ABSTRACT. Using a nonrandomized pretest-posttest control-group design, we tested whether paraprofessionals can improve text-sourced academic vocabulary for secondary students with reading difficulties, including multilingual learners (MLLs). Forty-five students (23 MLLs) received twice-weekly small-group instruction embedded in a 70-minute multicomponent intervention; controls learned comparable, non-text-sourced words. Researcher-developed receptive measures were administered around three thematic units, across three school years (2022-2025) in total. MLLs showed significantly larger gains on taught words (~10% higher), and tutor effects were nonsignificant.

ABSTRACT. Three experiments examined whether informational uncertainty promotes strategic source use during evaluative reading. Undergraduates judged whether expert fictitious posts should be published online. Low-certainty statements consistently reduced publication recommendations and increased source-based justifications, even when familiarity was controlled. However, uncertainty did not enhance delayed source–content recognition. Results indicate that informational uncertainty functions as a situational cue shaping trust evaluations and sourcing behavior but are inconclusive on its effect on source-content memory.

ABSTRACT. This study examined the role of the note-taking condition, the establishment of causal connections and modality of presentation in the comprehension of spontaneous discourse. We asked a group of Argentine college students to listen to or read the material, perform a note-taking task, and then to perform a note-review and a note-taking and comprehension difficulty rating tasks. Results indicated that establishing meaningful connections interplays with note-taking in written and spoken discourse comprehension.

ABSTRACT. This study examined whether text-to-speech (TTS) supports digital reading by reducing state mind-wandering in primary and lower-secondary students. Across self-paced and TTS conditions, students completed comprehension tasks and cognitive measures. TTS led to shorter completion times and fewer lookbacks, indicating greater reliance on memory rather than rereading. TTS also reduced state mind-wandering, which fully mediated higher comprehension accuracy. Working memory and comprehension skills independently predicted performance, underscoring key human–technology interactions in digital reading.

ABSTRACT. Learning from scientific text is a complex task. The overlapping nature of linguistic and cognitive processes involved in comprehension means offline assessments fail to capture how comprehension unfolds in real-time. This study constructs a model of scientific text comprehension using a battery of assessments alongside eye movement data from university students (n = 100). Linear mixed-effects models and structural equation modeling were applied to text- and reader-level features to predict and explain comprehension.

ABSTRACT. Linguistic features influence the extent to which people accept a message’s claim. However, it remains unclear which combination of linguistic features shapes message acceptance most. We analysed this using a large corpus of 2,567 pinned user comments on Dutch news articles, using the number of likes as a proxy for message acceptance. Results show that comments with many likes tend to be more persuasive, with little repetition, more reflexive pronouns, and greater use of intensifiers.

ABSTRACT. This presentation focuses on a multimodal discourse analysis of a digital video, "Dinosaurs," created by two students (age 14) in a classroom setting. The analysis centers on the film's use of floating and empty signifiers, concepts that are developed and clarified in order to examine their function in the focal video across the planes of meaning and subjectivity, thus offering new methodological tools and analytical units for multimodal discourse analysis.

ABSTRACT. What size groups are best for learning? To test this question, students in Research Methods courses were randomly assigned to work in groups of 2, 3, or 4, on discussion activities intended to support comprehension of short scientific reports. Results suggested that "Pair-and-share" activities may not be as effective as assumed. Groups of four appeared more likely than pairs to confer the intended advantages to weaker students from interacting with peers.

ABSTRACT. This study examines the discourse marker "well" in written fiction using Stance Theory and Dialogic Syntax. Drawing on 19 English novels, it analyzes how "well" functions as a meta-stance marker and compares its use in fiction with spoken interaction. Focusing on stance types, shifts, and dialogic features such as resonance and parallelism, the study shows that fiction preserves spoken-like local stance management in quoted dialogue while restricting broader narrative-level uses, revealing genre-specific constraints on stance-taking.

ABSTRACT. Political polarization is tied to science denial. This study explored how polarized political environments influence adolescents’ scientific reasoning. Students from liberal and conservative environments completed a task, connecting evidence to competing scientific explanations. After the task, adolescents in liberal environments showed more scientifically accurate evaluations. Adolescents in conservative environments were more likely to consider a less scientific explanation as plausible, potentially mirroring reasoning patterns observed in conservative adults, leading them astray from more accurate evaluations.

