The evaluation used a school-level cluster-randomized quasi-experiment (N = 50) that was powered only to detect large effects (d = 0.81)¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r1–r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Fifty pre-service teachers in 10 schools were assigned at the school level by simple lottery to experimental (5 schools, n = 25) or control (5 schools, n = 25) conditions within a quasi-experimental embedded mixed-methods design. → Ismael I 2026 q2 i? With N = 50, the minimum detectable effect size was Cohen's d = 0.81, so the study was adequately powered only for large effects and small-to-moderate effects may have gone undetected. → Ismael I 2026 (2)

Evidence¶

Ismael I 2026¶

Ismael I, Luo X and Li S (2026). An AI-supported E-mentoring model to develop EFL pre-service teachers' self-efficacy and emotional intelligence. Front. Psychol. 17:1853510. https://doi.org/10.3389/fpsyg.2026.1853510

q2 · i? · causal · r2

Methodology section (§4.2) describing the study's assignment procedure: an embedded mixed-methods quasi-experiment in which randomization occurred at the school level, with all pre-service teachers in a given school in the same condition. The article reports the assignment 'using a simple lottery method' with 5 schools and n = 25 per condition.

"the researcher randomly assigned the schools, as complete units, to the experimental condition (5 schools; n = 25 pre-service teachers) or the control condition (5 schools; n = 25 pre-service teachers) using a simple lottery method."

Ismael I 2026 (2)¶

Ismael I, Luo X and Li S (2026). An AI-supported E-mentoring model to develop EFL pre-service teachers' self-efficacy and emotional intelligence. Front. Psychol. 17:1853510. https://doi.org/10.3389/fpsyg.2026.1853510

q2 · i? · causal · r1

Sensitivity analysis reported in §4.2 using G*Power 3.1 (two-tailed independent-samples t-test, α = 0.05, power = 0.80). The article states the minimum detectable effect was "Cohen's d = 0.81 (a large effect)" because sample size was determined by availability rather than an a priori power analysis.

"the minimum detectable effect size for N = 50 was Cohen's d = 0.81 (a large effect). The study was therefore adequately powered to detect large intervention effects, whereas small-to-moderate effects may have gone undetected"

Discussion¶

Related Claims¶

• Clustering effects in education random-assignment trials vary by design but are typically large, requiring large school samples — a broader claim this one bears on
• Large school samples are required to achieve appropriate precision standards in clustered education experiments — a broader claim this one bears on
• Many education evaluations have sufficient power to detect precise impacts only for relatively large subgroups of sites — related
• Most education impact studies can rigorously address only broad questions due to power constraints — related
• Required numbers of schools differ across school, classroom, and student random-assignment designs — related
• The qualifying InsideTrack© study was an RCT across 17 lotteries in eight universities, with only seven 50/50 lotteries meeting WWC standards — related
• Sample size formulas for clustered designs with school- or teacher-level random assignment are derived using generalized estimating equation methods — related
• Edpuzzle qualifies for a Promising evidence rating based on one school-level randomized study — related
• Randomization does not ensure baseline equivalence in small-sample motor learning studies, so unbalanced pre-tests can overstate random-practice benefits — related