Finding: A critical review and z-curve analysis of 52 significant results across 18 articles finds that embodied motivation research (e.g., power poses modulating motivational states) has a ~22% estimated replicability rate, driven by outcome switching, analytical flexibility, and misinterpretation of non-significant results.
Why it matters: Directly relevant to behavioral science and persuasive technology teaching—this is a replication/credibility crisis paper that undercuts a popular class of interventions (posture-based motivation hacks) often cited in applied behavior change contexts.
Method: Z-curve analysis, which estimates the expected replication rate from the distribution of z-scores across published significant findings—a meta-scientific technique for detecting publication bias and inflated false-positive rates.
Jargon: *Z-curve*: statistical tool estimating true discovery rate from the distribution of significant p-values; *Analytical degrees of freedom*: researcher choices in analysis that inflate false positives without pre-registration; *Embodied cognition*: theory that mental states are shaped by physical body states.
Finding: This literature review synthesizes research on GenAI in higher education, covering mental health support, cognitive offloading, AI dependence, equity, and the bidirectional relationship between psychological distress and AI use.
Why it matters: The cognitive offloading and AI dependence findings are directly relevant to learning sciences and AI's impact on human cognition, and the equity/access dimensions matter for anyone designing AI-integrated educational experiences.
Method: Literature review synthesizing recent empirical and theoretical work on GenAI and student wellbeing.
Jargon: *Cognitive offloading* = delegating cognitive tasks to external tools rather than performing them mentally; *GenAI dependence* = habitual reliance on generative AI that may reduce autonomous thinking.
Finding: Descriptive review of AI applications in inclusive education for learners with disabilities, covering speech recognition, adaptive learning, and intelligent tutoring systems, while identifying barriers like bias, cost, and privacy.
Why it matters: Tangentially relevant to AI's impact on learning sciences and skill acquisition, but the disability/accessibility focus is outside the professor's core teaching areas.
Method: Descriptive (non-systematic) literature review — lower evidentiary value.
Claudia Calle Müller, Mais Kayyali, Patrizia Santaniello • Semantic Scholar
Finding: A survey of 135 Construction Management students identifies academic, financial, and social factors that drive or hinder motivation, recommending institutional interventions like peer mentoring, career fairs, and time management workshops.
Why it matters: Tangentially relevant to learning sciences and motivation, but focused narrowly on a specific vocational discipline (construction management) with generic findings unlikely to add theoretical or pedagogical insight for HCI, creativity, or behavioral science teaching.
Method: Single-institution survey (n=135) at Florida International University.
Finding: A literature review proposing the "I-DIAL" instructional design model for Islamic education that integrates AI, adaptive pedagogy, and digital technology with Islamic pedagogical principles.
Why it matters: Tangentially touches on learning sciences and AI in education, but the focus is narrowly on Islamic religious education curriculum design, making it minimally applicable to HCI, creativity, behavioral science, or general learning sciences research.
Method: Comparative conceptual analysis of existing instructional design models (ADDIE, Dick and Carey, ASSURE, Backward Design, Kemp, Design Thinking).
Johan Klarin, Eva Hoff, Daiva Daukantaite • PsyArXiv
Finding: Students with executive functioning difficulties are more likely to rely on generative AI for schoolwork, mediated sequentially through perceived usefulness and habitual use; habitual use showed the strongest direct association with reliance (β = .60–.69).
Why it matters: Directly connects AI adoption, habit formation, self-regulation, and behavior change mechanisms in an educational context—relevant to understanding how AI tools become dependency-forming and what that means for learning and skill development.
Method: Structural equation modeling with sequential mediation tested across two large community samples of secondary school students (N = 735 and 839 GenAI users).
Jargon: *Executive functioning (EF)*: cognitive control processes like planning, inhibition, and working memory. *Reliance*: dispositional preference for AI over own effort or teacher support. *Sequential mediation*: indirect effect through two mediators in sequence (EF difficulties → perceived usefulness → habitual use → reliance).
⚠️ No control group mentioned | Single study (not replicated)
Zoe Fowler, Clin KY Lai, Roger Beaty et al. (4 authors) • PsyArXiv
Finding: Two studies (N=162) testing whether collaborative vs. individual future imagination affects divergent thinking (measured via semantic distance in spoken narratives) found mixed/null results, with social connection showing a complex, negative relationship to divergent thinking scores.
Why it matters: Directly relevant to teaching about team creativity, collaborative cognition, and collective intelligence—the null/mixed results are themselves instructive about when social context helps or hurts creative output.
Method: Divergent Semantic Integration (DSI) quantifies divergent thinking by measuring semantic distance between all words in a participant's narrative—a computational, language-based operationalization of creativity worth knowing.
