councils, which are the subjects of the IPR Act, are compelled to, as Kruss and others argue, “balance multiple mandates”.

The University of Pfizer: How Pharma Uses Translational Science to Offload Risk to Universities

ABSTRACT. Background & Rationale

Since 2005, translational partnerships with universities helped Pfizer save nearly hundreds of millions of dollars as part of its cost-cutting strategy. This financial feat was achieved, in part, by the way that translational partnerships help Pfizer’s short-term R&D strategy. Yet, what this fact shows is that there is a separate history that marks the explosive growth of translational research. A dive into this history reveals that existing analyses of translational research largely misunderstand how it is actually unfolding on the ground.

Translational science and medicine has significantly changed the contours of academic biomedicine and global science innovation. Most research on this issue concern the pace and efficiency of translation, which processes support ease of collaboration between industry and academia, and the myriad legislative and financing concerns involved. Yet, could these be the wrong questions? Corporate partners now seek out translational partnerships to achieve short-term financial goals and as a remedy for problems relating to larger corporate strategy.


The result of a multi-sited, multi-year ethnographic research project alongside interview data and market analyses, this project is among the first on-the-ground studies of translational research.

Results & Significance

Taken together, the evidence from this study shows that despite policies focused on accelerating the development of safe and effective therapeutics, the on-the-ground reality of translational research indicates an entirely different reality. Additionally, the injection of translational imperatives is slowly transforming laboratory practices and experimental designs. Thus, a more comprehensive view of translation requires understanding elements that lay beyond the university and its laboratories, including transformations in the pharmaceutical sector and the way that the quest for short-term shareholder value finds its way into the agendas of university laboratories.

Revisiting the agent roles of the technology transfer office in Taiwan
PRESENTER: Shih-Hsin Chen

ABSTRACT. Under the influence of the U.S. Bayh-Dole Act, Taiwan also followed this global trend and legislated the “Fundamental Science and Technology Act” in 1999. Since then, a series of laws were enacted to allow university and academic scientists to privatize the outcomes of government-funded research. While many studies focused on the determinants of performance and outputs, little attention was emphasized on the trilateral interactions among the three major players in university commercialization activities: the university administration, the faculty, and the Technology Transfer Office (TTO). Drawing upon the agency theory, the dilemma and obstacles of TTO as a dual agent are examined in this qualitative study with 29 interviewees from nine universities in Taiwan. The results of this research have practical and theoretical implications for university technology transfer policy and the limitation of the incentive system for TTOs built within the universities.

10:30-12:00 Session 7F: Changing Frames for Research on Research
Location: GLC 324
Discourses, politics and practices of UK research on research

ABSTRACT. NB. This paper is intended to form part of the panel (see paper 226) on “Changing ingredients of research on research: old wine in new bottles, or a fresh recipe for success?”

Launched in April 2018, the new mega-agency UK Research and Innovation has pledged to create an evidence-informed “culture of evaluation” at its heart. A dedicated team of data analysts will lead this work through a “UKRI Data Hub”.

For a country that channels in excess of £6.5 billion a year of public funding through UKRI, the UK invests little in analysing how effectively our research system is working, testing different approaches and learning from innovations elsewhere. Individual research councils within UKRI have all grappled with these issues, and some have built their own in-house capacity. Outside of government, Nesta has contributed new frameworks for measuring and making sense of the wider innovation landscape. And the Research Excellence Framework (REF) is, of course, a large and resource-intensive process of evaluation, although it has other purposes too.

Systematic analysis of the research system remains patchy and underfunded in the UK – and is overly reliant on thinly researched reports by consultants, or reviews that lean on the authority of their chairperson in lieu of actual data and evidence. There has never been a UK equivalent of the NSF Science of Science and Innovation Policy programme in the US.

However the promise of UKRI is the greater strategic coherence it will bring to policies and priorities across the system. This will require more analytical firepower within UKRI itself – but also more distributed and sustainable capacity across the research system. Against this backdrop, this paper uses the idea of discourse coalitions (as developed by Maarten Hajer) to map and make sense of the growth and institutionalisation of RoR, and understand the synergies and tensions between competing perspectives.

‘Lost in translation? Post-Brexit futures for UK research’

ABSTRACT. Even with the rise of research economies such as China and the strengthening of research across the Eurozone, the UK continues to be one of the world’s most productive and high-impact research economies. However, political and cultural instabilities introduced as a result of the approach taken by the UK government to the UK’s exit of the European Union puts much of the UK research base at risk. A 2016 report by Digital Science shows that EU funding has become structural [1]. Deficits in the UK’s research funding model combined with the reluctance of business to invest in R&D have left the UK lagging with only 1.66% of GDP being invested in R&D, well behind the EU (1.96%) and OECD (2.37%) averages [2].

Digital Science’s Dimensions database allows analysis of the full research lifecycle, from grant, through publication and attention to patent, policy and beyond. A study of emergent and young research areas in the UK highlights concerning trends. Namely, that while the UK performs well on the international stage in both volume and citations for high-risk or early-stage research, this classification of research tends to attract, and hence to gain its predominant support from EU rather than UK-based sources. It remains to be seen whether the UK government will be able to live up to its commitment to match or exceed the funding gap generated by its departure from the EU. However, it is not merely a question of amount of money but also of funding areas of research that show the greatest promise. After Brexit, there is a growing risk that many strengths of the UK research base are lost in translation.

1 Hook, DW and Szomszor, M (May 2016) Examining Implications of Brexit or the UK Research base. Digital Research Reports. doi: 10.6084/m9.figshare.3383368.v5 2 2017 data. Accessed 15th March 2019.

A new agenda for Research on Research: combining disciplines, mixing methods
PRESENTER: Sarah de Rijcke

ABSTRACT. A new agenda for Research on Research: combining disciplines, mixing methods

Our understanding of (changes to) steering mechanisms in science has benefited greatly from the research done in Science Policy Studies, Science and Technology Studies, and Scientometrics. However, to date there is no interdisciplinary convergence in the sense of a strong collaborative agenda. In this presentation I will discuss how combining the toolboxes of Science Policy Studies, Science and Technology Studies, and Scientometrics may lead to a more fundamental understanding of both macro-level developments in science policy and practices of knowledge production at the organizational and individual level. A collaborative agenda will also help develop better policies and practices for promoting responsible research evaluation in research policy and research organizations.

10:30-12:00 Session 7G: Green Innovation
Location: GLC 158
Supporting New Ventures in Clean Energy Technology: The Role of ARPA-E
PRESENTER: Anna Goldstein

ABSTRACT. The private sector underinvests in clean energy technologies due to a double externality: large social benefits from both knowledge spillovers and environmental protection that cannot be captured by individual companies or investors. There is an opportunity for public support to fill this gap and assist clean technology (cleantech) startup companies over the “valley of death.”

In 2009, the Advanced Research Projects Agency – Energy (ARPA-E) was created within the U.S. Department of Energy (DOE). ARPA-E is unique among funding offices at DOE in its high-risk, high-reward approach to R&D on advanced energy technologies. ARPA-E program staff are empowered to craft solicitations around specific technological challenges, creating a bottom-up R&D portfolio. Evidence on ARPA-E to-date has shown enhanced productivity in terms of publications and patents, but these metrics are insufficient to measure progress toward market adoption of new technologies. The question remains of how well ARPA-E investments have performed in terms of stimulating private sector innovation.

In this study, we examine the trajectories of ARPA-E-funded startup companies, compared to startups funded elsewhere in DOE and the universe of similar cleantech companies. We take advantage of the surge of DOE funding in the economic stimulus package, which led to a large number of startups (companies founded within the previous 5 years) funded in fiscal year 2010. We collect data on patenting activity and private equity fundraising for all companies funded by ARPA-E, the Office of Energy Efficiency and Renewable Energy (EERE), and the DOE Small Business Innovation Research (SBIR) program in 2010. For a comparison group, we also collect data on cleantech startups in the i3 Cleantech platform from the same time period, as well as a set of startups that applied for—but did not receive—ARPA-E funding in 2010. We categorize each company in our dataset according to their particular subsector within cleantech. Controlling for sector and founding year cohort, and using a variety of specifications, we use regression analysis to compare the outcomes for ARPA-E awardees to other startups.

Initial analyses show dramatic differences between ARPA-E startups and other companies in terms of inventive activity. Startups that were funded by ARPA-E in 2010 had filed more successful patents as of 2014, compared to similar companies. We also find that these companies had filed more successful patents prior to 2010, the year they were selected for funding. Nonetheless, even controlling for 2010 patenting activity, ARPA-E awardees still had an advantage in post-award patenting, filing 3-4 times as many successful patents on average from 2011-2014.

This relatively high rate of patenting from ARPA-E startups is not shared among startup awardees from other DOE offices. Startups funded by EERE in 2010 roughly matched the performance of similar companies in terms of patenting, while those funded by SBIR patented much less than similar companies. Rejected ARPA-E applicants also patented at the same level as similar companies, demonstrating that the advantage for ARPA-E awardees did not arise from self-selection of ARPA-E applicants.

We also find that the private funding trajectories for ARPA-E’s 2010 startups differ significantly from the comparison set of cleantech companies. ARPA-E funded startups were more likely to have already raised private equity prior to 2010. Controlling for this effect, however, we measure no significant difference between ARPA-E startups and similar companies in terms of private financing from 2011-2017. Meanwhile, startup awardees from EERE and SBIR raised significantly less private money post-2010, so the fundraising activity of ARPA-E companies represents a relative advantage among DOE awardees.

Analysis of the technological focus of these companies demonstrates another important difference between ARPA-E and other DOE offices. Energy storage, which has been identified as a high priority for energy R&D spending, was fully one-third of ARPA-E’s portfolio of 2010 startups, compared to 3% for EERE. Energy storage companies were a relatively small portion of the cleantech startup population at that time. This suggests that ARPA-E’s bottom-up portfolio approach allows them to pay special attention to technology areas that may be neglected by private investors.

This presentation will explore our results in detail and discuss the implications of our findings for government programs that support entrepreneurial ventures in energy technology. The evidence on ARPA-E adds to the recent literature showing positive impacts of government intervention on knowledge generation and cleantech business success.

Mission-oriented research policies impact to research themes: case fuel cell technology

ABSTRACT. There is a significant body of literature analyzing the impact of science and technology policy has on research. Central to it are the concepts of mission- and diffusion-oriented policy (Ergas 1987) that frame the approach used by the national innovation system towards research. Notwithstanding the changes that have occurred in what is perceived as mission-oriented policy, particularly the shift towards ``grand challenges'' (Cagnin 2012), there is a clear dynamic that mission-oriented impacts research agenda more narrowly, whilst diffusion-oriented strives towards promoting network creation.

Studies have looked in detail how policy, particularly on research funding, impacts science. Research has focused on the impact of personal development grants on future research output and citation. Studies have also looked at the increase in research collaboration created through particular funding. Literature on how communities of research grow and change thematic orientation based on funding are less prevalent in literature. Although literature on research motivation suggests that in addition to the prime motivations of gold, ribbon and puzzle, a significant portion of researchers migrate and adjust their research agenda based on funding opportunities. Particularly in the case of mission-oriented research this should be central as the policy intervention is built on the notion of favoring a specific development path while putting alternative research trajectories in a disadvantage (Cantner 2001).

If we look at the research agenda setting and research uptake in isolation, without the policy intervention, Ayres (1969) put forward a framework of a self-organizing dynamic process of research. The framework, adapted from Holton (1962) is founded in the idea that the progression of technology and selection of research topics by the research is a function of the availability of novel ideas to solve. Holton describes that there is a finite space, a lode, of interesting idea within a field. As scientist opens a new space through scientific discovery, we can expect a 'gold rush' (Ayres 1969) where scholars "defect from their old field, in search for greener pastures." However, as the finite pool of ideas in a field becomes nearly exhausted, new discoveries become scarce and challenging researcher again search for new branches in the tree-of-knowledge.

It is clear that the process of a researcher pursuing a topic is much more complex than the framework by Ayres (Laudel 2014). We, for example, know that there is a counterforce, where a researcher looks for diversity in order to differentiate from other scientists. Prior work have also focused on research community building through co-authorship (e.g Suominen 2014), co-citation (e.g. Boyack 2010) and bibliographical coupling (e.g. Jarneving 2007). Studies have addressed thematic orientation through authors sharing terminology and create persistent new topics that might be emerging topics (e.g. Suominen 2015). The number of entrants and the cohesiveness of a research field has also been studied (Suominen 2013). However, the dynamics of research activity and thematic orientation of research particularly in the case of a clear mission--oriented policy intervention has seen little research.

The fundamental question in this study is, how mission--oriented policy shifts research participation and thematic agenda of research? As an empirical case, fuel cell technology was chosen, as it is a technology that has undergone several phases of high and low research policy priorization. In particular, the study focuses on the shifts in US policy, with its shifting priorization of fuel cell technology. Based on our analysis, we show how, mission driven policies change thematic orientation and pick-up in research participation, but also how a shift in policy is quickly responded by the scholarly community. The data used in the study the thematic changes of fuel cell research is sourced from the ISI Web of Science by using a query ``fuel cell'' or ``fuel cells'' being mentioned in the title, abstract, descriptors or identifiers of a publication. The data for the analysis has been retrieved in partitions; first search was done in 2009 and subsequently updated. The last update for the study has been done in the fall of 2016, which explains the low publication volume for the year. This process resulted in a data set of 75 479 articles, from which the bibliographical data was downloaded.

The analysis femployes Latent Dirichlet Allocation (LDA) to classify publications in themes. LDA, is a unsupervised learning algorithm that creates an outcome solely relying on its formal framework. Studies have shown LDA to uncover patterns from text collections. Since the original findings of (Blei 2003, LDA has been shown to produce coherent results with S&T relevant data sources such as scientific publications (Yau 2014).

In practice, LDA produces a user defined number of themes, based on optimization using KL divergence, represented by word probabilities. Word probabilities describe the content of each theme. To qualitatively evaluate the content of topics, a wordcloud was created for each theme. Wordclouds have been seen as a practical method to communicate the content of the LDA topic (Suominen 2015, Suominen 2016} and by qualitatively analyzing the wordclouds, the content can be reduced to a label representing the thematic content of a topic. For this study 24 were created and drawn as a wordcloud, there after labelled by the researcher. By using the pivot table of topic probabilities per year, topics are ranked by their significance on each year of the time series.

