Compatibility Between AI Mediation and Learner Proximity: Testing Question (Q) vs Answer (A) Orientations in AI-Driven Learning
How to cite
Malik, Tariq. 2025. "Compatibility Between AI Mediation and Learner Proximity: Testing Question (Q) vs Answer (A) Orientations in AI-Driven Learning." AEA RCT Registry. AEARCTR-0017006. https://doi.org/10.1257/rct.17006-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
This study investigates how artificial intelligence (AI) interacts with human learning proximity—specifically, whether AI-driven instruction is more compatible with Question (Q) or Answer (A) proximity orientations of learners. Building on proximity theory and cognitive load frameworks, the experiment examines whether AI aligns more naturally with the learner’s tendency to construct knowledge through questions (Q-proximity) or to consolidate understanding through answers (A-proximity). The study employs a randomized 2×2 factorial design manipulating (1) proximity orientation (Q vs A) and (2) AI compatibility level (High vs Low). Participants complete microlearning tasks within an AI-assisted learning environment where AI either mirrors or mismatches their proximity orientation. The primary outcomes include engagement, arousal, uncertainty reduction, and learning performance. The results will clarify the cognitive–affective mechanisms by which AI complements or conflicts with human learning orientations, offering empirical guidance for the design of adaptive and personalized AI learning systems.
Primary Research Question
Is AI-driven instruction more compatible with Question-oriented (Q-proximity) or Answer-oriented (A-proximity) learners, and how does this compatibility influence engagement, arousal, and performance?
Hypotheses
H1: AI–learner compatibility (matching AI mode with learner proximity type) leads to higher engagement and performance than incompatibility.
H2: AI’s answer-generative nature makes it more compatible with A-proximity learners, improving near-term learning outcomes.
H3: Q-proximity learners benefit more when AI adopts a Socratic or exploratory questioning mode, supporting curiosity and transfer.
H4: The effects of compatibility on performance are mediated by arousal and uncertainty reduction.
Experimental Design
A between-subjects randomized controlled trial (2×2 factorial):
Factor 1: Learner proximity orientation (Q-proximity vs A-proximity)
Factor 2: AI compatibility (High: AI mirrors learner orientation; Low: AI mismatches it)
Each participant engages in an AI-guided learning module (e.g., concept learning in economics or statistics).
AI responses are scripted to simulate either Socratic questioning (Q-compatible) or direct answer tutoring (A-compatible).
Treatment Arms
Q-Compatible AI: AI adopts a questioning style (Socratic prompts) aligned with Q-oriented learners.
Q-Incompatible AI: AI provides direct answers to Q-oriented learners.
A-Compatible AI: AI provides structured answers aligned with A-oriented learners.
A-Incompatible AI: AI prompts learners with open questions, mismatching their A-orientation.
Primary Outcomes
Learning engagement: Behavioral (interaction rate), affective (interest/arousal), cognitive (depth of inquiry).
Learning performance: Immediate quiz, far-transfer, and one-week retention.
Mediators: Arousal and uncertainty reduction.
Expected Contribution
This study contributes to understanding human–AI compatibility in educational contexts. By isolating whether AI naturally favors answer-based or question-based learners, it informs the design of adaptive pedagogical AI capable of shifting between inquiry and explanation modes to optimize learning engagement and performance.
Key Details for Registry Form
Experimental unit: Individual learner
Sample size: N = 300 (balanced across four cells)
Randomization: Computerized block randomization
Setting: AI-enabled online learning platform (simulated or live environment)
Analysis: GLM and SEM with mediation (Arousal → Uncertainty → Performance)
Primary outcome timeframe: Immediate and one-week delayed post-tests
IRB: From Liaoning University
Intervention
1) Research aim & hypotheses
• Aim: Test whether brief, music-based prompts can shift students’ learning mode toward question-ownership (vs. answer-ownership) during an interactive session with teacher/peers/AI.
• H1 (primary): Treated students will show a larger increase from pre to post on a Question-Ownership Index (QOI).
• H2 (secondary): Treated students will report higher excitement and confidence, and lower anxiety about learning challenges.
2) Design overview
• Participants: ~60 undergrads + 10 postgrads you already teach (n≈70).
• Arms (between-subjects RCT):
o Treatment (Q-lyrics): 1 song version emphasizing curiosity, asking, exploration (lyrics/questions).
o Control (A-lyrics): Same melody/length/tempo, lyrics emphasizing getting correct answers/solutions.
• Exposure: 6–8 minutes total (one full play + short reprise), identical volume and environment.
• Timing: Pre-survey → music exposure → 10–12-minute interactive task (pair + AI or peer + teacher mini-rounds) → Post-survey.
3) Measures (pre & post)
Use your 20-item instrument as the backbone. Map items into outcomes:
(A) Primary outcome: Question-Ownership Index (QOI)
• Build from Q1–Q11 (dyads among Teacher/Student/Peers/AI).
• Create 2 subscales:
o Who owns the Question? Q-items where the left role asks (e.g., Student–Teacher means S asks, T answers).
o Who owns the Answer? Q-items where the right role provides answers.
• Scoring rule: For each item, higher agreement = stronger endorsement of the presented configuration. Code items where the question is owned by the learner as +1 direction; configurations where the answer is owned by others as −1 direction (reverse-code so that higher always = more question-ownership). Then average and rescale 0–100.
• Report QOI total and (optionally) two facets:
o Personal Q-ownership (student-as-questioner items: S–T, S–Peers, S–AI).
o External A-ownership (teacher/peers/AI-as-answer items; reverse-coded).
(B) Affective & self-efficacy outcomes
• Excitement (Q18), Confidence (Q19), Anxiety (Q20, reverse); rescale 0–100.
(C) Manipulation checks (post only; add 3 quick items)
• “The song made me want to formulate questions.”
• “The song made me focus on getting correct answers.” (reverse)
• “During the activity I generated new questions.”
(Use 5-point Likert; 0–100 rescale.)
Demographics/covariates (pre only): degree (U/P), major, gender, university, GPA.
Use the exact wording from your sheet for Q1–Q20 to preserve validity.
Experimental design
Design overview
Participants: ~60 undergrads + 10 postgrads you already teach (n≈70).
Arms (between-subjects RCT):
Treatment (Q-lyrics): 1 song version emphasizing curiosity, asking, exploration (lyrics/questions).
Control (A-lyrics): Same melody/length/tempo, lyrics emphasizing getting correct answers/solutions.
Exposure: 6–8 minutes total (one full play + short reprise), identical volume and environment.
Timing: Pre-survey → music exposure → 10–12-minute interactive task (pair + AI or peer + teacher mini-rounds) → Post-survey.
Design details
Cited by
No citation data available for this registration yet.