ABSTRACT. This study examined adult learners’ ability to resolve lexical ambiguity in non-literal and ambiguous sentences. We found that the ability to identify semantic relationships (i.e., via Word Class) strongly predicted non-literal sentence comprehension. Ambiguous sentence comprehension is predicted evenly by vocabulary knowledge. Findings will be discussed in terms of understanding the interplay between vocabulary knowledge, complex texts, and adult learner abilities. Findings may also inform future interventions that support complex text structures for adult learners.

ABSTRACT. Comprehension often requires strategic engagement with texts. Adults that range in proficiencies in the foundational skills of reading may have different strengths and challenges in how they strategically engage with texts. In the present study, adult literacy learners, less proficient college readers, and more proficient college readers thought aloud while reading narrative texts. The think aloud responses revealed the strengths and challenges in terms of comprehension strategies across these three samples.

ABSTRACT. We examined whether college students’ intellectual humility (IH) indirectly links metacognition and reading comprehension. Students completed measures of IH, metacognitive awareness of reading strategies, reading comprehension, and self- and other-judgments of performance. Mediation analyses showed that IH significantly mediated relationships between general metacognitive awareness and comprehension, and between judgments of others’ performance and individual outcomes. Weaker indirect effects emerged for self-judgments. Results suggest that greater IH may support lower-skilled readers in developing more accurate meta-comprehension.

ABSTRACT. Many college students arrive underprepared for the comprehension demands of university coursework. Learning-by-explaining can improve understanding, but struggling students often need support that is difficult to provide in large, introductory courses. GenAI chatbots may fill this gap, but risk promoting cognitive offloading. We present findings from in-class studies in introductory biology examining how chatbot design and instructional prompts shape students’ interactions with the bot and how these interactions affect comprehension strategies, explanation quality, and comprehension.

ABSTRACT. Adults who do not excel at academic skills exhibit writing difficulties, an area not typically addressed in adult literacy (Graham & Perin, 2007). We surveyed 33 instructors about their level of preparedness to teach writing, and their instructional practices as well as assessed 86 adult learners to explore their digital writing strengths and weaknesses. This presentation will share the results as well as provide implications for professional development of instructors and instruction for learners.

ABSTRACT. AI tools are increasingly used to adapt literary texts according to reading level. How popular models interpret adaptation prompts and the metrics they use are unknown. We systematically compared how ChatGPT, Claude, and Gemini adapted excerpts from three canonical works using prompts for different reader populations. Analyses using multiple metrics of text features in discourse processing found that the output of models vary widely, reflecting different adaptation strategies or assumptions about readability and accessibility.

ABSTRACT. This study builds on previous work examining the fairness of LLMs in terms of non-standard dialect evaluations. Specifically, it examines 394 pairs of student responses: (a) with naturally produced African American English and (b) with the AAE feature standardized. It also examines feedback across a baseline prompt (1) and one providing specific instructions about AAE (2). Results show the feedback provided by prompt 2 scored significantly lower on three readability metrics compared to prompt 1.

ABSTRACT. Learning to read within a discipline is an important goal for a university degree. Students need support to learn in psychology, and one impasse is that introductory textbooks require knowledge-based inferences in their coverage of theories and studies. In this study, we found that students learned significantly more of the theory and studies from a revised version of the textbook with knowledge inferences filled in than the published version.

ABSTRACT. Task-unrelated thought (TUT) is negatively associated with learning, yet little work has examined whether moments of inattention lead to further inattention. Across two studies, we examined how starting with difficult lecture material influences perceived difficulty and TUT. We found that experiencing initial difficulty was associated with easier material being perceived as more difficult and higher rates of TUT. These results suggest that subjective difficulty is a key driver of sustained inattention.

ABSTRACT. This study investigated how multilevel textual features impact reading behaviors during parallel text processing (original and simplified versions side-by-side). Participants (n=100) read the story while their eye gaze was recorded via webcam. Linguistic features were analyzed against reading behaviors including number of switches between versions and fixation duration. Results showed that more difficult words and high sentence complexity significantly increased switching, indicating that readers utilize the alternate text strategically to mitigate increased linguistic challenges.

ABSTRACT. Mind wandering (MW), a cognitive state in which thoughts move freely, is shown to negatively influence comprehension. It is theorized that to experience MW one must disconnect from external stimuli i.e., perceptually decouple. We challenge this by presenting evidence that MW can occur while being perceptually coupled. Perceptually coupled MW also increased creativity levels, suggesting that such an experience might improve comprehension by helping readers make better elaborative inferences.