To understand changes in the thematic orientation of the fuel cell research, the topics dynamics was analyzed for time series breakpoints and clusters. For breakpoint detection the approach developed by Bai et al. (2003) implemented in the R package strucchange. Their approach uses an optimal partitioning approach to find one or more breakpoints in the time series. The breakpoints were used to understand if there are strong thematic changes with in the time series topics. The findings of this research make clear that mission oriented policy agenda has a strong impact in both researcher themes. During the time period analyzed, fuel cells were center stage in the US mission agenda with president George Bush Jr. Freedom car initiative. This time coincides with a strong increase with strong upward trend in “vehicle” related themes. To further emphasize the impact, these themes decline as president Obama’s administration cuts funding for fuel cells. There seems to be a strong policy push towards a particular theme of fuel cell research.

In comparison to technology push or market pull, policy push is a seldom researched topic. In the context of innovation economics and eco-innovations, the existing traditional innovation economics discussion on technology push and market pull has been extended to include a regulatory aspect (e.g. regulatory push/pull or policy push/pull) (e.g. Cleff 1999). The technological trajectories of policy driven technologies, eco or not, have several underlying challenges which should be further studied.

Governing Competing Technological Innovation Systems for Sustainability Transitions – Lessons from the German Renewable Energy Feed-in Tariffs
PRESENTER: Carsten Schwäbe

ABSTRACT. Background

The new mission orientation in innovation policy puts emphasis on technological openness for the resolution of societal challenges. Hence, several technological options are considered and potentially supported in order to create experimental market niches. However, this support cannot be permanent: Returns from specialization and economies of scale require a decision of the state for a lock-in to one option or to a specific mixture of several options. The German feed-in tariffs for renewable energies (RE) are a typical example for such a mission-oriented innovation policy instrument at the demand side [1, 2]: Different technological options for a green energy transition are subsidized by technology-specific feed-in tariffs aiming at achieving specific market share objectives at the German electricity market. In order to avoid a too early lock-in or an excess support, the feed-in tariffs are dynamically adjusted to the evolution of the specific innovation processes. Furthermore, theory and empirical analysis of mission-oriented policies for sustainability transitions lack of the consideration of both sides of the coin: innovation considering several technological options and exnovation aiming at the phase-out of dominant, but unsustainable technological paths such as nuclear or fossil-fueled energy technologies [3].

Research Questions

In the context of sustainability transitions, comprising both innovation and exnovation processes, our project addresses three research questions for a better understanding of dynamic innovation and exnovation policy interaction at the example of the German energy transition: 1. What are crucial points-of-decisions, which policy makers are forced to address in the course of socio-technological sustainability transitions? 2. Did the evolution of the German RE feed-in tariffs react on the dynamics within the related technological innovation systems in competition? (innovation side: new. vs. new) 3. Has the demand-sided RE support been accompanied by exnovation policies overcoming the power production from nuclear and fossil-fueled energy sources? (exnovation side: old vs. new) The proposed project is divided in a theoretical part providing a conceptual approach answer to the first research question and an empirical contribution aiming at applying the conceptual propositions to the second and third research question. Theoretical concept Our project conceptualizes socio-technological transformations as a competition between old and new technologies or entire technological systems. For the case of the German energy transition, it can be argued that entire Technological Innovation Systems [4] share a competition relation: Old, established TIS such as coal or nuclear energy technologies compete with the RE alternatives, which developed as well a systemic structure. Furthermore, the support of several innovative options for a sustainability transition suggests that the new TIS are in competition, too. Due to the inherent fundamental uncertainty about which new technological alternative(s) will succeed in the future, policy makers must govern both competition types simultaneously. The have to decide continuously how long experimentation through the support of several alternatives is justified or should be stopped in order to lock-in to one or several new technological options and phase out of the conventional paths [5]. As innovation processes for every single TIS are dynamic and can take place at different stages of the Science-Technology-Cycle [6], the governance of competing TIS must depend on the specific challenges of the TISs at the different stages of the related innovation and exnovation processes. Our conceptual propositions provide a heuristic for innovation and exnovation processes following the idea of double-boom cycles rather than s-curves, because double-boom cycles include the possible ups and downs in the development and diffusion of innovations. In particular, the exnovation process heuristic considers the different reaction patterns of actors related to the established TIS from trying to avoid the transition at all to a lobbying for a late phase-out. We identify several points-of-decision for dynamic policy making within sustainability transitions. All of them require new modes of tentative governance mechanisms for the mission-oriented policy mix.


We demonstrate the need for a dynamic decision making according to the specific dynamics of the focal technologies by answering the second and third research question of the proposed project. Using the experience of German RE feed-in tariffs for the theoretical sustainability transition debate, our methodological approach combines the central TIS indicators with a qualitative case study research methodology [7, 8]. The project follows the idea of Hoppmann et al. [9] analyzing the policy dynamics of feed-in tariffs for photovoltaic deployment in Germany, but extends the analysis to all relevant technological trajectories affected by the sustainability transition at the German power market. Main data sources for the policy dynamics of the German feed-in tariff scheme are the Renewable Energy Law from 2000 and its amendments. Further periodically published public monitoring and evaluation reports and data sets complement the analysis, especially for implemented exnovation policies. The observed time period has been defined from 2000 to 2017 because our projects focuses on the diffusion phase of the German energy transition. In order to validate the analysis and the drawn propositions, qualitative expert interviews within the policy making authorities in Germany (including the German Federal Ministry of Economic Affairs or the Federal Environmental Agency) are conducted. These are the other important empirical data source of this project.

Preliminary findings

Our preliminary findings suggest that conflicting objectives are one reason for a lack of policy coherence in the innovation and exnovation policy mix of the German energy transition. Specific objectives exist, for example, for the nuclear phaseout and the RE diffusion, but not for the phaseout of hard and brown coal power plants. Moreover, the implemented policies at several points-of-decisions did not necessarily lead to a sustainability transition, but to a stabilization of some of the dominant fossil-fueled technological paths. Regarding the new-vs.-new competition, it can be argued that feed-in tariff adjustments often directly addressed innovation process dynamics captured by our conceptual ideas on phases and points-of decisions within our heuristic on innovation and exnovation process dynamics.

References 1. Boon, W. and J. Edler, Demand, challenges, and innovation. Making sense of new trends in innovation policy. Science and Public Policy, 2018: p. scy014-scy014. 2. Edler, J., 10. The impact of policy measures to stimulate private demand for innovation, in Handbook of Innovation Policy Impact, J. Edler, et al., Editors. 2016, Edward Elgar: Cheltenham. p. 318-354. 3. David, M., Moving beyond the heuristic of creative destruction: Targeting exnovation with policy mixes for energy transitions. Energy Research & Social Science, 2017. 33: p. 138-146. 4. Hekkert, M.P., et al., Functions of innovation systems: A new approach for analysing technological change. Technological Forecasting and Social Change, 2007. 74(4): p. 413-432. 5. Dreher, C., M. Kovač, and C. Schwäbe, Competing technological innovation systems as a challenge for new mission orientation-insights from the German Energiewende. International Journal of Foresight and Innovation Policy, 2016. 11(1-3): p. 43-72. 6. Schmoch, U., Double-boom cycles and the comeback of science-push and market-pull. Research Policy, 2007. 36(7): p. 1000-1015. 7. Yin, R.K., Case Study Research: Design and Methods. 2009, Los Angeles and London: Sage Publications, Thousand Oaks. 8. Eisenhardt, K.M. and M.E. Graebner, Theory Building from Cases: Opportunities and Challenges. The Academy of Management Journal, 2007. 50(1): p. 25-32. 9. Hoppmann, J., J. Huenteler, and B. Girod, Compulsive policy-making—The evolution of the German feed-in tariff system for solar photovoltaic power. Research Policy, 2014. 43(8): p. 1422-1441.

Public Policy of S,T&I and Territory: Energy Sectorial Fund in Pernambuco/Brazil

ABSTRACT. Because they are fundamental factors in the process of capitalist accumulation, Science, Technology and Innovation have been increasingly distinguished as elements of the territory, integrating the dynamics of production and social reproduction of space. These elements determine to what extent and under what circumstances a territorial system is prone to participate in the learning economy, making it more or less attractive to capital intensive in S,T&I, which is a criterion for the territorial division of labor (SASSEN, 1998; MORGAN, 2001, SANTOS, 2008, FELDMANN, 2009). It reinforces the importance of the construction of skills and conditions necessary for the production and social and economic application of knowledge which, in undeveloping capitalist economies and with an immature innovation system, is largely produced and accumulated in public universities , emphasizing the importance of stimulating University-Company collaboration as a way to strengthen the scientific-technological connections essential to the stimulation of innovative dynamics, decisive to business and territorial competitiveness in the Learning economy (LUNDVALL, 1996). The incentive to this process of scientific-technological interaction, especially in the case of regions and backward countries where such connections are less dense, can be provided through S,T&I public policies. These, in turn, comprise the part of the State apparatuses, in which are inserted laws, programs, actions and strategies, dedicated to the induction, development, regulation and protection of Science, Technology and Innovation activities and their results (products, processes, resources and benefits) in the territory, as well as the adaptation of the territorial standards of production, appropriation, diffusion and use of S,T&I for the international criteria and demands for cooperation with foreign actors and participation in foreign markets, therefore, this policies regulate, finance and foster cooperative efforts in each Territorial Innovation System. To this end, in order to guarantee regularity in investments in science, technology and innovation in the country in the face of the threats posed by the dilapidation of the national patrimony by the privatization of state-owned enterprises, the Brazilian government created between 1999 and 2004 the Sectoral Funds of Science and Technology (CT Funds), inspired by the experience of CT-Petro, created in 1997. Based on a set of legal devices, the State has led public and/or private companies to allocate part of the revenue obtained from royalties, financial compensation, licenses and authorizations generated by the use and exploitation of natural resources owned by the Union, for the financing of R&D projects in strategic sectors, carried out in cooperation with Universities/Scientific and Technological Institutes throughout the country. These projects, from perspective of this Public Policy , represent a path to produce technological solutions of an innovative character, essential for economic growth and expansion of national competitiveness in the world market, considering the dissimmetries related to the internal market context, the Brazilian economic situation and the characteristics of each sector. In this context, the Energy Sector Fund (CT-Energ) was crieted through law 9.991/2000. This fund which establishes the obligation of electricity generation, transmission and distribution companies (GDTs) to invest 1% of their ROL in projects R&D+I with National University and Scientific-Technological Institutions, aiming to stimulate research aimed at finding new alternatives to reduce costs and increase efficiency in electric energy generation, increase the competitiveness of national industrial technology in the international context, the promotion of formation of human resources in the area and technological training. Likewise, CT-Energ's objectives include contributing to the mitigation of Brazilian regional inequalities, for that the strategy of allocating 30% of the fund's investments in less developed regions of the country was formulated. However, since, as Ramalho and Fernandes (2009) affirm, the simple regulatory framework does not guarantee the effectiveness of public policy, its evaluation is essential for the identification of the results obtained in the implementation of the strategies front of the objectives proposed by it ( ALA-HARJA and HELGASON, 2000), serving as an instrument to define the extent to which the efforts undertaken by the State are directed strategically efficient to the demands and objectives listed as priorities within the set of needs and interests at stake in the territory, allowing the development and continuous adaptation of forms and instruments of public action (Frey, 2000). This poses the challenge of understanding how such experiences, historically constructed from a hierarchical, hereditary and multiscale perspective (COSTA, 2018), materialize in each specific context, which brings with it particular potentialities and challenges that interfere with the expected results and benefits, that can only be understood from a systemic analysis. In light of the above, this article analyzes the results and benefits generated by the Energy Sectorial Fund in the context of the Pernambuco Innovation System, situated in the peripheral region of the immature Brazilian innovation system, under the perspective of the role of the Fund for strengthening the dynamics interactivity, as well as in the construction of developments that effectively contribute to reducing the regional inequalities that permeate the national territory. This object was selected by the representativeness of CT-Energ as a significant source of resources for scientific and technological production in the sector. Between 2001 and 2014, more than R$ 7 billion were invested primarily in cooperation projects between companies GDTs and Universities / Research Institutes. This appears as an essential factor for the GDTs of Electric Energy to be among the most interactive companies in the country: according to Righi (2007), in 2004, of the twenty most interactive companies in Brazil, nine belonged to the electric sector, according to Census of the Directory of Research Groups (DGP) CNPq, what the author attributes to stimulus provided by the Sectorial Energy Fund. For Costa and Fernandes (2009), this governmental incentive is also fundamental for the highlight of the electric power companies within the framework of interactions in the State of Pernambuco, especially the São Francisco Hydroelectric Company (CHESF), the fourth most interactive company in the country, which gathered 66% of the firms' relationships with Pernambuco research groups. The methodology of this research included a reconstruction of the trajectory of changes and (dis)continuities of the S,T&I policy in Brazil, encompassing the Sectoral Funds of Science and Technology, highlighting CT-Energ for the amounts collected, number of partnerships made and division management of resources. Based on the systematization of data from R&D projects supported by the scope of the Fund and interviews with leaders of research groups and representatives of companies in the electricity sector with projects funded by CT-Energ, this paper apresents the results and benefits promoted by the public policy through the interactions between the agents involved in face of the objectives of the law. As a result, it was observed that, in addition to aspects generally found in these interactions, specific determinants present in the SPIN interfere with the SUI stimulated by CT-ENERG, among which we highlight the low presence of innovative organizational and institutional factors within companies of energy in Pernambuco and the cognitive and institutional distance between the GDTs and the Universities with which they interact, suggesting that the identification of elements of the territorial systems of innovation are relevant to the improvement and to the effective attainment of the objectives of such policies.

13:15-14:45 Session 9A: Interdisciplinarity & Human Capital (SciSIP)


Location: GLC 233
Policy Extensions and Pathways: Engaging the Science of Broadening Participation and Inclusion in Science and Engineering

ABSTRACT. The science of broadening participation (SoBP) has been invoked as a means for developing a comprehensive understanding of pertinent issues for creating pathways toward greater diversity, inclusiveness, and productivity in science and engineering (S&E) disciplines and occupations. Posing questions and challenges for increasing the participation and inclusion of underrepresented groups — including women, minorities, and persons with disabilities — in S&E fields, policy objectives and analytical matters are explored relative to the delineation and determination of SoBP research and context-based approaches for pursuing related goals and considering their broader implications and impacts across academia, industry, and government sectors and society in general.

