"Effects of Artificial Intelligence and Virtual Reality Assisted Instruction in Rehabilitation Education: A Meta-Analysis" (analysis code and extracted data)

It contains: (1) study-level extracted data and effect sizes (Hedges' g / OR) for 29 included studies (13 on artificial intelligence and 16 on virtual reality, comprising 25 randomized controlled trials and 4 quasi-experimental studies) across four outcomes (theoretical knowledge, practical skills, higher-order abilities, and teaching satisfaction); (2) complete Python analysis code implementing random-effects pooling, subgroup analysis with Bonferroni correction, Knapp-Hartung adjustment, sensitivity analyses (S1–S3), outlier and leave-one-out analyses, Egger's and Begg's tests with the trim-and-fill method, meta-regression, and Peto cross-validation, yielding results consistent with the R metafor package; (3) environment specifications and a reproduction guide.

Running the code reproduces all pooled effect estimates, forest plots, and funnel plots reported in the manuscript and its supplementary materials.