Matheus Gallas-Lopes @matheusglls.bsky.social · 31/07/2026But replicating the review process was much harder. None of the three full replications identified exactly the same set of included studies as the original review. 100
Matheus Gallas-Lopes @matheusglls.bsky.social · 31/07/2026Teams also found differences in data processing, meta-analysis and risk-of-bias judgments. Missing or ambiguous methodological details often meant that replicators had to make their own decisions. 110
Matheus Gallas-Lopes @matheusglls.bsky.social · 31/07/2026Access was another issue: not all raw quantitative data were available, and none of the original reviews shared their analysis code. Documenting decisions and sharing the full analytical workflow could make future reviews easier to verify and reuse. 100