(Based on project supported by NSF SciSIP Awards 1551904 and 1551880)

Regulations that do not help: The effect of rule and regulations on scientific collaboration and productivity
PRESENTER: Heyjie Jung

ABSTRACT. This study investigates how institutional controls on the inputs to science affect collaboration structures and science production. It is motivated by two observations. First, despite wide recognition that access to resources is a key rationale for scientific collaboration (Katz and Martin, 1997; Melin, 2000; Thorsteinsdottir, 2000; Beaver, 2001; Heinze and Kuhlman 2008; Wagner, 2005, Bozeman and Corley, 2004, van Rijnsoever et al., 2008), much, if not most, science policy research considers resource inputs to be under the control of scientists (Shibayama and Baba, 2011; Wagner, 2005). Nevertheless, policies are establishing rights over resource inputs to research reducing scientists’ ability to access them. Second, significant concerns exist about how institutional controls on resource inputs may negatively affect scientific research (Jinnah and Jungcurt, 2009; Martinez and Biber-Klemm, 2010; Atlas, 2003; De Greef 2004; Grajal, 1999; Welch et al. 2013; Welch et al. 2018). However, systematic research investigating how institutional constraints affect scientists’ collaboration networks and productivity is limited.

We focus on one type of resource input, biological materials, which are regulated by multiple global, national and organizational rules that control access, exchange and use (Bretting, 2007; Sebastian and Payumo 2006; ten Kate, 2002). Policy rationales vary widely but include fairness and equity in response to biopiracy concerns, safety and security in the face of bioterrorism, health and safety, and intellectual property rights (Rosendahl, 2006; ten Kate and Laird, 1999; Thornstrém, 2012; Esquinas-Alcázar, 2005; Toledo and Burlingame, 2006). The policies respond to important stakeholder demands but also control domestic and international flows and uses of biological materials (Segger et al., 2013; Morgera et al., 2012). In this paper, we ask two main questions: 1) How do institutional controls on material inputs to science affect collaboration structures? and 2) How do the controls affect grant production?

Traditionally, material resource exchange occurs within networks that link scientists with other scientists and with provider organizations. In this new context of contested resources, access to and exchange of biological materials are jointly determined by the network structure and relationships within which researchers are embedded and these new authority structures that govern them. Scientists are likely to strategically alter their collaboration networks to improve access to and use of materials. Because it is in scientists’ self-interest to maximize value for research and minimize transaction costs of exchange related to resource inputs, they can adopt one of two strategies to access needed materials: exploit existing network ties or explore new sources of material and data (Levinthal and March, 1993). An exploitation strategy would seek to rely on members of one’s existing network to obtain materials and data. An exploration strategy would develop new network ties to individuals and organizations to ensure access to needed materials and data or to gain access to novel materials.

We expect that adoption of exploration and exploitation strategies will be influenced by the costs associated with the resources. Two types of costs are likely: social exchange costs and exchange transactional costs. Social exchange costs (SEC) include expectations of reciprocity based on the contextual norms that guide science collaboration (Shibayama et al. 2011; 2012). Exchange transactional costs (ETC) are attached to material and data because of the exogenous constraints set by resource provision policies, formal agreements or contracts, tracking and reporting requirements, monetary and non-monetary returns, no third party exchange, and limited use provisions, to name a few. Both sets of costs interact and likely help determine exploitation and exploration strategies.

Individuals explore new ties when the existing network structure and composition does not align with their resource needs and when costs of access are too high. However, new ties are uncertain. They require significant investment of time and energy to establish and they may not contribute the level or type of resources expected. By contrast, the payoff from exploration strategies may be high as new ties could generate otherwise unattainable opportunities, resources and ideas. We expect that given the benefits, exploration strategies will be associated with a higher number and dollar amount of submitted grants. However, the uncertainties associated with exploration strategies will result in lower grant success.

We test our hypotheses using a unique national survey dataset of tenured and tenure-track academic scientists in three fields – marine sciences, entomology and ecology. Of the total 3,406 scientists from US universities classified as “extensive” or “intensive” research activity (2010 Carnegie Classification) who were invited to online survey 1,353 (39.7%) responded. The analysis in this paper focuses on the subset of respondents who maintain active material exchange networks . Of the respondents, just over 43% have exchange networks in which the respondent (ego) identified individuals (alters) in their networks from whom they have obtained biological material. Biological material is defined as: “Biological material includes all biological materials that contain functional units of heredity (DNA/RNA) that are used in your field of research. The term ‘biological material’ includes plants, breeding animals and their eggs/semen, and pathogen isolates-bacteria, fungi, and viruses.” Generated alter names generated were piped into later name interpreter questions asking about the dyadic relationship in general and the material exchange relationship specifically.

Results from this study should provide significant insights into how policies on the material inputs of science affect both the structures of science networks and their outcomes.

This research was funded by the US National Science Foundation (Grant 1360166), Science of Science Policy 2014-2018) (PI, Dr. Eric Welch).

Basic Research on the Scientific Frontier: The Influence of the Disciplinary Knowledge Variety of Scientists’ Publications and their Citation Impact on Subsequent Interdisciplinary Publications
PRESENTER: Beverly Tyler

ABSTRACT. In recent years, efforts by academia, government agencies, and industry to promote interdisciplinary research have significantly increased the amount of interdisciplinary research conducted (Jacobs & Fickels 2009; Yegros-Yegros, Rafols, and D’Este 2015).  Interdisciplinary research involves integrating the research expertise of different specialized disciplines to address common problems, in contrast to multidisciplinary research where researchers from different disciplines work on problems relevant to their own discipline without integration (Birnbaum 1981). Interdisciplinary research–defined here as scientific problem solving that integrates analytical strengths of two or more scientific disciplines–has gained popularity because of the belief that many important societal issues such as healthcare, food safety, and national security require interdisciplinary research (Aboelela et al. 2007; Carr, Loucks, and Blöschl 2018; Vogel et al. 2017; Vogel and Tyler 2019).  For example, universities encourage interdisciplinary research by providing seed funding, assistance with efforts to write grants, and incorporating interdisciplinary research as an important criteria for promotion and tenure (Bloschl, et al. 2012). While interdisciplinary research is typically conducted by interdisciplinary teams, individual scientists are also evaluated for promotion and tenure based on their research productivity as reflected in the number of their interdisciplinary publications (Birnbaum 1981; Klein and Falk-Krzesinski 2017), in addition to other criteria.

Due to these developments, individual scientists in diverse disciplines and academic fields have begun conducting research to solve problems that address important social issues (Jacobs 2013), while simultaneously addressing the challenges associated with interdisciplinary research. Some scientists are motivated to conduct interdisciplinary research, because of their commitment to the social issues their research addresses. Other scientists see interdisciplinary research as an opportunity to leverage their distinctive knowledge domains to solve problems that will enhance their publication record (Klein and Falk-Krzesinski 2017). Scientists must also consider the costs associated with interdisciplinary research. For example, to be successful with interdisciplinary research scientists must bridge different world views, learn new vocabularies, and manage their research within their organizations’ political and structural context (Dougherty 1994; Kaplan, Milde, and Cowan 2017).

Interdisciplinary research in the innovation literature has primarily focused on the contributions of diverse knowledge domains to problem solving in organizational creativity and innovation (Kotha, George, and Srikanth 2013).  For example, research at the organizational level has considered how enlarging the knowledge pool creates more knowledge combination opportunities to produce new products (Katila and Ahuja 2002), how diverse and strong ties facilitate the generation of creative ideas (Sosa 2011), the double-edged sword of recombination in generating breakthrough innovation (Kaplan and Vakili 2015), and the challenges of conducting interdisciplinary research (Kaplan, Milde and Cowen 2017).  However, research has also found individuals were better able to combine diverse knowledge than teams (Taylor and Greve 2006), and individuals that accesses more diverse knowledge through distinct interactions with others from various disciplines are better able to generate creative solutions  than those with less interaction (Sosa 2011) .  Yet, little empirical interdisciplinary research has considered how individual scientists access and recombine diverse knowledge domains to solve problems.

The Knowledge Based View (KBV) recognizes the key role individuals play in knowledge generation and recombination in solving problems (Grant, 1996; Felin & Hesterly, 2007; Simon, 1991), yet so far has contributed primarily to an understanding of problem solving in hierarchies and markets (e.g., Caner, Cohen, & Pil, 2017; Jeppesen & Lakhani, 2010; Leiblein & Macher, 2009) with little to say about hybrid organizations (Jackson & Nickerson, 2004).  Hybrid organizations, such as universities and research institutions, which represent a unique context within which to study individual problem solving. Accordingly we ask: What factors contribute to individual scientists’ problem solving performance in hybrid organizations?

We integrate the interdisciplinary innovation literature with KBV principles to propose theoretical logic of how individual scientists solve problems in hybrid organizations. We develop theory to posit that scientists’ disciplinary knowledge variety is positively associated with the number of their subsequent interdisciplinary publications up to a point, but then becomes negatively related, based on an understanding of how decision rights are determined, communication established, and incentives work in hybrid organizations such as universities and other public institutions (Felin & Zenger, 2014; Jackson & Nickerson, 2004). We define scientists’ knowledge variety as the dispersion of their publications across different domains of knowledge (Yegros-Yegros, Rafols & D’Este, 2015). Furthermore, we propose that the opportunity for scientists that work in hybrid organizations to participate in interdisciplinary research is also likely to be related to scientists’ scientific reputation, because scientists with different domain knowledge will be incentivized to work with scientists that have strong reputations. We define scientific reputation as the impact of scientist’s publications on future research as represented by citations (Hall, Jaffe and Trajtenberg, 2005; Hamermesh 2018;). Therefore, we posit that scientists with stronger scientific reputations are more likely to solve interdisciplinary problems leading to publications.

We test these hypotheses in the context of eight US nanomedicine development centers (NDCs) funded by the US national Institute of Health (NIH).  The NDCs provide an appropriate context to access scientists participating in interdisciplinary research (Blöschl et al., 2012; Franks et al., 2007; Viterbo, 2007). We analyzed a dataset containing 2,393 publications and 21,666 citations for 169 researchers participating in the eight NDCs using random effects negative binomial model with the conditional likelihood estimator. The empirical analysis and results provide support for our hypotheses. 

Our study contributes to interdisciplinary research and knowledge-based theory literature in two ways.  First, we contribute to interdisciplinary innovation research on the value of knowledge variety for individual scientists’ interdisciplinary research performance. We find that a moderate level of disciplinary knowledge variety provides the highest contribution to scientists’ subsequent interdisciplinary publication performance. Second, we inform the interdisciplinary literature on individual scientists by showing that scientists the stronger scientists’ reputations for publishing the more likely they will be asked to participate in interdisciplinary research, confirming prior research on the value of reputation for scientists. Third, we contribute to the KBV by extending the problem solving perspective to include a consideration of individual problem solving in hybrid organizations.

13:15-14:45 Session 9B: Gender & Productivity


Location: GLC 235
Female-authored papers are more often in high impact journals but less cited
PRESENTER: Ulf Sandström

ABSTRACT. The problem

Differences between women and men in science are said to be gradually narrowing. However, the latter gender is still ranked higher and has a better reputation in the scientific community. Xie & Shaumann (1998) indicated a process of narrowing, but Van den Besselaar & Sandström (2018) showed that this is not the case, also 20 years later the overall situation is still that women researchers perform two-thirds of what is expected from male colleagues. Actually, Cole & Zuckerman (1984) introduced data from the U.S. which gave the same relation between genders ever since the 1930s.

In this paper we will dig a bit deeper into the differences of today (2008-2011) and will answer the following questions: - Are papers with (predominantly) female authors cited less than papers with (predominantly) male authors? - Are papers with (predominantly) female authors published in lower impact journals than papers with (predominantly) male authors?

The method

The work starts here: a name disambiguated file with all publication with a Swedish address over the years 2008 2011, consisting of some 74,000 articles (Articles, Letters, Proceeding Papers and Reviews) from the Web of Science databases (SCI-E, SSCI, A&HCI). We concentrate on Swedish articles with Swedish authors only as these are the article shares that are disambiguated and gender has been attached to these names. There are 14 760 Swedish articles with Swedish author addresses only.

Papers are classified into “gender” categories: 1, 2, 3, 4 and 5 female authors; 1, 2, 3, 4 or 5 male authors; and various combinations such as 1M(an)1W(oman), 1M2W, 2M1W. We then compare the citation impact and the average journal impact factor between those groups, to find out whether a gendered publication and citation patterns emerge.

In this abstract, we only use papers with Swedish authored papers only, but for the next version, we will use for the international co-authors in order to increase the sample considerably.

In order to measure citation impact, we have constructed reference values for publications by field. Based on these, we can calculate for each paper the citation impact. That is done using the percentile model. For each percentile category we give points to a paper: top1% (field and year normalized) highest cited papers = 100 points; top5% = 20; top10% = 10, top25% = 4; top50% = 2; lowest 50% = 1). This seems big differences between the percentile groups, but the probability of having a top1% paper is ten times lower than a top10% paper, and that is reflected in the scores. Measures based on percentiles have the advantage of not being affected by causes of bias in citation distributions (Rousseau 2005). In certain disciplinary areas, a few publications with very numerous citations otherwise boost the mean, which can result in 70% of the articles in the area being below this mean (c.f. Campbell 2017 [STI 2017 paper], c.f. Thelwall 2019).

Field delineation is maybe the most important issue here. After some experimentation with various approaches, we found that the suggested macro fields in the Science Metrix classification would fulfil the requirements that were needed, e.g. to distinguish a category of applied science fields. So, we had in the final round five different clusters (fields); humanities, social science and economics, applied sciences, health sciences, and natural sciences.

The Results

The about 15,000 papers with Swedish address and Swedish authors consist of about 32,000 article shares. The mean is a bit more than 2 authors per paper. Of these article shares, 7,400 are from papers written by a mixed gender team (1W&2M or 2W&1M). Almost 5,000 (4,784 papers) are written by women only in different constellations (1W, 2W, 3W, 4W & 5W) and almost 20,000 (19,528) were authored by male persons only. This indicates that the selections of papers are to some extent similar to the distribution of papers over gender. In the overall data of Swedish researchers, there are about 40% women and 60% men for the total author shares (with a Swedish address).

