Objective learning gains did not differ significantly by condition in the primary ANCOVA, but KBD showed a significant positive effect on learning gains in exploratory regressions accounting for learner covariates¶

Claim · All claims
Evidence · 1 study (3 entries) · causal r1–r2 · q2–q3
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? In the ANCOVA with pretest as covariate, the KBD group's mean learning gain (0.75 points) versus the control group's (0 points) was not significant. → Stephan Vonschallen 2026 q3 i? In exploratory regressions with covariates, the KBD condition coefficient was positive and statistically significant (B = 1.714). → Stephan Vonschallen 2026 (2) q3 i? Participants who preferred learning in groups had significantly greater learning gains than those preferring to learn alone (B = 0.667). → Stephan Vonschallen 2026 (3)

Evidence¶

Stephan Vonschallen 2026¶

Stephan Vonschallen, Karim Kaufmann, Dominique Oberle, Friederike Eyssel, Theresa Schmiedel. (2026). Teachy Mini: Development and Preliminary Evaluation of a Knowledge-Based Generative Social Robot for Higher Education. arXiv. https://arxiv.org/abs/2607.22345 doi:10.48550/arxiv.2607.22345

q3 · i? · causal · r2

ANCOVA on the 7-item posttest with pretest score as covariate; the overall model was significant (F(2, 21) = 4.506, p = .024, 23.4% variance explained), but the condition effect was "not significant".

"The KBD group showed a mean learning gain of 0.75 points (SD= 1.54), compared to 0 points (SD= 1.54) in the control group. However, this effect was not significant,B= 0.447,t= 0.967,p= .344,95% CI ={−0.514, 1.407}."

Stephan Vonschallen 2026 (2)¶

Stephan Vonschallen, Karim Kaufmann, Dominique Oberle, Friederike Eyssel, Theresa Schmiedel. (2026). Teachy Mini: Development and Preliminary Evaluation of a Knowledge-Based Generative Social Robot for Higher Education. arXiv. https://arxiv.org/abs/2607.22345

q2 · i? · causal · r1

Post-hoc exploratory linear regressions adding gender, age, education, prior robot experience, baseline competence, general motivation, group learning preference, and robot attitudes as covariates; the KBD coefficient was "positive and statistically significant".

"Further, after accounting for these covariates, the coefficient for the KBD condition was positive and statistically significant (B= 1.714,t= 2.355,p< .05,CI= [0.142, 3.285])."

Stephan Vonschallen 2026 (3)¶

Stephan Vonschallen, Karim Kaufmann, Dominique Oberle, Friederike Eyssel, Theresa Schmiedel. (2026). Teachy Mini: Development and Preliminary Evaluation of a Knowledge-Based Generative Social Robot for Higher Education. arXiv. https://arxiv.org/abs/2607.22345

q2 · i? · causal · r2

In the same exploratory regressions, "Only learning preference had a significant impact on objective learning effectiveness", favoring group-preferring learners; an unadjusted moderator finding from a small sample.

"Only learning preference had a significant impact on objective learning effectiveness, such that participants who preferred learning in groups compared to learning alone had significantly greater learning gains (B= 0.667,t= 2.251,p= .042,CI= [0.027, 1.308])."

Discussion¶

Related Claims¶

• KBD configuration produced no significant between-condition differences in system acceptance, intrinsic motivation, or subjective learning effectiveness in a single tutoring session — related
• The Expert Mathematician shows a positive but not statistically significant effect on middle school math achievement (effect size 0.35, improvement index +14 percentile points) — related
• Cognitive apprenticeship instruction produced no significant differences from traditional instruction on post-treatment problem-solving, final, and standardized mathematics exams in community college technical mathematics — related
• Random assignment in the Baker (1997) trial produced pretest differences favoring the comparison curriculum, controlled by using the pretest as a covariate — related
• Students perceive a KBD-configured generative social robot tutor as more aligned with responsible tutoring behavior than an identical non-KBD control robot — related
• Exploratory ANCOVA controlling for mastery probability favored the adaptive condition, but is not interpretable as a causal treatment effect — related
• Students in the non-adaptive condition reached significantly higher average mastery probabilities than students in the adaptive condition — related
• Students' initial knowledge level (midterm scores) did not significantly influence improvement in lab scores, indicating it did not confound the feedback-source difference — related