ABSTRACT. Situation model theories are highly influential—but the majority of accounts remain comparatively underspecified at the formal level. We propose a model using possible world semantics, treating worlds as hypothetical spaces rather than metaphysical entities. Updating is operationalized as a weighted Bayesian-style updating procedure, combining an agent-specific probability function with a multi-parameter relevance function. This framework potentially allows us to formalize dynamic changes in graded beliefs as the narrative unfolds.

ABSTRACT. We conducted a meta-analysis of 11 educational intervention studies, from 1,742 screened, examining how spatially enriched computational tasks (e.g., robotics and block-based coding) affects computational thinking (CT) learning outcomes. Using Ainsworth's (2006) DeFT framework, we found stronger effects when spatial and computational elements were aligned within the same task versus treated separately. This suggests that deliberately coordinating spatial and computational elements may be more effective for strengthening both CT and programming proficiency.

ABSTRACT. Event schemas support the activation of causal antecedents and consequences. Experiments manipulating causal direction revealed a robust consequence bias: participants verified consequences faster and more accurately than antecedents. Explicitly instantiating the schema significantly strengthened this bias. These findings suggest that the activation of event schemas associated with goal-directed actions supports causal prediction more so than causal bridging.

ABSTRACT. This usage-based study examines how 660 Hebrew speakers from fourth grade through adulthood represent events in written personal narratives. The analysis contrasts simplex and complex predicates, showing that simplex predicates remain stable across development, while complex predicates show significant growth. Two developmental patterns emerge: a "juvenile pattern" of linear event representation and a "mature pattern" of hierarchical, multidimensional event construal. Eleventh grade marks a key phase in this syntactic-cognitive reorganization of narrative event representation.

ABSTRACT. AI misconceptions are widespread in education, and chatbot-based refutations have shown promise for misconception correction. Drawing on the Elaboration Likelihood Model (ELM), this study examines how empathetic and simple refutational chatbot designs engage users in different information-processing routes. Using linguistic analysis of chatbot-user dialogue, results suggest empathetic chatbots elicit more affective language, while simple chatbots elicit more analytic language, with similar cognitive engagement. These findings indicate different, albeit effective for revision, processing routes.

ABSTRACT. This study examined whether a brief instructional video could help students to more effectively use a chatbot to support explanation writing. Undergraduates (n = 95) were randomly assigned to instructional video (guided by ICAP principles, control) and then used ChatGPT to write a biology explanation. Results suggest the instructional video promoted more constructive and interactive engagement and yielded significantly higher causal words, suggesting a promising approach to supporting explanation tasks in large-enrollment courses.

ABSTRACT. This mixed-methods study analyzes 200 newspaper and magazine articles to identify salient indicators of misinformation and disinformation in generative artificial intelligence (GenAI). It examines how these indicators contribute to the distortion of factual claims and the reconfiguration of epistemic practices. Through the application of text-analysis software to the collected articles, the findings emphasize the importance of pedagogical frameworks in GenAI literacy and discourse studies that foster the critical evaluation of AI-amplified information ecosystems.

ABSTRACT. The study investigated graduate students’ reflections on the use of GenAI in learning tasks as guided by the Ed-AI Literacy framework (Allen & Kendeou, 2024). We employed case-based learning to explore students’ analysis of a scenario depicting the framework components (knowledge, evaluation, collaboration, contextualization, autonomy and ethics). Results show learners’ understanding of genAI and whether it reflects current theoretical issues in AI literacy and education.

ABSTRACT. Language evolves alongside communication technologies, which shape not only vocabulary but norms of expression and interpretation. This position paper examines how large language models introduce a new influence on writing by turning common stylistic features into signals of suspected AI authorship. Using the em dash as a case study, we argue that writers increasingly engage in defensive writing, strategically avoiding certain constructions to maintain legitimacy. This shift adds cognitive burden and risks stylistic homogenization.

ABSTRACT. This qualitative study examines how Black primary caregivers navigate family structure and socioeconomic challenges while supporting their children’s reading at home. Using Bowen Family Systems Theory, thematic analysis, and cross-case analysis, findings highlight how family and economic factors shape home literacy practices. Additional findings show how family system characteristics buffer these challenges. Results challenge deficit narratives and offer insight into strengthening family-centered approaches to literacy development in socioeconomically marginalized contexts.