If we now order the author constellations in terms of the average number of points per paper, we get the following list from the highest number of points per paper (3.8) to the lowest number of points per paper (2.2): 5M, 4M, 5W, 3M, 2M, 1W, 1W&2M, 1M, 2W&1M, 2W, 3W, 4W. This clearly shows (with a few exceptions) that the more men, the higher the number of average points per paper, and the more women, the lower the average points per paper.

And what when we look at the average quartile score of the papers in each of the gender author constellation groups? We score every paper in terms of the Journal Impact Factor quartile the journal where the paper is published in belongs to. Quartile 1 consists of the 25% journals with the highest JIF. Ordering the author constellations from low values which is based on mean quartile per paper than the JIF average for the female-dominated constellations. We get the following order from lower to higher: 1M, 1W&2M, 5M, 2M, 4M, 1W&2M, 3M, 4W, 2W, 2W&1M, 3W, 5W. This suggests that female-authored papers tend to end up in high impact journals. Consequently, we find that the publication venues are not the explanation to the pattern found regarding actual citations.

Conclusions and discussion

This leads to the conclusion that there seems a clear gendered order in that ‘male papers’ are more cited although they appear on average in lower impact journals than ‘female papers’. Why is this the case? Is it the effect of gender bias in journal review processes, leading to fewer review-based barriers, a sort of positive discrimination for ‘female papers’ in high impact journals? Or is there gender bias in citations implying that ‘female papers’ are taken less seriously? Or do women send papers to higher ranked journals resulting in fewer citations? To sort out the explanations is an issue for further research. At least two empirical issues need to be solved before the conference: First of all, we will attribute gender to the non-Swedish co-authors, as we then have a much bigger sample. Above that, it is well known that internationally co-authored papers are better cited and more often in ‘better’ journals. So including these papers may change the picture. Secondly, the data cover publications over a four-year period (2008-2012). In order to find out whether the patterns we found are referring to real selection phenomena, we will redo the analysis with a similar cleaned set of papers, covering Sweden between 2012 and 2015.

Field differences in gender bias: the ERC starting grants case

ABSTRACT. Introduction

There is a longstanding discussion on whether gender bias influences grant selection processes, and the literature shows contradicting results. However, there are three main problems with most research: (i) Most studies explain in fact only differences between success rates of men and women. However, these success rates are only meaningful when taking possible quality differences of male and female researchers into account. If these exist, gendered differences in success rate could partly or fully an effect of those quality differences and not of gender bias. To solve this problem, we have collected data to measure various dimensions of past performance, which are included in the analysis. (ii) Most studies depend on information only about the successful applicants, but not on the rejected – as the latter data are generally accessible for investigators. However, in this study we do have the data about successful and rejected applications. (iii) Bias emerges from the decision-making process, and this is often done at the level of review panels. In contrast, most studies focus on a higher level of aggregation, such as the funding instrument, or at the level of the discipline. We include here an initial analysis at panel level. We do detect gender bias, in contrast to recent reviews.

We answer the following questions: 1. Does gender influence the panel score and application success? 2. Do we find field differences in gender bias? 3. Do panel characteristics influence the occurrence of bias?

In an earlier paper (presented at STI 2018 in Leiden), we presented the results for the life sciences. In this paper, we include the other disciplines within Physics and Engineering and in the Social Sciences and Humanities.

Approach, data & methods

We aim to predict the applicants scores and application success, using a set of independent variables related to performance (productivity, impact, previous grants, quality of the collaboration network) and to the person (age, nationality, research field, and of course gender). As decision-making on grants is done in panels, the effect of the panel is considered too – using multi-level approach. The following data were collected: - Age, gender, date of PhD, nationality, field of research - Earlier and current other grants. - Collaboration network - Median of the position (in the Leiden ranking) of the organizations in the network. - Position of the Host institution in the Leiden Ranking - Productivity (fractionalized) - Impact - Organizational proximity (cronyism) - Cognitive proximity - Panel review scores of the applications - Decision

We investigate the 2014 ERC Starting Grant scheme, and have access to the relevant data about the 3,030 applicants (about 95 %) that gave informed consent. We selected this case, as it is the most prestigious grant that exist in Europe for early career researchers (up to seven years after the PhD), and it is expected to strongly contribute to career opportunities of those getting the grant. As the whole process is weakly formalized and codified, one may expect bias entering the selection process.

We collected the bibliometric data for all applicants, which is the most labor-intensive task in the project. The unique nature of our data is that we can combine (advanced) bibliometric indicators with a large set of other variables.

First findings

We used the bibliometric indicators mentioned above, the variables on the quality of the network and the host institution, and the number of grants the applicant has already acquired. We also include whether a panel member is at the host institution of the applicant, and gender. In this abstract, we present some preliminary results.

For the Life Sciences we find the following: - Running an ordinal regression, and after manually stepwise deleting variables that did not work, eight variables remained in the model, which resulted in a pseudo R-square (Nagelkerke) of 0.308. Table 2 shows the result. - Factors that help to get a better score are papers in high impact journals, the quality of the network, measured as the median ranking of the organizations in the network of the applicant, average number of international coauthors, and the number top 10% most cited papers (fractionally counted). Negative works the average number of coauthors, as that may suggest a lower level of independence. Finally, we do find effects of sexism and nepotism: women score some 0.35 points lower than male (on a five-points scale), and when the candidate has a panel with a panel member that is at the proposed host institution, this gives almost a 0.6 point bonus. - This means that from a performance perspective, only one variable plays a role (the number of top cited papers). The other variables that influence the score are reputation based (journal impact related; ranking related) and network based (number of (international) co-authors). Also, the number of earlier grants has a positive effect on the score; and these grants partly can be considered as performance, but at least also partly as reputation-related. Finally we find two bias factors: after controlling for the performance and reputation variables, sexism and cronyism still have an effect on the scores the applicants get. - These results are on the domain level, but within the Life Sciences, 9 different disciplines are distinguished. And the gender bias patterns are not the same between the LS-disciplines. Gender bias in favor of men is in six of the nine panels, covering 78 % of the female applicants in the life sciences. The other three have bias in favor of women, with 22% of the female applicants.

For the Physics and Engineering (PE) domain, we find exactly the opposite (based on a sample of some 65% of the applicants). Although there are a few panels with gender bias against female applicants, the overall picture within PE is gender bias in favor of female applicants.

Within the Social Sciences and Humanities (SSH), we find no gender bias at the SSH domain level, but this is the average of panels with gender bias in favor of women and panels with a bias against women. Also here the analysis uses a sample of 65% of the applicants at the moment.

Currently we are redoing the analyses at the domain level, and we will do a multilevel analysis that includes the panels as a second level.

Conclusions and further work

Our first analysis shows that gender bias occurs in the life sciences, but not in all parts of the field in the same way. In most panels we find bias against women, but in three panels it is the opposite. The first analysis suggests that in physics and engineering the situation is the opposite, and the the social sciences and humanities it is mixed. In the paper we will elaborate on these differences between the domains. What social and intellectual differences explain the differences in gender bias?

Panels play an important role, therefore we will also include characteristics of the panel in the model. What panel characteristics do lead to gender bias? For example, we found a negative correlation between the number of female panel members and the female success rate. A multilevel analysis should bring new insights.

How inclusive is AI? A bibliometric study on gendered scientific and technological development

ABSTRACT. Introduction

The growing hype around artificial intelligence (AI) and its potential societal benefits have burgeoned a significant focus of government spending worldwide. In 2016, the AI market accounted for $4,065.0 million worldwide, and the forecast for 2025 is $169,411.8 million (Allied Market Research 2018). AI holds a huge promise to alleviate global concerns over global warming and healthcare access, providing advances in digital technologies (King and Roberts 2018; Nolan 2018). This notwithstanding, many scientists and policymakers have raised concerns about the ethical, societal and legal aspects of AI due to its fast-paced advancements, spanning from human-machine interactions and autonomous cars (Elsevier 2018). However, these ethical concerns mostly align with inequity from the perspective of the distribution of benefits from the use of AI and do not address inequity concerns related to the capacity for AI development and commercialization. Along these lines, the concerns over “inclusive AI” also revolves around the former rather than both aspects, which sheds light accounting for diversity (of gender, age, race, origin, etc.) in algorithms, datasets, and other learning mechanisms. This research tries to fill this void and address gender inequity concerns from the latter perspective: the scientific and technological development of AI.


This study tries to address gender-related concerns around the scientific and technological development of artificial intelligence. It also provides a large scale, cross-country and disciplinary homogeneous gendered analysis of the scientific and technological productivity and impact of researchers involved in the development of the artificial intelligence technology. More specifically, this study maps to what extent women are involved in AI-related scientific and inventive activities by years, disciplines, high impact research, country and sector of affiliation, and collaboration types. Moreover, this study analyzes gender differences in the citation and journal impact AI-related articles when women are assigned to leading authorship positions.


Article data is extracted from the Web of Science (WoS) database using the AI-related keyword dictionary (consulted with AI experts from the Institute for Data Valorisation (IVADO)) and patent data is extracted from the United States Patent and Trademark Office (USPTO) using the International Patent Classifications concordance for US class 708 (Data Processing: Artificial Intelligence). Field classification for articles is based on journals discipline by the U.S. National Science Foundation’s (NSF) Science and Engineering (S&E) classification scheme, which assigns each journal to only one discipline. Gender will be further assigned to authors and inventors using the universal and/or their affiliated country name and gender lists, including U.S. Census, WikiName, Wikipedia, African lists, France and Quebec lists and other country-specific lists (more details can be found in (Larivière et al. 2013)). Author’s affiliations and patent’s assignees will be further categorized into academic, governmental, and industrial sectors using the keywords introduced in (Ghiasi et al. 2015). Since the full name of authors is only available starting from 2008 in WoS (which is essential to assign gender), we analyzed AI-related publications from 2008 onward.

This research relies on bibliometric indicators of scientific (and technological) output and impact and uses the number of scientific publications (and patents), normalized citations, and normalized journal Impact Factor (IF) to evaluate, respectively, scientific and technological productivity, scientific (technological) and journal impact. The proportion of the scientific (or technological) output of authors of each gender is defined as a fractional count of papers (or patents), according to which each author (or inventor) is given a 1/x count of authorships (or inventorships) where x account for the number of co-authors (or co-inventors) for which gender is assigned on given paper (or patent). This research defines authorship (inventorship) as these gendered fractions while being aggregated at country, discipline, and sector level. The citation impact of a paper (or patent) is measured as the average yearly number of citations received by a paper (patent) divided by the average yearly number of citations to all the papers from the same year, in the same discipline (or the same IPC) and of the same document type. The normalized journal IF is defined similarly, considering the IF of the journal in which a paper is published.


In this study we identified 69,675 total AI-related articles (from 2008-2017) and 83,520 AI-related patents (from 1976-2016). Our findings reveal that women account for 23% of AI-related authorships. This share has constantly been increasing over the past few years (from 21% in 2008 to 25% in 2017) and is more pronounced in the field of Electrical engineering and electronics, applied mathematics, and neurology and neurosurgery. When women are in leading positions, either the first author or corresponding author, their publications are published in lower impact journals and receive lower citation rates. However, gender differences in journal ranks are less pronounced than those in citation rates. Most interestingly, women first-authored papers receive a higher rate of citations and are published in higher ranked journals than men first-authored papers, when the corresponding author (a proxy for supervisor of the research project) is a woman and vice versa. Our analysis of patents reveals that although there is a substantial increase in the number of patents, female inventorship is increased insignificantly (3% increase in 15 years). A higher share of women is involved in patents assigned to universities.

Our results show that women are at a disadvantage when conventional scientific indicators (publication, patent and citation measures) are applied for evaluative purposes in the field of artificial intelligence. Although policy efforts around inclusive AI have become more significant in the science and technology discourse, initiatives to involve and retain women in AI R&D have not been effective. This study could serve as a baseline from which to strengthen gender mainstreaming in AI-related scientific and inventive activities.


Allied Market Research. (2018). Artificial Intelligence (AI) Market to Garner $169,411.8 Million, Globally, by 2025. Allied Market Research. Accessed 15 March 2019

Elsevier. (2018). ArtificiaI Intelligence: How knowledge is created, transferred, and used. Accessed 14 March 2019

Ghiasi, G., Larivière, V., & Sugimoto, C. R. (2015). On the Compliance of Women Engineers with a Gendered Scientific System. PloS one, 10(12), e0145931.

King, R. D., & Roberts, S. (2018). Artificial intelligence and machine learning in science. In OECD Science, Technology and Innovation Outlook 2018 (pp. 121–136). OECD. doi:10.1787/sti_in_outlook-2018-10-en

Larivière, V., Ni, C., Gingras, Y., Cronin, B., & Sugimoto, C. R. (2013). Bibliometrics: Global gender disparities in science. Nature News, 504(7479), 211. doi:10.1038/504211a

Nolan, A. (2018). Artificial intelligence and the technologies of the Next Production Revolution. In OECD Science, Technology and Innovation Outlook 2018: Adapting to Technological and Societal Disruption. OECD. doi:10.1787/sti_in_outlook-2018-en

Scientific production and impact in Africa – Does collaboration mitigate the gender bias?

