📚 Research Digest

Finding: Feedback (both trial-by-trial and summary) significantly improves children's ability to strategically seek information to fill knowledge gaps, outperforming their own subjective uncertainty judgments as a guide for self-regulated learning.

Why it matters: Directly relevant to learning sciences and skill acquisition, with implications for designing educational technology (tablet-based learning), feedback mechanisms in games/gamification, and behavior change systems that scaffold metacognition when self-monitoring is weak.

Method: Two experiments with 203 children aged 5-9 using a tablet-based word-learning task, measuring information-seeking behavior and uncertainty calibration after different feedback conditions.

Jargon: *Metacognitive monitoring* = awareness of what you know/don't know; *metacognitive control* = adjusting behavior based on that awareness; *knowledge gap* = difference between current and target knowledge state.

Finding: Introduces "metacognitive ownership" — a framework for how learners govern AI contributions to their learning (standards, evaluation, and action decisions) — arguing that accepting AI advice without examining it is distinct from meaningfully directing it.

Why it matters: Directly relevant to teaching about AI's impact on learning, skill acquisition, and human-AI collaboration, particularly how AI tools may undermine or support learner agency and self-regulation.

Jargon: *Metacognitive ownership* = learner governance over what criteria to use, how to evaluate work, and what actions to take when AI assists; *evaluative judgment* = the capacity to assess quality of one's own work; *human-AI regulation* = shared control of learning processes between human and AI system.

Finding: Blue-collar manufacturing workers conceptualize human-AI collaboration as a two-step process (AI suggests → human verifies/acts) grounded in verification and control rather than trust, with AI literacy being a key moderator of collaboration quality.

Why it matters: Directly relevant to AI impact on human collaboration and work, and the trust/verification dynamic parallels themes in the "Collectively Negotiating Trust and Mistrust in Generative AI" paper the user found relevant.

Method: Qualitative interview study (n=11) with Finnish manufacturing workers, using transparency cues and explanations as stimuli to probe human-centered AI feature preferences.

Jargon: *Industry 5.0*: EU framework emphasizing human-centric, sustainable industry alongside automation. *Transparency cues*: design features that make AI reasoning visible to users. *AI literacy*: task-specific understanding of how AI works and its limitations.

Finding: Brief exposure to a forgiving, idealized AI partner in a trust game *reduced* subsequent cooperation with a standard partner, consistent with a negative contrast effect (expectations set too high by the idealized partner made the normal partner seem worse by comparison).

Why it matters: Directly relevant to designing AI collaborators, persuasive technology, and behavior change interventions — the finding that "idealized" agent design can backfire has implications for gamification, AI tutors, and any system attempting to use positive relational experiences to shift behavior.

Method: Randomized experiment (N=206) using Hidden Markov Model (HMM)-based artificial agents simulating human-like trust dynamics in repeated trust games; within-subject baseline/exposure/post-exposure design.

Jargon: *Rejection sensitivity (RS)*: a personality trait characterized by anxious expectations of interpersonal rejection. *Internal working models*: mental schemas for how relationships work, from attachment theory. *Negative contrast effect*: a drop in valuation of a target stimulus after exposure to a superior comparison stimulus. *Trust game*: an economic game measuring cooperative investment between partners.

Finding: Video gamers report motivations centered on competence, mood management, social connection, and positive disengagement, while toxicity, unfairness, and stress contribute to negative psychological outcomes.

Why it matters: Provides qualitative texture on psychological need satisfaction/frustration in games, relevant to gamification design and understanding player engagement mechanisms including addictive vs. restorative dynamics.

Method: Reflexive Thematic Analysis on 15 semi-structured interviews — small sample, qualitative only, limited generalizability.

Jargon: *Basic psychological needs* — from Self-Determination Theory: competence, autonomy, relatedness; *Reflexive Thematic Analysis* — a qualitative method where themes are actively constructed by the researcher rather than discovered.

Finding: Emotional intelligence (EI) correlates positively with perceived teamwork effectiveness in Gen Z students (r=0.36–0.57), with empathy and social skills being comparatively weaker components; self-motivation strategies including goal-setting and role clarity emerged as key collaboration supports.

Why it matters: Touches on team dynamics, goal-setting, and emotion regulation in collaborative learning contexts, which are relevant to courses on teams and behavioral science.

Method: Convergent mixed-methods design combining ANOVA/Pearson correlations on 206 students with reflexive thematic analysis of 15 semi-structured interviews.

Jargon: *CATME* = Comprehensive Assessment of Team Member Effectiveness, a peer-rating instrument for teamwork; *reflexive thematic analysis* = qualitative approach where researcher reflexivity is foregrounded in theme development.

Finding: In people with depressive symptoms (dysphoria), temporal self-comparison feedback (how you've improved over time) dominates self-assessment more than social comparison feedback, reversing the typical pattern seen in healthy populations.

Why it matters: Has modest relevance to behavioral science, motivation, and feedback design—particularly for understanding how feedback framing (social vs. temporal) affects self-evaluation and performance, which matters for gamification, goal-setting, and learning system design.

Method: False feedback experiment (N=159) using a visuospatial 2-back task with manipulated concurrent social and temporal comparison information.

Jargon: *Dysphoria*: subclinical depressive symptoms; *HAPSCI*: "higher assigned priority of social comparison information"—the typical finding that people weight social comparisons more heavily than self-comparisons over time; *2-back task*: a working memory task requiring participants to identify stimuli matching those presented two steps earlier.

Finding: An AR framework combining brain-based learning and cultural values (Tri Kaya Parisudha) for early childhood character education showed positive impacts on prosocial, critical thinking, and communicative behavior in preschool children.

Why it matters: Marginally relevant to learning sciences and AR/interactive media design, but the focus on early childhood character education in a specific cultural context (Balinese/Indonesian values) is far from the professor's core interests in HCI, creativity, teams, or behavioral science.

Method: Quasi-experimental design using the ADDIE instructional design model with expert validation.

Finding: Perceived LMS quality (across system, information, instructional, interaction, and learning quality dimensions) explains 35.1% of variance in student engagement (r=0.593) among 45 students using Moodle.

Why it matters: Marginally relevant to learning sciences and HCI/UX in educational technology contexts, but the tiny sample (n=45), narrow context, and purely correlational design limit applicability.

Method: Simple linear regression with cross-sectional survey; effect size is large (Cohen's f²=0.542) but n=45 is severely underpowered for generalization.