Chatbot preference groups differed in total EBAPS and Axis 4/5 scores at the nominal level, with small effect sizes, but not after Bonferroni adjustment¶

Claim · All claims
Evidence · 1 study (2 entries) · associational r2 · q2 · i0 negligible
Currency · newest study 2026 · checked against its sources 2026-10-20 (automated, 2 of 2 entries) · not reviewed by a person

Subclaims¶

q2 i0 Kruskal-Wallis tests showed at least one preference group differed on total EBAPS scores and on Axis 4 and Axis 5, with small effect sizes (total 0.004; Axis 4 0.007; Axis 5 0.006). → Amogh Sirnoorkar and Omkar Mamidpalliwar 2026 q2 i? Under a Bonferroni-adjusted significance level (α=0.0083), the omnibus differences for the total score and Axes 4 and 5 were no longer statistically significant. → Amogh Sirnoorkar and Omkar Mamidpalliwar 2026 (2)

Evidence¶

Amogh Sirnoorkar and Omkar Mamidpalliwar 2026¶

Amogh Sirnoorkar and Omkar Mamidpalliwar. (2026). Students' Epistemological Beliefs and their Chatbot Preferences in AI-mediated Physics Learning. https://arxiv.org/abs/2607.29385 doi:10.48550/arxiv.2607.29385

q2 · i0 · associational · r2

Kruskal-Wallis tests (α = 0.05, scores not normally distributed) on 1048 overlapping EBAPS/activity responses. Table III prints effect sizes of 0.004 (Total), 0.007 (Axis 4), and 0.006 (Axis 5); the article states "the small effect sizes indicate that chatbot preference accounted for only a modest proportion of the variability".

"The results indicate that at least one pair of chatbot preference groups (Inquiry, Answer, and Combination) differed significantly in their overall EBAPS scores and in their scores on Axis 4 (balance between extremes of absolutism and relativism) and Axis 5 (success in science through innate ability versus effort)."

Amogh Sirnoorkar and Omkar Mamidpalliwar 2026 (2)¶

Amogh Sirnoorkar and Omkar Mamidpalliwar. (2026). Students' Epistemological Beliefs and their Chatbot Preferences in AI-mediated Physics Learning. https://arxiv.org/abs/2607.29385

q2 · i? · associational · r2

Robustness check applying a Bonferroni-adjusted significance level (α=0.0083) across the six omnibus Kruskal-Wallis tests; the printed nominal results did not survive the correction.

"Under this more conservative criterion, the omnibus differences observed for the total score and Axes 4 and 5 no longer highlighted statistical significance."

Discussion¶

Learner Variables¶

• Learning-Style Preferences — predictor: learners reporting different styles differed here; not a test of matching instruction to style

Related Claims¶

• Reflective thinking dimensions correlate positively with learning process/expert knowledge, and most correlate positively with learning effort and negatively with innate/fixed ability and certainty of knowledge — the general claim this finding bears on, with the studies behind it
• Students preferring combination-style chatbots had higher total EBAPS scores than students preferring direct-answer chatbots (nominal significance only) — a narrower finding that bears on this claim
• No statistically significant EBAPS differences were observed for Axes 1–3 or between guided-inquiry and answer-oriented chatbot preferrers — related
• Combination-preferring students scored higher than answer-preferring students on evolving-knowledge (Axis 4) and source-of-ability (Axis 5) EBAPS dimensions — a narrower finding that bears on this claim
• Most introductory physics students preferred a combination chatbot (guidance first, answers when stuck), while few preferred pure guided Q&A — related
• The focus-on-students effect was not robust: significant in the ANOVA but not in non-parametric tests — related
• Instructional approaches do not differ significantly in monitoring or reflecting, and omnibus differences in orienting and planning show no significant pairwise differences — related