ABSTRACT. Publishing, as a universal tool of measuring scientific production, remains a yardstick for academic promotion, even in academic contexts that do not appear to support research, as is often the case in Africa. African research currently accounts for a fraction – less than 3% – of scientific publications, much smaller than is desirable if the potential contribution of Africa’s researchers is to be realised for the benefit of its populations. At the same time, one of the most serious gaps that African universities need to close if African countries are to fully utilise their human potential is the gender gap in research participation. Globally, women account for a minority (28·8%) of the world’s researchers and approximately 24·0% of researchers in African countries. More importantly, and the focus of this paper, is the consistent finding globally that the scientific production of this minority of female researchers is lower than that for men. However, research on scientific production of researchers in Africa, especially in gender-disaggregated form, are scarce, partly because, until recently, women scientists were “so rare in Africa as to be nearly invisible.” A review of the literature shows that even after half a century of empirical research on gender differences in scientific production conducted in developed countries, no single explanation or group of explanations satisfactorily accounts for the phenomenon, aptly referred to by Cole and Zuckerman as “the productivity puzzle”. In our research, we compare the scientific production of men and women, and then, by controlling for a variety of publication- and gender-related variables, provide a more refined portrait of the production of women in relation to that of their male peers. As such, we respond to what has been referred to as “an urgent need for more gender-disaggregated data, and more refined statistical analysis”, to facilitate the formulation and implementation of effective institutional change strategies concerning gender equality in research participation and production. To provide a more refined portrait of the scientific production of women in relation to that of their male peers, we conducted a web-based survey in 2016 of individuals across all African countries that had co-authored at least one scientific article in a journal indexed by the Web of Science (WoS) in the preceding ten years. In total, 7,515 individuals responded, which constitutes a response rate of approximately 10%, once multiple email addresses of the same individual were excluded. When we removed individuals that are not working in Africa, and those that had not completed the questions relevant for this study, we were left with 5,032 valid and complete cases, of which, 3,263 are in Science, Technology, Engineering and Mathematics –STEM– (68·3%) or health scientists (31·7%). The respondents’ number of publications, mean normalised citation scores, number of co-authors were extracted from the WoS data for the period 2014-2016. In addition, we characterised the co-publication network of these individuals, using the data extracted to build the sampling frame – i.e. all the individuals with an African affiliation within WoS, and their direct co-authors. Because of the very large number of individuals in the database, we built networks per domain (9 categories in STEM and health) and by field (2 fields – STEM and health). The resulting 3,263 valid questionnaires from STEM respondents were analysed using multivariate regressions on number of articles published in the preceding three years, and on their mean normalised citation score. Controlling for a variety of publication- and gender-related variables, we identified the factors that most affect scientific production, focusing on the moderating effect of gender. Our results show that, while women’s scientific production is positively affected by age and collaboration, it is negatively impacted by number of children, care-work and household chores, as well as by limited mobility. When women devote the same number of hours to academic tasks and raise the same amount of research funding as their male colleagues do, they are as prolific. More importantly, the influence of children is temporary, as women “catch up” later on in their careers. Contrary to these unfavourable indicators, we find that the difference between men and women in regard to the amount of funds at their disposal is not significant. Hence our results do not give credence to the hypothesis that African women are less productive because of a lack of funding or because of the amount of time they devote to raising research funds; in fact, the more time women spend fundraising, the more productive they become, compared to their male colleagues. For all other academic tasks (teaching, supervision, research, administration), when men and women work an equal number of hours, the impact on scientific production is the same. While increased frequency of collaboration (with colleagues ranging from those in one’s own institution to international colleagues) contributes to increasing scientific production, both men and women perform equally given the same opportunities. In addition to self-reported collaboration, the social network indicators, degree centrality and betweenness centrality both reinforce the importance of collaboration. These do not however entirely mitigate the gender bias identified. In contrast, mobility does not appear to influence the number of papers published until we discriminate by gender (16): women who have not been mobile in the preceding three years are significantly less productive than both their male colleagues (whether they have been mobile or not) and their female mobile colleagues. The standard academic career model, relying on continuous publications and regular promotions, assumes a masculine life cycle that reinforces the general perception that women with discontinuous careers are less capable than their male colleagues are, and systematically disadvantages women. Our results lead us to argue that the solution resides beyond women’s empowerment, i.e. in the broader institutions of education and the family, which have an important role to play in fostering men’s equal contribution to household chores and care-work.

13:15-14:45 Session 9D: Evaluation in Practice
Location: GLC 225
Government funding and international collaboration in scientific research

ABSTRACT. Government funding plays a significant role in the development of science, and have attracted extensive interests of academic community. Scientific discovery increasingly relies on wide-spread collaboration including international collaboration. Positive effect of either government funding or international collaboration has been proved in many studies. What might happen when both government funding and international collaboration are considered? This puzzle will be cleared up in the following questions: (1) What is the contribution of funded research to science in different countries? (2) What is the effect of government funding support on citation impact of different countries? (3) What is the role of international collaboration in different countries? Does authorship affect collaboration? (4) Is funding support bound to add up the effect of international collaboration?

Data are extracted from publications indexed in the CWTS-licensed version of the Web of Science database (WoS) of Clarivate. To ensure representativeness of our study, we choose China, Brazil, South Africa, Germany, The Netherlands, and the US so as to represent publication activities of countries from Asia, South America, Europe and North America, or developing and developed countries. National funding agencies supporting basic research of the target countries including the National Natural Science Foundation of China (NSFC), the National Council for Scientific and Technological Development of Brazil (CNPq), National Research Foundation of South Africa (NRF), German Research Foundation (DFG), The Netherlands Organization for Scientific Research (NWO), and US National Science Foundation (NSF). These organizations are named as focal agencies in the later text. Funded publications are harvested from CWTS funding organization database originated from WoS index fields FO and FT. To clarify if funding support from the focal and non-focal organizations generate different effect on citation impact, we further classify publications of a nation into two categories – publications funded by the focal agencies or non-focal agencies.

Although highly productive, the area Clinical Medicine is not covered in the current analysis because it is not major area supported by the National Science Foundation (NSF). Thus, six subject categories including Basic Life Sciences, Biological Sciences, Chemistry & Chemical Engineering, Environmental Sciences & Technology, Mathematics as well as Physics & Materials Science will be our focal areas.

To measure the effectiveness of funding or international collaboration, we use indicators measuring publication output including percentage of funded publications, top-1% highly cited publications. CWTS indicator Mean Normalized Citation Scores (MNCS) is used to measure citation impact. Citation window is variable.

In view that effect of collaboration with developed countries in science might be different from that of collaborating with developing countries, we will analyze the effect of collaboration with developed and developing countries in science. Countries of Group 7 (i.e., Canada, France, Germany, Italy, Japan, the United Kingdom, and the United States) are defined as developed countries, the rest as non-G7 countries. Collaboration with US will be analyzed individually because of its absolute leading position in science.

The importance of funding support in scientific research has been proved in the current study. With absolute high percentage of publications supported by the focal agencies, China and Brazil have more centralized government resources supporting basic research than the rest four countries. The NSFC of China supports basic research in almost all areas, while in the US, the responsibility for supporting basic research is spread across the National Science Foundation (NSF), National Institute of Health (NIH), and Department of Energy (DOE). The percentage share of funded publications in all publications of a country has kept growing, and each country shows its own way of growth. China has the highest percentage of funded publications, and the percentage of NSFC in the funded total has kept growing, which is in stark contrast to the focal agencies in the other five countries, especially DFG and NWO with a decreasing share. In other words, Chinese researchers rely heavily on NSFC funding, whereas US, Germany, Dutch and South African scholars have more options in seeking funding supports.

With regard to citation impact and top-1% highly cited publications, developed countries (i.e., The Netherlands, US and Germany) perform better than developing countries (i.e., South Africa, China and Brazil), and Netherlands performs best. Funding support plays an important role in raising citation impact, although country variation exists. In supporting high-quality research in terms of citation impact, some focal agencies like NSF are more efficient than non-NSF funding support, some (i.e., NWO and NSFC) do not show much difference, and some (i.e., DFG, NRF, and CNPq) even are less efficient.

The positive effect of funding support in raising citation impact of publications is further confirmed in OLS regression. Citation-raising effect of most of the focal agencies (i.e., NSFC, NSF, NWO and CNPq) are higher than that of non-focal agencies. Nonetheless, it is not true in German and South African situation. The situation in South Africa is most unique – there is no significant change of citation impact of publications funded by NRF whereas non-NRF funding can raise citation impact.

Statistically, not all countries get benefit from international collaboration in terms of raising citation impact. The developing countries benefit more than the developed. Collaborating with developed countries especially the US can be a first option for choosing foreign partners. Although citation impact of the developed countries might be lowered in collaborating with developing countries, such collaboration still has values in complementing shortage of human resources and helping young scholars from developing countries grow in their academic career. Authorship in international collaboration matters significantly. Citation impact can be raised if collaboration is led by developed countries, and the opposite is true if the developing countries lead a collaboration.

Situation would become more complex when both funding support and international collaboration are considered. The effect of focal-agency support is different from that of non-focal-agency support. Focal-agency support has significant impact in four countries – positive impact in the US and Germany and negative impact in China and Brazil, whereas non-focal-agency support only affects two countries – positive in Brazil and negative in Germany.

A Value Assessment Engine for International Space Station Research and Development

ABSTRACT. The International Space Station (ISS) is an orbiting platform operated jointly by the U.S. National Aeronautics and Space Administration (NASA), European Space Agency (ESA), Japan Aerospace and Space Agency (JAXA), and Russian Space Agency. Among U.S. government research user facilities, the ISS is a unique asset. By providing experimentation in an environment where gravity is no longer a predominant force, the ISS can enable research investigations that cannot be performed in any other location. These investigations span the full range of R&D activities, from purely curiosity-driven basic research to applied research to experimental development. In addition, once the human crew is included as part of the ISS infrastructure, the facility accommodates studies across many disciplines, including physical sciences, biological sciences, biomedical research, and social science. This tremendous diversity—encompassing the studies conducted aboard the ISS and technologies developed to support the operation of the ISS itself—means that any measure of the value of the outputs of the ISS must also accommodate tremendous diversity.

In 2017, the ISS Program Office at NASA Johnson Space Center commissioned RTI International to develop a system for identifying, tracking, and depicting the value of the results of research and development activities (R&D) conducted aboard or for the ISS. This paper summarizes our initial work to develop this system. Our approach first developed methods for gathering the evidence of the value contributed by the ISS. Starting with data from the ISS program on the research experiments conducted aboard the ISS, we collected and analyzed three types of R&D outputs: scholarly articles, patents, and research datasets. Second, we reviewed existing literature on methods for generating measures of the value of each type of output. We then designed and produced different indicators of value, taking the three forms of evidence collected and visualizing summary information about that evidence base in ways that could address the interests of the ISS program management.

Assembling the evidence base. The documents published by researchers about their work related to the ISS do not necessarily cite the ISS explicitly. Since the ISS Program provides a research platform but not funding, there is no policy requiring specific acknowledgement of the ISS Program or NASA in those documents. We constructed a highly-targeted Boolean keyword search to find both publications and patents likely to be attributable to ISS-based research. In the process, we developed a classification system describing attribution of results to ISS research as a continuum. We also note current limitations on the ability to link the use and re-use of datasets to the ISS.

Specifying research value. The value of research should be framed in terms that go beyond financial valuation. We examined the different pathways by which ISS research generates value for different communities of stakeholders. In the process, we derived important observations about the nonlinear relationship between the research and technology development and the value derived from related activities.

Designing and generating visual indicators for research value. We developed a series of discrete interactive graphical representations of different aspects of value, derived from the metadata on ISS-hosted experiments and research outputs. We followed a systematic process for determining how such indicators could help ISS program management understand the nature of the portfolio of R&D activities associated with the ISS. The indicators we generated assist decision-making by ISS management while also informing stakeholders on different modes by which communities gain value from ISS research.

In this paper, we explain (1) the different aspects of “value” generated by the ISS, (2) the methods developed to identify and analyze these three types of R&D output, and (3) novel indicators developed to provide a summary representation of different perspectives on the scientific and technical value of ISS R&D activities.

Evaluating an Innovative Basic Research Program: The Case of the Emergent Phenomena in Quantum Systems Initiative

ABSTRACT. Philanthropic foundations contribute substantial funding for fundamental scientific research at universities and research institutions in the U.S., although the exact scale of that contribution is uncertain. (Kastner, 2018; Committee on National Statistics, 2016). Data show that the U.S. federal government provides the overwhelming share of funding for university-based basic research. How can philanthropies generate a disproportionate impact from a smaller portion of the funding stream? In particular, could we expect that a fundamental research program with a novel design and operating model might be able to accelerate science and generate transformative progress in particular fields?

In 2013, the Gordon and Betty Moore Foundation (the foundation) launched a funding initiative on Emergent Phenomena in Quantum Systems (EPiQS). The EPiQS Initiative provided $90 million in funding through 2019 to support research and related activities in the field of quantum materials. The central motivation for EPiQS is encapsulated in its “Theory of Change,” expressing how the initiative intends to make an impact on the field:

“By providing top scientists with access to the best materials and experimental probes, the resources and freedom to innovate, and opportunities to collaborate, EPiQS will catalyze transformative change in the science of emergent phenomena in quantum materials.”

In 2017, RTI International was commissioned by the foundation to conduct an external evaluation of the EPiQS initiative. The purpose of RTI’s evaluation was to:

1. assess whether the strategies pursued by the EPiQS initiative were appropriate for achieving the desired outcomes and implemented effectively; 2. identify the effects thus far of the EPiQS initiative’s strategies on the quantum materials field of research; and 3. generate observations on the initiative and its consequences, and how that might inform future investments in this field (or in comparable fields of research).

To do this, RTI engaged in three different streams of analysis:

1. Our qualitative analysis used structured interviews and document review to construct a retrospective narrative and analysis of how the EPiQS initiative developed and how its activities progressed over the past six years, and synthesized the opinions and recollections of grantees and other key stakeholders. 2. Our quantitative analysis involved the deployment of a web-based survey to grantees and postdoctoral researchers supported by EPiQS. We gathered information about these individuals’ activities and attitudes related to the initiative, providing trends and patterns in how EPiQS affected their research activities and accomplishments. 3. Our scientometric analysis, conducted primarily by our research partner Observatoire des Sciènces et Technologies, used both state-of-the-art and exploratory methods of bibliometric and text analysis to make inferences about changes in research topics, publication behaviors, and community development in the quantum systems field.

Separately, the foundation engaged an international expert panel of leading scientists in quantum materials. That panel evaluated the quality and impact of the scientific research undertaken by PIs and to identify potential future areas for scientific investment. The results of the expert review were also incorporated into RTI’s evaluation.

In this paper we review the methods used in the evaluation and how those led to key findings provided to the foundation. In particular, we found that:

• EPiQS aided in establishing “quantum materials” as a distinct field of science. • EPiQS grew the pipeline of researchers entering the field of quantum materials research. • EPiQS raised the standing and significance of materials synthesis as a key part of the U.S. research community. • EPiQS contributed to decisions to hire new faculty in materials synthesis. • Other research sponsors are interested in supporting quantum materials research, especially in tandem with EPiQS.

Our paper deals with the challenges of conducting a formative evaluation of a funding initiative where the scientific impact is likely to take decades to be realized fully. We also discuss what the evaluation revealed about the benefits of the novel design elements incorporated into EPiQS, and what we might learn from how philanthropic funders can work synergistically with government funders to enable the emergence of new fields and to produce potentially transformative research results.