Jargon: *Divergent Semantic Integration (DSI)*: a computational measure of how conceptually spread-apart the words in someone's speech are, used as a proxy for divergent (creative) thinking. *Co-imagination*: imagining the future collaboratively rather than individually.
⭐ Includes replication ⚠️ No control group mentioned | Convenience sample (e.g., MTurk, students)
Finding: Proposes the EbD4ESG framework, positioning UI/UX design decisions (including dark patterns, consent mechanisms, and inclusivity) as governance mechanisms within ESG accountability structures.
Why it matters: Directly relevant to HCI, dark patterns, ethical persuasion, and choice architecture — the argument that interface design shapes user autonomy and informed consent connects to behavioral science and persuasive technology ethics.
Method: Structured conceptual synthesis drawing on design ethics, HCI, and ESG governance literature; no empirical validation.
Jargon: *Ethics-by-design* = embedding ethical principles into products during design rather than retrofitting them; *ESG* = Environmental, Social, and Governance — corporate accountability framework; *dark patterns* = manipulative UI designs that exploit cognitive biases to benefit the company at user expense.
Finding: MAOV, a closed-loop HCI framework using multimodal behavioral data (five synchronized streams), machine learning clustering, and an Interaction Bottleneck Index, significantly reduced task completion time (31–50%), error rates (57–65%), and cognitive load (NASA-TLX −40.6%) in smart home interaction redesign.
Why it matters: The behavioral granularity approach and closed-loop optimization methodology are directly applicable to UX design research and demonstrate how combining behavioral data streams with user segmentation can systematically identify and fix interaction friction points.
Method: Within-subjects experiment (N=40) comparing baseline vs. optimized prototypes across seven tasks; behavioral archetypes derived via K-means++ clustering on neural network embeddings from multimodal data fusion.
Jargon: *SUS* = System Usability Scale (standard UX questionnaire, max 100); *NASA-TLX* = Task Load Index (cognitive workload measure); *IBI* = Interaction Bottleneck Index (composite friction metric per interaction node); *AMBF-Net* = neural network fusing eye-tracking, gesture, touch, audio, and physiological signals.
⭐ Longitudinal design ⚠️ Small sample (n=40) | No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Finding: Affora is a design system that makes interfaces legible to both human users and AI agents simultaneously, showing that substantial visual variation is possible while preserving machine-readable interaction meaning.
Why it matters: Directly addresses HCI and UX design in the context of AI agents operating software—relevant to teaching about AI's impact on interface design and the emerging challenge of designing for dual audiences (humans + AI).
Method: Three controlled studies examining component implementations, visual variation, and interaction-design principles, plus evaluation on independently authored interfaces.
Jargon: *Computer-use agents* = AI systems that autonomously operate software UIs (clicking, typing); *affordance* (implied by "Affora") = the perceived action possibilities of interface elements; *interface representation* = how an interface is encoded/perceived by a machine reader (e.g., accessibility tree, screenshot).
Alyssa H. Sinclair, Mallory Harris, Danielle Cosme et al. (9 authors) • PsyArXiv
Finding: Interactive behavioral interventions (text, quiz, map-based) communicating the economic impact of science funding cuts reduced approval of those cuts and increased action intentions across political divides, with active engagement (quizzes, maps) selectively boosting behavioral intentions beyond attitude change.
Why it matters: Directly relevant to persuasive technology design, choice architecture, and behavior change — specifically how interactivity and surprise drive engagement and action, with implications for nudge design and civic tech.
Method: Two preregistered experiments (N=5,177) plus a naturalistic website study (N>24,000) testing text vs. quiz vs. map interventions with politically-balanced samples.
Jargon: *Action intentions* = self-reported likelihood of taking civic actions (e.g., contacting representatives); *preregistered* = study design and hypotheses publicly committed before data collection to prevent p-hacking.
⭐ Pre-registered ⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Patrick Wegerer, Bernhard Schenkenfelder, Rudolf Ramler et al. (5 authors) • Semantic Scholar
Finding: LLM-generated low-code configurators can automatically produce multiple interface styles (wizards, dashboards, profile-based guidance) from software product line feature models, reducing cognitive load for non-technical users navigating complex variability constraints.
Why it matters: Touches on HCI/UX design principles (reducing cognitive load, interface generation), AI's impact on software development workflows, and accessibility of technical systems to non-experts—relevant to AI-augmented work and UX design topics.
Method: Exploratory experiments across proxy scenarios combined with expert interviews from SPL and HCI specialists.
Jargon: *Software Product Lines (SPLs)*: families of related software systems sharing managed variability; *Feature models*: structured representations of configurable options and their dependencies; *Low-Code Development (LCD)*: visual/automated approaches to building software with minimal hand-coding.
