A meta-analysis pools data from several studies of the same question to get a single, more precise estimate. It usually follows a systematic review that finds the studies, and reports a pooled effect size with a confidence interval.
Psychedelic examples show both its use and its traps. A 2012 analysis of six LSD trials for alcoholism, with 536 participants, reported an odds ratio of 1.96 in favour of LSD (Krebs and Johansen, 2012). A 2024 BMJ meta-analysis of psilocybin for depression was later corrected, after errors in calculating effect sizes and data from the same study counted more than once (BMJ correction, 2025). The correction describes the research as preliminary (BMJ correction, 2025).
A meta-analysis can’t remove bias shared by every study it includes, such as functional unblinding. One 2025 meta-analysis compared control groups across psilocybin, SSRI and esketamine trials, and its authors note such cross-trial comparisons are indirect (Hieronymus et al., 2025).
Sources
- Krebs and Johansen, Journal of Psychopharmacology, 2012
- BMJ correction to Metaxa and Clarke, 2025
- Hieronymus et al., JAMA Network Open, 2025
Last checked .