Professional standards in bibliometric research evaluation? A meta-evaluation of European assessment practice
PRESENTER: Arlette Jappe

ABSTRACT. Professional standards in bibliometric research evaluation? A meta-evaluation of European assessment practice

Background and rationale

Research organizations and research funding agencies have a growing demand for practicable methods of research evaluation, including metrics based on publication and citation data. Such bibliometric indicators remain controversial among the scientific communities affected by performance assessment, and there are presently no widely accepted methodological standards. In recent years, several studies have reviewed scientific developments in the area of evaluative citation analyses. However, there is little overview regarding which methods are actually used in bibliometric assessment practice. This gap is addressed in the present paper.

In the literature, the expression ‘meta-evaluation’ is commonly used to denote systematic reviews of evaluation studies with regards to methodological quality and results. In the present study, we analyse the methodologies of existing evaluation studies from a meta-perspective. However, rather than evaluating published studies according to predefined methodological standards of good practice, our purpose here is investigate two main research questions. First, what were the prevailing methodological standards, referred to as professional de facto standards, in the field of research assessment practice during a certain period? Second, if certain de facto standards of bibliometric research assessment can be identified, which social actors have defined them?

This paper is part of a project conducted with the aim of understanding the development of bibliometric assessment methods from the perspective of Andrew Abbott’s sociological theory of professions. We selected this theory to investigate how particular methodological choices become socially established as professionally legitimate means of handling certain evaluation problems. More specifically, this framework is used to address the issue of professional control in bibliometric assessment. Applying Abbott’s terminology, the increasing demand for practicable and efficient assessment of academic performance constitutes a problem amenable to expert service. Research assessment is potentially within the jurisdiction of professional experts who can define the nature of assessment problems, and offer solutions that effectively address clients’ needs. A prior paper by the same authors investigated reputational control in the academic sector of evaluative citation analysis.

Our initial assumption was that leading organizations within the expert field would be influential in defining professional de facto standards—first, because they have a high market share of assessment services and, second, because they serve as a legitimate role model that is imitated by other bibliometric experts. Our findings support this assumption but also highlight the importance of data access and data distribution for the operative establishment of de facto standards.

Methods and Data

The focus of this study is on the measurement of citation impact. We excluded other topics of bibliometric assessment, such as emerging research topics, research profiles, and international collaboration. Moreover, this investigation only includes ‘real’ evaluations, i.e. assessments conducted for purposes of decision-making in research policy or research management. We applied several complementary search strategies, and identified 82 individual studies published during the period 2005–2014, which evaluated either research organizations (RO) or research funding instruments (FI) from 14 European countries plus EU framework programs. We analysed the bibliometric design of each individual study using a scheme of 37 coding questions. Most items involved a nominal level of measurement, i.e. non-ordered qualitative characteristics, with some open items for more detailed descriptions.


1. Bibliometric research assessment is most frequently used in the Netherlands, the Nordic countries, and Italy.

We found instances of bibliometric evaluation in many European countries, but the most regular use of bibliometric assessments during the observation period was concentrated in a few countries. Approximately 60 % of the studies come frome the Netherlands, Italy, Sweden, Norway, Finland, and Denmark.

2. The Web of Science (WoS) is the dominant database for the assessment of public research in Europe.

The bibliometric evaluation of public research in Europe during the observation period was largely based on the citation indices contained in the WoS. Of the total sample set, 90% of studies relied on WoS, while only 15% used Scopus or a combination of WoS and Scopus. Some studies employed designated databases, such as PubMed or MathSCInet, but these alternative citation databases exist for only a few disciplines.

3. Expert organizations invest in cleaning and improvement of WoS citation data and are able to set technical standards with respect to data quality.

Raw citation data as provided by WoS or Scopus require considerable processing before they are adequate for the assessment of authors and research organizations. The main issues are the ambiguity of author names and institutional addresses as well as the unambiguous assignment of authors to research institutions. Expert organizations such as the Dutch CWTS and the Norwegian NIFU, but also the Italian CNR-IASI, the German Max Planck Society and the German Competence Center Bibliometrics currently deal with this situation by buying raw data from database providers (Clarivate Analytics, formerly Thomson Reuters, sometimes complemented by Scopus Elsevier) and constructing in-house databases with improved data quality.

4. Field averages are the most frequent frame of reference for citation impact in professional evaluation studies.

In order to analyse how the frame of reference was conceived in bibliometric studies, we analysed their evaluation objects and choice of metrics in more detail. Concerning evaluation objects, ROs were differentiated according to scale (number of institutes or universities) and scope (mono- vs. multi-disciplinary), while FI were distinguished according to funding units (research projects, scientists, ROs, portfolios). Concerning the choice of metrics, we distinguished impact metrics based on “international field averages” from impact metrics based on “national rankings” and “other”. Metrics of the type “international field averages” are is used in 65 % of the study sample. National rankings are used only in the evaluation of Italian university departments. Other frames of reference include mainly journal impact based measures and h-type indices, but also some quasi-experimental group comparisons.

5. The WoS classification of science fields (SCs) functions as a de-facto reference standard for research performance assessment.

Among 69 studies that used some form of field normalization, 84 % rely on the WoS classification of science fields (WoS subject categories), and an additional 3 % used the related Essential Science Indicators classification by Clarivate Analytics. The Scopus science classification is mentioned mainly in the Italian VQR reports. 13 % of field normalizations are based on self-defined journal sets, or keywords in combination with self-defined journal sets. To date, alternative field taxonomies proposed in the academic literature have had little influence. It can be concluded that the WoS classification of science fields has attained the status of a de facto reference standard.


The theoretical framework used raises questions regarding professional control over citation data and bibliometric indicators. Abbott asks how corporate control of expert commodities (in this setting, citation databases) will affect the future development of professional knowledge and practice. It seems likely that an open-access regime of citation databases would support the development and broad diffusion of more sophisticated bibliometric techniques for research assessment. Therefore, it seems promising to explore more explicit connections between methodological debates in evaluative bibliometrics and the ongoing open access transformation of the scientific publication system. Further research and policy discussions should focus on whether and how open access to citation data could be provided in Europe.

13:15-14:45 Session 9E: Technologies Emerging Globally

Global South

Location: GLC 222
Timing of Science, Technology, and Innovation Policy for Country-level Innovation
PRESENTER: Jennifer Woolley

ABSTRACT. Innovation is a policy imperative in many countries, backed by large budgets and considerable public attention. Science, technology, and innovation policies (STI policies) comprise a variety of government initiatives designed to support basic research, innovation and commercialization of inventions (Lundvall and Borras, 2005; Woolley and Rottner, 2008). Governments institute STI policies to heighten innovation and economic growth (Gault, 2011; Schot and Steinmueller, 2018; Flanagan et al., 2011; Morlachi and Martin, 2009) and cultivate new markets (e.g. Uyarra et al., 2014). As governments evaluate ways to become technology leaders, knowing when to enact STI policies has increased in salience. The timing of policy implementation has largely been neglected in the literature. Determining the optimal moment for policy implementation poses significant challenges to policymakers since STI policies require long-time horizons for assessment (Jaffe, 2011). This begs the question: Do countries that enact STI policies earlier during the emergence of a nascent technology see greater innovation? This research examines the advantages and disadvantages of enacting STI policy investments early in a nascent domain of activity. We apply the first–mover perspective to STI policy to gain insight into policy timing. We then examine the consequences of STI policy timing in nanotechnology across 53 countries. Lastly, we discuss theoretical contributions and policy implications.

Background STI policy encompasses government programs sponsoring scientific research, technology development, and innovation commercialization. Recent work has begun to examine the influence of policy timing. Woolley and Rottner (2008) find that U.S. states with the first STI policies for nanotechnology had higher rates of nanotechnology-related entrepreneurship. Beaudry and Allaoui (2012) found an exponential relationship over time between STI funding and related academic scientific productivity. Further, Ruan and colleagues (2014) show that STI policies co-evolves along with industry growth. These studies suggest that temporal dimensions affect the outcomes of STI policy initiatives, but the area remains under-examined in terms of innovation.

First-mover Advantage Perspective & STI Policy First-mover firms are pioneers that enter a market earlier than other firms. Lieberman and Montgomery (1988) proposed that firm-movers may earn higher economic profits than other firms due to the advantages of technological leadership, scarce asset preemption, and heightened buyer switching costs (see also Suarez and Lanzolla, 2007). First-movers are also susceptible to four types of disadvantages: free-ridership, technological uncertainty, enabling new entrants, and incumbent inertia. Analogous to first-mover firms, first-mover STI policies are pioneering public policy initiatives that jurisdictions (e.g. city, state, region, or country) enact before others to stimulate progress in a nascent science, technology, and innovation domain. As forerunners, first-mover STI policies support technologies at their earliest stages to increase the innovative output of the area. In contrast, later-stage policies tend to support incremental technological change and development. Not surprisingly, the foundation provided by successful first-movers is likely to influence the attributes of these later-stage policies. By applying the first-mover perspective to STI policy, we shed light on some temporal dynamics that previous research has left unexamined.

Advantages of First-Mover STI Policy As with firms, jurisdictions with first-mover STI policy may benefit from technological leadership. Early policies support basic scientific research when the specific avenue of research that will succeed is unknown. First-mover STI policies, therefore, often support a broad base of research and innovation (see Van de Ven and Garud, 1993) that can create a forerunner stronghold of tacit knowledge and intellectual property. First-mover jurisdictions may enable the lock-in of critical scare assets ahead of other areas, particularly if the asset is married to the geographic area or is difficult to. Organizations within those jurisdictions may benefit from lower switching costs through a reduced likelihood that firms or human capital will relocate (Pierson, 2000).

Disadvantages for First-Mover STI policy Similar to first-mover firms that endure free-riders, early STI policies expend considerable resources to launch infrastructure for further innovation development (Van de Ven, 1983). Those areas enacting later-stage policy may benefit by not needing to make as many investments, instead exploiting (or free-riding) the resources and systems developed by first-movers. The inherent uncertainty surrounding emerging technologies poses other disadvantages for first-mover locations if they fund technology that fails to develop. Likewise, policymakers may offer ineffective policy solutions as they struggle to identify the appropriate support for the new domain (Flanagan et al., 2011). The lack of legitimacy for new technological arenas constrains policymakers and their policies. Relatedly, nascent technologies are understood by fewer people leading to HR struggles. Those areas with first-mover STI policy can suffer from incumbent inertia by applying outdated approaches to an emerging field.

Methods To gain insight into the role of timing in STI policy on the emergence of a nascent technology, we focus on national STI policies that support nanotechnology development established between 1981 and 2012. These data were gathered from over 20,000 pages of archival data from over 1,000 documents and websites. In total, 53 countries were identified to have STI policy initiatives for nanotechnology before 2012. Country innovation level was measured by the country’s share of worldwide nanotechnology patents derived from the U.S. Patent and Trademark Office, European Patent Office and OECD. Independent variables include: year of the country’s first national nanotechnology initiative; the amount of government nanotechnology STI funding by each nation in 1990; the nation’s nanotechnology STI funding in 2003. Several economic controls were included. We estimated generalized linear models using logit.

Results & Discussion First-movers countries had higher related patenting activities. Whether we consider the year of STI policy “entry” or the level of investment, the findings remain qualitatively the same. Moreover, the amount of early investment was not as important as simply enacting early STI policy. Thus, first-mover advantages occur when STI policy begins before the technological uncertainty resolves, but the amount of the early investment is less important than simply being active in the fledging domain. Countries with later STI policy saw greater cumulative and long-term patenting. Thus, countries need not invest relatively higher amounts than others to reap late-mover benefits. Funding spent later in the development of a new domain garners greater innovation output than earlier funding. Together, these results reveal that the disadvantages of later-stage STI policy initiatives may be attenuated by higher amounts of funding. The effects of early and late STI policies are not simply additive. Countries with high amounts of nanotechnology STI funding in both 1990 and 2003 had lower nanotechnology patenting. These models indicate a later-mover advantage that is greater than the first-mover advantages. With globalization, firms easily move innovation activities offshore. Policies directly influence the infrastructure and environmental munificence available to firms and are important for managers to consider. Firms working on nascent sciences and technologies may want to locate in countries with early STI initiatives; whereas firms focusing on applications or sub-field of new technologies may find other countries with later, specific STI policies to be more supportive. This paper attempts to further the discussion about the temporal dynamics of STI policy by exploring its advantages or disadvantages in a nascent domain. We advance theory, provocatively using the first mover advantage perspective to consider the influence of the timing of STI policy on innovation. By better understanding the temporal elements of STI policy, policymakers may more effectively support the advancement of technology and foster innovation systems within their region.

Boundary Work and Transversality: A Framework for the North-South Governance of Emerging Technologies
PRESENTER: Susan Cozzens

ABSTRACT. Background

The US National Nanotechnology Initiative (NNI) set in motion a large-scale and long-term investment in nanotechnology bolstered by promises of commercial and societal potential across the globe (Roco 2001). This enthusiasm coalesced in the “next industrial revolution” and raised the prospects for a new “techno-economic paradigm” (Drechsler 2010, Perez 2012). That zeal was not restricted to the Global North as nations in the Global South, such as South Africa (RSA 2006). Many highlighted the potential benefits for societal challenges in the Global South (Salamanca-Buentello, 2005). However, the outcomes have not been commensurate with either the promise or the scale of the developmental challenge. Many researchers have argued that nanotechnology-based innovation is widening socio-economic disparities when the relevant communities are not explicitly included in the process (Invernizzi and Foladori 2006). This motivates us to ask: What theoretically-informed diffusion models might enable emerging technologies to better address developmental challenges in the Global South? How might North-South knowledge exchanges elucidate opportunities for improving living standards? Theoretical Grounding Our study engages with two theories: transversality from innovation studies, and boundary work derived from science and technology studies (STS). Innovation systems can be analyzed at the macro level (national innovation systems), meso level (sectoral, regional, technological systems), or at the micro level (private firm or organization). The concept of “transversality” offers a way to analyze a multi-level innovation system that may promote the diffusion of technologies through the interaction of various actors in diverse settings Cooke (2012). The notion of transversal governance in the public policy literature, however, integrates aspects of sub-national, national and transnational levels in a primarily consensus-oriented, rather than state or market-driven manner (Paquet 1999). Our first proposition is the existence of a “transversal innovation system” with developmental outcomes as a primary goal.