⚠️ No control group mentioned | Single study (not replicated)
Abejide Ade-ibijola, Damola Olugbade, S. Oyelere • Semantic Scholar
Finding: A quasi-experimental study found that students using an AI chatbot ("Ovayo") for sustainability education scored significantly higher on conceptual knowledge post-tests than traditional instruction or control groups, and qualitatively reported the chatbot as interactive and adaptive.
Why it matters: Provides empirical evidence on AI-mediated personalized learning and chatbot-supported instruction, relevant to AI's impact on learning sciences and skill acquisition.
Method: Quasi-experimental mixed-methods design with three groups (chatbot, traditional instruction, control) using pre/post knowledge assessments plus semi-structured interviews.
Jargon: *DigCompEdu* — European framework for digital competence of educators, covering pedagogical use of digital tools.
Finding: Theoretical analysis of psychological determinants of goal-directedness in university students, focusing on volitional traits, academic motivation, self-regulation, and self-actualization.
Why it matters: Touches on goal-setting and motivation topics relevant to the professor's interests, but appears to be a general educational psychology overview without novel empirical or practical contributions.
Method: Theoretical/methodological analysis only — no empirical data.
Jargon: *Goal-directedness* — the psychological orientation toward purposeful, sustained pursuit of objectives; *volitional traits* — willpower and self-control characteristics.
Rachel J. Anderson, J. H. C. McClure, J. Shevchenko et al. (5 authors) • Semantic Scholar
Finding: Episodic simulation (mentally imagining future events) increases feelings of controllability over personal goals, particularly for individuals with high depressive symptoms, but does not improve likelihood estimates, importance ratings, or anticipated emotions.
Why it matters: Has marginal relevance to goal-setting and motivation research, but the findings are narrow and largely null, limiting practical application for teaching behavioral science or goal-setting frameworks like OKRs.
Method: 2x2 experimental design comparing goal-related vs. goal-unrelated episodic simulation against neutral imagery and letter-search controls, with repeated measures of expectancy ratings.
Jargon: *Episodic simulation* = mentally pre-experiencing a future event in detail; *expectancies* = beliefs about likelihood, controllability, and emotional outcomes of future events.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Novia Setyoningtyas, S. Utaminingsih, Ahmad Hariyadi • Semantic Scholar
Finding: Researchers used the ADDIE instructional design model to develop a Scratch-based arithmetic learning app, finding that stakeholders valued intuitive interfaces and color/image-based visualization, while recommending peer collaboration features and better reward mechanisms.
Why it matters: Touches on gamification, learning sciences, and motivation design in educational tools, which overlaps with the professor's interests in serious games and skill acquisition.
Method: Qualitative stakeholder interviews (n=4) analyzed descriptively — very small sample limits generalizability.
Finding: Mastery-oriented classroom goal structures improve student academic engagement partly through boosting self-efficacy, with self-efficacy mediating ~50% of the relationship.
Why it matters: Tangentially relevant to goal-setting and learning sciences, but focused narrowly on secondary school motivation in Pakistan rather than HCI, design, teams, or behavioral intervention.
Method: PLS-SEM with 10,000 bootstrapped resamples across 300 students; standard mediation analysis.
Ni Made Citariani, K. Agustini, I. K. Suartama et al. (6 authors) • Semantic Scholar
Finding: An AR framework combining brain-based learning principles with Balinese cultural values (Tri Kaya Parisudha) for early childhood character education showed strong expert validity and positive impacts on children's prosocial, critical thinking, and communicative behavior.
Why it matters: Tangentially relevant to learning sciences and educational technology design, but the focus on early childhood character education in a specific cultural context is far from the professor's core interests in HCI, teams, creativity, or behavioral science.
Method: Quasi-experimental design using the ADDIE instructional design model with expert review validation.
⚠️ No control group mentioned | Single study (not replicated) | Limited statistical detail in abstract
Juliastuti Juliastuti, I. Budiana, Amita Diananda et al. (4 authors) • Semantic Scholar
Finding: Curriculum innovation integrating technology and adaptive learning in Islamic secondary schools shows four key challenges: curriculum transformation, teacher digital readiness, infrastructure disparities, and adaptive learning implementation for personalized learning.
Why it matters: Touches on adaptive learning and technology integration in education, but the Islamic education context and Society 5.0 framing make this peripheral to the professor's core interests.
Method: Descriptive qualitative with semi-structured interviews, observations, and document analysis with triangulation — straightforward and not particularly novel.
Finding: A graduate BIM course used adaptive instructional strategies (Bloom's taxonomy, scaffolded instruction, project-based learning) to support students with varying technical backgrounds across three learner profiles.
Why it matters: Marginally relevant to learning sciences and skill acquisition, but the domain (building information modeling) and focus (inclusive pedagogy logistics) are too narrowly applied to add much to the professor's core teaching areas.