Assessing computational reproducibility makes sense. I was recently asked to review the statistical analysis in a revised paper. Most analyses were reproducible. However, going through the R code showed that the main hypothesis test relied on a model ignoring dependencies of repeated measures.