The second concept of boundary work was first introduced by Gieryn (1983). It attributes certain methods, values, types of knowledge and the organization of work that is within “science”, while socially differentiating it from other activities that are deemed to be “non-scientific”. Gieryn (1983) did not consider contexts where the asymmetry of privilege between “elite” professionals and other knowledge holders in non-industrial settings is so wide that the “defeated knowledges” of the latter are generally undermined to the detriment of livelihoods and life styles – thus the need to acknowledge “alternative sciences” (Visvanathan 2006). We adopt the perspective of Juma and Ojwang (1989) that technology from the informal sector in the Global South and its reciprocal integration with modern industry, represents a legitimate pathway for innovation in the Global South. Our second theoretical proposition is that a “transversal innovation system” is a system that bridges the bifurcation between socio-technical know-how in industrial settings and its manifestation in “resource poor” settings. This offers a framework with a vertical governance dimension, which cuts across local, regional, national, and international realms and horizontal dimension, which traverses the boundary between “low tech” and “high tech” knowledge for developmental purposes. Methods A 10-year research program on nanotechnology, equity, equality and responsibility was originally driven by Georgia Tech and Arizona State University researchers to understand the differential impact of nanotechnology research on societal outcomes in the US and the Global South Africa. Publications between 2002 and 2017 identified cases and then interviews, observations, and site visits gathered rich qualitative data. The cases involve US-based nanotechnology researchers working with Global South researchers and artisans to address clean water challenges.


The bibliometric analysis of publications update findings from (Cozzens et al. 2013). Interview data findings with Mexican, South African and US researchers revealed counter-intuitive innovation processes. One case reveals governance challenges across multiple system levels. Another shows how commercialization opportunities were sought for “niche” applications for affluent customers in the Global North, rather than for the Global South. US nano-scientists were interested in advancing the frontiers of knowledge in their disciplines, yet were also inspired by solving water challenges in the Global South.

Case 1: South Africa – Nano-enabled tea bag water filter. Although the innovation was first identified as useful in poor communities in South Africa, commercialization efforts re-aimed it at niche applications for “nature sport” practitioners in Europe. This shift reflects transversality in the vertical dimension, albeit through “reverse” (Global South to North) innovation. Transversality in the horizontal dimension was based on finding an application in another sector (sports), but it did not have any “low tech” dimension.

Case 2: South Africa – Nanomembrane brackish groundwater treatment plant. It was implemented at a school in a village in the Northwest Province. The vertical dimension of transversality was therefore demonstrated (diffusion within the water sector and from national-level research to the village level) but the project was not sustained due to challenges associated with governance at multiple levels. There was no evidence of transversality in the horizontal dimension.

Case 3: USA – Nano-enabled ceramic water filter solution partly developed in the US was implemented in two South African provinces, thereby demonstrating the vertical dimension of transversality. Horizontal dimension also demonstrated through the leveraging of South African potters’ skills for making the filters. This is the only case with a solution that is not patented. However, this case was based on a US researcher who established a not-for-profit organization in South Africa to diffuse the team’s nanotechnology-enabled point-of-use clean water solution. The original set of South African researchers with whom they collaborated, work on infectious diseases rather than on water. The know-how for manufacturing the ceramic water filters came in part from women potters in a Northern Province of South Africa, who make clay pots for ornamental purposes.

Case 4: USA – Nano-dust solution developed in the US for cleaning Arsenic in contaminated water was implemented in a pilot project in Mexico. Vertical dimension of transversality was demonstrated although evidence of sustainability is yet to be determined. Furthermore, knowledge exchanges between US and Mexican nano-researchers were established. However, there is no evidence of the horizontal dimension of transversality, whether in the high tech domain or with respect to “low tech”.

Case 5: South Africa (Possibly) – Non-nano case involving Durban latrine. The municipal water company was charged with providing a standard amount of clean water free to Zulu households in a remote suburb. The Afrikaaner water engineer respectfully worked with the multiple wives in the Zulu families to create cisterns that respected family relationships and filled overnight with “city water.” The working relationship allowed adjustment of ventilated pit latrines for these households as well, greatly reducing the risk of cholera epidemic for the whole city of Durban. Local Zulu workers were trained in the relevant skills as part of the project.

Significance This work furthers reconciliation efforts between Innovation Studies and STS (Martin 2012). Integrating transversality and boundary work offers a framework that traverses the Global South-North) in the vertical dimension with the horizontal dimension attending to knowledge exchange between “frontier” science and local knowledge. Theoretically, the concept of “boundary work” from the STS literature is expanded in a manner that enables us to reflect on what constitutes “science” and “non-science” as it relates to development collaboration between “high tech” university-based researchers and “low tech” community-based practitioners. This study offers suggestions for future research that addresses our broader concerns about nanotechnology and development.

Regional Technological Systems in Transition - Effects and Development of Technological Portfolios in China
PRESENTER: Henning Kroll

ABSTRACT. In recent years, the degree of variety and relatedness in regions' technological portfolios has received increasing attention as a determinant of techno-economic pathways and thus, ultimately, options and rationales for regional development (Isaksen & Trippl, 2016; Grillitsch, Asheim, & Trippl, 2018). Since Frenken et al. first proposed the notion of related variety (Frenken et al. 2007) in the industrial space various methodologies to capture technological diversity and relatedness have been further developed and arrived at a notable level of sophistication (Hidalgo et al., 2018; Balland et al., 2018). Empirically, however, most research in this area has remained based on evidence from Europe or the United States. There is some evidence on China but - with a view to Boschma and Capone's (2015) findings on the strong role of institutions - it must remain uncertain if even its more fundamental propositions, that have become accepted as common ground (Content and Frenken, 2016), will apply under emerging economy conditions. While the existing literature never made active claims towards universality, an exploration of existing assumptions' relevance in the Chinese context thus seems of sufficient importance on its own, as the world's second largest economy transitions from an externally to an internally driven mode of development. As economic transition and institutional difference hardly affect China alone, efforts to understand the applicability of central propositions of the existing literature in the Chinese context merit attention. Against this background, this paper will analyse the role of established, basic aspects of regional variety under the framework conditions of an emerging economy. Different from earlier studies that mostly consider patent activities as a variable dependent on variety in the industrial or product space (Castaldi, Frenken, & Los, 2015; Kogler, Rigby, & Tucker, 2013), our work follows the recommendation put forward by Content and Frenken (2016) by establishing patterns of variety in the technological space as well as their fit with the regional economy, to later refer them back to regional economic performance and development. Intentionally, it seeks to analyse those effects for an economy that remains differently positioned in global value chains than Western economies (Fu et al., 2012; Liefner and Wei, 2014) and, over the past two decades, has experienced fundamental economic and institutional transformation (Breznitz & Murphree, 2011; Peyman, 2018). Adding to ongoing shifts in the regional balance of its innovation system (Fan, 1995; Liu, Gao, & Wang, 2018; Liu & White, 2001; Kroll, 2016), moreover, several of its provinces are now moving dynamically towards an innovation-driven mode of development (Liu et al., 2018). In this fundamentally different and dynamic context, technological variety may have (had) very different origins and thus implications than in Western market economies. Moreover, its relevance should be strongly contingent on the presence of a regional "fit" between technological and sectoral activities that allows technological achievements to trigger effects in the regional economy (Liu et al., 2018). Overall, there are thus two main reasons why China constitutes a relevant study case. First, China's past trajectory suggests distinct relations between technological variety and economic development that should help us to put established assumptions in perspective. Second, recent literature on China seems to imply that new sources of technological variety may be increasingly emerging, motivating a closer investigation if that is truly the case. Methodologically, it will do so in a threefold manner. First, by developing standard variety measures on an entropy basis, a more advanced coherence measure based on patterns of co-patenting and a measure of regional fit between technological and economic as well as scientific and technological structures. To the authors' knowledge, this information has so far not yet been available for the technological space in China. Second, by mapping those across China's provinces, including a detailed cluster analysis taking into account the various different aspects of regional technological portfolios outlined above. This allows us to gain a more structured understanding of China's provinces regional innovation systems as well as their general positioning on techno-economic development pathways. Third, by analysing in panel models if these diverse aspects of China's regions' technological portfolio have during the 2007-2016 period meaningfully related to regional GDP levels and regional GDP growth. Furthermore, we will explore to what extent the increasing emergence of certain types of variety have been dependent on specific characteristics of China's regional economies in post crisis technological uptake period following 2011. The comparatively clear finding of our research is that established wisdom on the role of technological variety does not seem to be applicable in the entirely different economic context of China. The GDP level depends positively on general specialisation, not variety. Nonetheless, there are first traces that the overall situation may be changing with economic growth reacting positively to technological diversity (advanced measure) and urban, science-based environments demonstrating increasing tendencies to develop related variety (entropy measure). Additionally, we find clear indication that the fact whether regional technological activities match the profile of local industries do influence its observable effect on both GDP levels and GDO growth notably. These findings are relevant for future research from a twofold perspective. First, they underline that while prior research has revealed an important general principle of regional path development, its relevance remains strongly context specific. At earlier stages of economic development, different development logics may apply and in countries positioned differently in global value chains and with different institutions, other factors may superimpose local trajectories. Second, it underlines that irrespective which processes are at play, their effect remains contingent on the local economic context into which they are placed. Even our first, sketchy consideration of this aspect through local "goodness of fit" variables illustrates this point very clearly.

The evolution of transnational R&D: the case of US MNEs and India and China
PRESENTER: Ajinkya Kamat

ABSTRACT. Background: With globalization transforming the the global economic landscape over recent decades, multinational enterprises (MNEs) are key drivers of globalization and major players in the global innovation landscape. In 2015-16, the world’s top 100 R&D-investing MNEs accounted for more than a fifth of the global expenditure on R&D [calculations based on (R&D magazine 2016) and (EU Scoreboard 2015)]. Especially over the last two decades, MNEs have been expanding their research and development (R&D) activities beyond North America, Western Europe, and Japan to developing countries. Globally, US MNEs dominate the MNE R&D activities: one in every three of the world’s top 2500 R&D-investing companies is a US MNE, significantly more than any other country (EU Scoreboard, 2015).

Particularly since the turn of the century, India and China have become two of the most attractive destinations for MNE R&D investments and their new R&D centers (Kekic, Lofthouse, & Whyte, 2004; UNCTAD, 2005). In both of these countries, US MNEs are dominant players in the MNE R&D landscapes. Studies on the globalization of R&D typically focus on international trends in MNE R&D investments or in patenting by MNEs, or they focus on R&D activities of MNEs in a specific country along with the implications for that country (with some studies focusing on a specific industry sector within a country).

In addition to exploring patterns in the R&D activities of US MNEs and their shift towards developing countries, this paper presents a comparative assessment of the MNE R&D landscapes in India and China -- two countries which have become more attractive destinations for R&D activities of US MNEs than many developed countries.

Data: To explore transnational patterns in the US MNEs’ R&D activities, we use data related to their R&D investments (an input to R&D activities), patents from their R&D activities outside the US (a proxy for outputs of inventive R&D), and other key innovation indicators. To discuss the evolution of MNE R&D landscapes in India and China, we use a number of indicators related to R&D investments by US MNEs in the two countries across industry sectors as well as patenting activity in and from the two countries (i.e. patents filed by Indian and Chinese organizations and MNEs at Indian and Chinese patent offices as well as outside these countries). This paper utilizes quantitative data from various sources, such as the US Bureau of Economic Analysis, USPTO, UNESCO, and OECD, along with interviews of researchers and research managers in industry and universities, as well as literature on MNE R&D and, broadly, on innovation systems in India and China.

Results: We show that US MNEs’ R&D investments have shifted towards emerging economies of India and China, much more than other developing countries. While Canada, Germany, and UK remain the top locations for these R&D investments, their share in US MNEs’ global R&D investments has declined, while India's and China's has increased.

Comparative analysis of the MNE R&D landscapes in India and China reveals three key findings. First, consistent with the literature on globalization of R&D, enormous growth of the research ecosystem and market opportunities in China can explain why it is a prominent destination for US MNEs’ R&D investments. On the other hand, although the scale of the research ecosystem and market in India is much smaller than China (but comparable a few other developed countries), both the countries receive equivalent R&D investments from US MNEs. In fact, Indian R&D centers of US MNEs contribute to more US patents than the R&D centers in any country other than the US.

Second, US MNEs’ R&D investments in both India and China are concentrated in two industry sectors namely the professional, scientific, and technical services and the manufacturing sector (within manufacturing specifically the computer and electronics products and the chemicals manufacturing subsectors). While India receives more R&D investments in the services sector, China is ahead in the manufacturing sector.

Third, MNE R&D activities have a significant share in inventive R&D in both the countries, more so in India than in China. 85% of all US patents with India-resident inventors filed during 2010-2014 were owned by foreign (not Indian) MNEs, whereas only 47% of such patents with China-resident inventors were owned by foreign (not Chinese) MNEs. 8 out of the top 10 organizations, which owned the largest number of the patents with India-resident inventors, were MNEs. In contrast, the corresponding number for China was 2 out of the top 10 organizations. A significant number of patents co-owned by Chinese organizations with foreign MNEs indicate the prevalence of joint-ventures between Chinese firms and foreign MNEs. The number of patents co-owned by Indian firms with foreign MNEs is negligible.

Discussion: We further explore the determinants of some of these patterns through interviews of researchers and heads of MNE R&D centers in India and researchers at Indian universities. They point to the interplay between a few factors that have made India and China equally attractive to MNE R&D activities, despite significantly different scales of innovation ecosystems and market opportunities in the two countries: - Foreign direct investment policies in India let MNEs retain majority or full ownership of their subsidiaries and hence intellectual property, whereas China’s policy emphasizes joint-ventures with and technology transfer to Chinese firms. - The intellectual property (IP) laws, their enforcement, and overall lower propensity of IP laws in India, compared to China, is one reason why MNEs favor R&D in India. - English being a common language of instruction in higher education and of business communication in India, the Indian R&D centers of MNEs do not experience language gap with their counterparts in the US and Europe; in China such language gap is more likely. - The growth of market and research ecosystem in China, however, seem to counteract the factors discussed above and provide enough motivation for US MNEs to invest heavily in R&D in China. Given the much larger manufacturing ecosystem in China, which serves Chinese as well as global market, Chinese affiliates of US MNEs generate significantly higher sales in the manufacturing sector than the Indian affiliates.