ABSTRACT. This paper aims to unpack self-regulated learning processes in learning from multiple digital resources. Guided by MD-TRACE and Winne and Hadwin’s self-regulated learning models, it conceptualizes multiple-resource learning as five phases: task model formation, goal setting and planning, initial single-resource use and processing, multiple-resource use and processing, and formation and assessment of task product. For each phase, the paper articulates underlying cognitive processes and outcomes and explains how these can be monitored and regulated.

ABSTRACT. This qualitative case study examines ways in which a middle school science teacher uses scaffolds in order to support students with disability labels as they evaluate competing explanatory models of socioscientific issues. We found that scaffolding questions, discourse scaffolds, and teacher expectations emerged as key instructional components. This suggests that dialogic talk and scaffolding can play an important role in supporting students with disability labels to make sophisticated judgments about socioscientific issues and models.

ABSTRACT. Students’ demographics are reported to higher education systems. These systems often offer limited lists of demographic categories not fully capturing the variance in demographics in the student population. The current study compares students’ (n=521) self-described responses to an open-ended demographic questionnaire and their institutional demographic records. Results showcase the fact the institutional demographic data does not capture the variance in the student population. We discuss implications and offer recommendations for improving demographic data collection.

ABSTRACT. One hundred and six undergraduates read contradictory claims with source information (author credentials, publication venues). Using a “judgment of source memory” paradigm, students rated their likelihood of remembering sources for each text read. Subsequently, they completed a source memory test and intellectual humility (IH) survey. Results indicated greater IH leads to more accurate source metamemory judgments and enabled readers to accurately calibrate high confidence for both beneficial and detrimental text source metamemory judgements.

ABSTRACT. This study examines how perceived topic importance is reflected in accuracy- and affirmation-directed reasoning across scientific contexts. In a reading and writing activity, students rated topic importance as personal, community, both, or neither, and selected reasoning justifications for four science topics. Results showed that topic importance judgments varied across contexts, whereas reasoning responses were predominantly accuracy-directed. Multinomial logistic regression further indicated that reasoning justifications did not differ systematically across categories of perceived topic importance.

ABSTRACT. This study examined accurate and inaccurate knowledge, and confidence in that knowledge, about the learning sciences and college teaching and learning in 145 undergraduate students enrolled as psychology, education, or other majors at a regional university in the Midwest United States. Confidence scores were higher than knowledge across all majors overall. Compared to other majors, psychology majors demonstrated both higher accuracy regarding learning science misconceptions and self-rated knowledge about effective learning in college.

ABSTRACT. Feedback helps students improve their performance and guide self-regulated learning. However, feedback is ineffective if students do not attend to it, trust it, or believe they can act on it. Drawing on theories of feedback-seeking as cost-benefit decision making, we examined the roles of growth mindset and self-efficacy in feedback-seeking. The findings indicate that growth mindset indirectly influences feedback-seeking through enhanced academic self-efficacy, suggesting the need to consider domain-specific appraisals in growth-oriented academic behaviors.

ABSTRACT. How people coordinate attention and language depends not only on task demands but also on how information is distributed across collaborators. We report two eye‑tracking studies examining how symmetric versus asymmetric roles shape gaze and linguistic alignment during joint route planning (Study 1) and visual search (Study 2). In route planning, performance tracked role-dependent temporal patterns of gaze alignment; visual search showed no such modulation, consistent with divide-and-conquer strategies. Ongoing transcription analyses will clarify coordination.

ABSTRACT. 81 half-hour conversations from the CANDOR corpus (Reece et al., 2023) were analyzed for instances of successful interruptions, in which one member of a dyad takes the conversational floor away from the other. Nearly six hundred such episodes were identified and analyzed to identify underlying systematicity. Gender was not related to successful interruptions, but older participants were more likely to successfully interrupt, suggesting that a speaker’s age affects whether a floor exchange will take place.

ABSTRACT. Digital platforms allow instructors to embed rubrics into formative tasks. This study examined whether the order of rubric performance levels (ascending vs. descending) influences engagement and task performance in a writing assignment. Fifty-seven students completed the task while their attention to the rubric was tracked. Students focused most on the highest performance level, which predicted better outcomes. Placing the highest level first appeared to draw attention to it more strongly than when it appeared last.