These findings and insights are valuable for policy makers, particularly the R&D funding agencies, administrators of higher education and research institutes, and domestic industrial enterprises in India and China to facilitate exchange of knowledge, technology, and financial and human resources, among the MNE R&DCs and the rest of the innovation systems in these countries. There are also considerable prospects for MNEs to diversify their research portfolio in the rapidly growing Indian and Chinese innovation systems.

13:15-14:45 Session 9F: Space Policy
Location: GLC 324
Influence of Commercial Satellite Operator Data Sharing and Emerging Space Nation Stakeholder Preferences on Space Traffic Management System Architecture
PRESENTER: Miles Lifson

ABSTRACT. Space sustainability and mechanisms for the coordination of space traffic are drawing increasing attention from the associated technical and policy communities. This presentation discusses work to understand the needs of two key stakeholder groups, commercial satellite operators and nations with emerging space programs. It then analyses how their preferences influence and constrain the design-space for a space traffic management (STM) system. Work is underway to begin designing a U.S. STM system, which will powerfully influence subsequent efforts to integrate STM globally. The conclusions of this work help assess the usefulness of different strategies for STM coordination to inform future research, uncover inherent system design trades, and promote the development of a more just and effective system.

Background and rationale The use of the Earth’s orbits for satellite operations is a classic environmental commons problem. All nations have an equal right to access and use outer space under international law, but orbiting satellites occupy a portion of finite orbital volume, with certain orbits being particularly desirable due to inherent physical properties. When satellites collide, or fail to be disposed of responsibly at end of life, they produce long-lived orbital debris which poses a collision risk for other satellites. This debris and the need to expend fuel maneuvering to avoid potential collisions are negative externalities associated with use of the space commons. Coordination to avoid collisions also requires sharing information about a satellite’s properties, operations, and intentions, which many national security and commercial operators are reluctant to do. Increasing demand from current and emerging space actors, plans for novel very large constellations of satellites, and clear consequences from past debris generation events have motivated global recognition that existing regulatory and coordination structures are inadequate to safely handle the projected increase in space traffic. Work towards STM is ongoing at both technical and policy levels, within the U.S. and internationally. Nevertheless, a variety of factors make such work incredibly challenging. As space sustainability expert Brian Weeden argues, “space traffic management is a ‘super wicked’ public policy problem that involves balancing an indefinable set of technical, legal, and economic variables; conflicting interests and worldviews of many stakeholders; and a complex political environment with diffuse responsibilities and authorities.”

Methods The systems architecture framework provides a useful mechanism to understand and design complex engineer systems. Developing from systems engineering work for the aerospace and semiconductor industries in the mid-20th century, the framework decomposes complex problems and associated system designs into objectives, stakeholders, forms, and functions. Prior work by the author has applied the systems architecture framework to STM and identified two major gaps in the literature with significant impacts on ultimate system architecture: 1) an understanding of the nature of commercial operator objections to information sharing to enable orbital coordination and 2) an understanding of the views of countries with emerging space programs. This work seeks to address these two forms of STM stakeholder uncertainty with the aim of understanding how associated stakeholder preferences constrain and enable potential STM system designs. Understanding the needs of these stakeholders introduces limits on the STM system design-space and make it more manageable. The views on data sharing of commercial operators are critical due to their political clout, their dominant role in overall space utilization (including several potential mega-constellations), and because system-wide coordination to prevent orbital collisions must be able to respect and function under requirements for limited information distribution. The views of nations with emerging space programs (and especially developing nations with emerging space programs), is critical to ensuring that an STM system is just and reflects the needs of all actors, not just entrenched players. Existing decision-making processes and inertia already ensure the inclusion of established major space operators, but there is no similar guarantee for nations with emerging space programs. Technical and standards development work favors the viewpoints of those who participate in such processes, which skew towards better-resourced groups with entrenched interests. Work to actively solicit, understand, and include emerging space nations earlier in the process allows for system architecture choices that may involve little cost early in system design, but would be prohibitively expensive if added later in the process. Satisfying the needs of emerging space nations will drive wider and earlier adoption, ultimately enhancing the overall effectiveness, as STM depends on widespread participation to be effective. For this work, semi-structured interviews were conducted with a wide variety of stakeholders from both groups and are continuing. The interview questions and protocol were vetted with subject matter experts prior to the stakeholder interviews. The authors also conducted an extensive academic, policy, and technical literature review. Emerging space nations were selected using a set of criteria built on Wood’s space participation metric. Countries were then binned across regions, human development index, and level of space activity. A representative set of major commercial satellite operators was selected to ensure representation of low Earth orbit (LEO), geosynchronous (GEO), and both remote sensing and communications operators.

Anticipated Results Privacy views differ significantly between commercial space operator niches. GEO communication satellite operators are most restrictive regarding SSA information sharing, but primary need to coordinate with a limited number of orbital neighbors due to the physics of geosynchronous orbits. LEO and remote sensing operators tend to be much more willing to share SSA information. Operators generally agree on the set of potential competitive threats posed by SSA information sharing, but diverge on how much those factors are relevant to their particular business. The work will also explore the extent to which views vary between functional units in a company (i.e. government relations, corporate strategy, flight dynamics). LEO presents the largest challenge for STM coordination and would benefit from different coordination mechanisms than GEO featuring different levels of data sharing and different trust structures. Developing nations with emerging space programs view space sustainability as important and generally welcome policy and technical work to safeguard space sustainability. They are concerned about the potential technical and financial burden of international guidelines or an international STM system. They are generally supportive of mutually agreed capacity building by established space nations and the continued provision of free SSA services. There is a belief that nations who contributed to the existing debris landscape should shoulder the burden of mitigating it. At a more detailed level, views differ significantly across countries, regions, and development levels. Countries differ in whether they would support a globalized STM system (as opposed to some level of international standardization of interfaces between national systems). Nations also differ in the extent to which they wish to be involved in the definition of a process for STM development or in the actual development process. Analysis to understand how these preferences limit the STM system design-space is still ongoing.

Significance This work fills in gaps in public understanding regarding the needs of two important stakeholders in any future STM system. By doing so, it constrains potential system designs and futures for consideration, helping to guide technical research directions and making system-associated policy dilemmas and trades more manageable and explicit. The overall approach of identifying and leveraging key stakeholder interests to restrict design-space consideration for complex engineering systems intended to address wicked problems has widespread applicability to other challenging contemporary dilemmas.

Success and Failure in National Security-Civil Cooperation: Examining the Case of Interagency Sharing of Satellite Reconnaissance Data

ABSTRACT. Successful cooperation between national security and civil agencies is notoriously difficult, fraught with organizational, security, and ethical challenges. Yet, when successful, these interagency efforts can result in significant benefits for society. Programs to share satellite reconnaissance data provide an interesting case study in this area. The intelligence community, which collects and analyzes data from United States reconnaissance satellites and other sources, has a long history of sharing this data outside the intelligence community to support civil government efforts. However, the motivations and organizational arrangements for doing so have varied significantly across projects and across time, with some efforts achieving greater success than others. This paper reviews the evolution of reconnaissance satellite data sharing programs over time and examines how these policies have affected the civil geospatial information community.

The United States began a reconnaissance satellite program at the very outset of the space age, successfully launching hundreds of national security surveillance satellites over the decades. Beginning in the late 1990s, data from these government-owned satellites was augmented with data purchased from the newly-established U.S. commercial remote sensing sector, increasing the repository even further. Although the purpose of these programs was to provide intelligence for national security purposes, and the data was classified, it was clear from the beginning that the data could also be valuable to civilian researchers and operators. High-resolution imagery could be used to update maps, provide assistance in recovery from natural disasters, support natural resource management, enable environmental research, and inform decision-making.

Since the U.S. government has already paid for these systems, and the data has already been collected (or is being collected), making the data available for additional purposes makes it possible to increase the benefits to society at a very low additional cost. These potential benefits were deemed to be large enough that it was worth making some effort to provide the civil community with access to the data. One of the primary ways that this sharing was undertaken was within the auspices of the Committee on Civil Applications of Classified Overhead Remotely Sensed Data, generally referred to as the “Civil Applications Committee (CAC).”

This paper traces the evolution of data sharing in the CAC over more than four decades, identifying key transition periods in which data sharing and use was expanded. Understanding the factors that led to these increases provides insight to help to improve future efforts to share national security satellite data with the civilian sector, ensuring that the U.S. government – and society as a whole – get the greatest possible benefit out of the investment in these systems, while remaining sensitive to potential security issues associated with sharing this data. Analyzing these models can also provide insight into interagency data sharing projects more broadly, providing a better understanding of what makes this type of effort more or less successful.

Helping the Little Guy: The Impact of Government Awards on Small Technology Firms
PRESENTER: Aleksandar Giga

ABSTRACT. The United States Small Business Innovation Research (SBIR), one of the pillars of government effort to enhance innovation, has offered research and development (R&D) awards to firms with 500 or fewer employees since 1982. The rationale of the program is straightforward; by providing early-stage financing to small companies that might not attract funding in the market, the program focuses on the population with the highest potential for growth, both in revenue and employment.

We explore the differential effects of the National Aeronautics and Space Administration (NASA) SBIR program on firms of various sizes. In particular, we study the second-stage (larger) SBIR award and its impact on the invention performance of so-called “microfirms” of 1-5 employees, comparing it with that of larger small businesses (6-500 employees). We find that microfirms with SBIR awards are approximately 60% more to produce patents following the award, and furthermore they generate twice as many patents compared with the microfirms without the awards; the program does not show the same effect for larger firms (6-500 employees). By using NASA’s internal ranking data embedded in the selection process, we explore the program’s ability to identify good quality projects, an important contribution, given the difficulty of assessing merits of small firms without prior track record.

Our results suggest that the size limitations of the eligible firms should be reconsidered, especially considering that only 15% of NASA SBIR Phase II funds went to microfirms. This is consistent with the change in the labor/capital ratio of large corporations over the last decades and is particularly true for high-tech companies that lead the market cap rankings. At a minimum, our results suggest that further research on the design of the SBIR program is necessary.

Active Space Debris Removal: Responsibility, Timeline, and National Security Implications

ABSTRACT. I. Background and Rationale

Since the historic launches of humanity’s first satellites in the late 1950’s, the number of objects orbiting Earth has steadily increased. However, only a small fraction of the man-made objects that remain in these orbits are operational. Low Earth Orbit (LEO), the region of space within 2,000 kilometers of Earth’s surface, is increasingly congested with orbital debris and decommissioned spacecraft that no longer serve a useful purpose. The rising popularity of small satellites and mega-constellations consisting of hundreds to thousands of satellites will severely compound the orbital debris situation. Without Active Debris Removal (ADR), the likelihood of on-orbit collisions and accidental damage to functioning spacecraft will continue to increase.

The majority of the global population utilize satellites in LEO, particularly those used for telecommunication and data transmission, on a daily basis. As a result, the “tragedy of the commons” situation applies. In other words, individual space actors behave according to thier own self-interest, and not the common good of all users, by polluting the shared resource that is the near-Earth space environment. The orbital debris problem has received increased attention in recent years, but deploying an ADR mission is an expensive and technologically complex undertaking.As a result, there is no public consensus on which space actor(s) should be responsible for implementing ADR.

The basic ADR mission concept involves a powerful spacecraft equipped with complex robotic technology, precise maneuverability, and advanced control software. The autonomous rendezvous and proximity operations capabilities needed to de-orbit a large piece of orbital debris could also be used for malicious purposes, such as de-orbiting functioning satellites. Regardless of intentions, any country or private entity who develops ADR capabilities in space poses a national security threat to other nations. Therefore, a successful implementation of ADR must also include measures to clearly define and communicate the scope and intent of the mission in an international forum.

II. Methods and Anticipated Results

In the absence of a profitable private investment model for ADR missions, the responsibility to mitigate and remove orbital debris is likely to fall on the government sector. However, ADR does not clearly fall into either NASA’s or the DOD’s organizational objectives, which makes securing U.S. government funding difficult. Expert panelists at ESA’s Clean Space meetings did not expect to rely exclusively on public funding for ADR. Therefore, public-private partnerships will be investigated in various forms. One option is a disposal levy system funding and governance model comparable to the Kyoto Protocol approach to reduce greenhouse gas emissions. A variety of incentive schemes to motivate the commercial sector will also be considered, such as an X-prize model to foster ADR technology development. Yet another option would be to impose a tax on new launches to fund ADR, forcing all space actors who contribute to the orbital debris problem to take responsibility for its solution.

To assess the immediacy with which the orbital debris situation in LEO should be addressed, reports of space objects tracked by the U.S. Strategic Command (USSTRATCOM) and NASA are considered alongside satellite launch forecasts. Ideally, the influx of new satellites launched into LEO would be counterbalanced by de-orbiting satellites that have been decommissioned. NASA’s Orbital Debris Program Office set an international timeline for this purpose: satellites should be de-orbited within 25 years after their operational life ends. However, adding the propulsion system needed to de-orbit a spacecraft is expensive and countries are not required to enforce the 25-year guideline. In 2015, 35 percent of satellites were out of compliance. The current debris environment, launch forecast, and expected de-orbit rate are the primary factors used to characterize the urgency with which an ADR system is needed.

In 2007, the Chinese anti-satellite (ASAT) test intentionally destroyed its Fengyun-1C satellite. This ASAT activity resulted in over 150,000 pieces of debris tracked by NASA, making it the largest space debris-creating event in history, and sparked debate among the international community about space weapons. These debates did not end in formal international agreements on the complex issue of space weaponization, but they highlighted the need for communication and transparency in space activities. In the context of ADR, the same system used to de-orbit large pieces of orbital debris in LEO could be used to de-orbit another nation’s valuable space assets. To prevent this outcome, ongoing and transparent international communication will be critical. The Committee on the Peaceful Uses of Outer Space (COPUOUS) is a prime setting for these international discussions.

III. Significance

Outer space is often referred to as a “global commons” and humanity benefits greatly from satellite technology on a day-to-day basis. The notion that outer space is so vast and free that it is immune to pollution by humankind is the same mindset that society once held towards Earth's oceans. This mindset resulted in irrevocable damage to the ocean ecosystem before the global community realized it needed to educate the public about the dangers of pollution, mitigate debris growth, and take aggressive action to remove debris from the ocean so that humanity can continue to reap its benefits. The growing congestion of LEO poses a significant risk to society’s ability to continue deriving benefits from space over the long term. As levels of orbital debris increase, risks to current and future satellites also increase. ADR is a promising mission concept to maintain the accessibility of space and reduce risks of on-orbit collisions.