Fixed & random effects
Inverse-variance pooling with DerSimonian-Laird random effects — the standard your reviewers already expect from RevMan.
META-ANALYSIS SOFTWARE
No exporting to a separate stats package. Pool effect sizes with the same models Cochrane reviews use — fixed or random effects, Mantel-Haenszel, Peto — and get forest plots, funnel plots, and heterogeneity statistics automatically.

3
Pooling Methods
10+
Statistical Outputs
Q, I², τ²
Heterogeneity Stats
Inverse-variance pooling with DerSimonian-Laird random effects — the standard your reviewers already expect from RevMan.
Dichotomous pooling methods that stay robust with sparse events, without needing continuity correction on zero-cell studies.
Interactive, exportable plots — including RevMan-style grouped subtotals when your studies split into subgroups, and log-scale funnel plots for ratio measures.
Test for subgroup differences with the same Q-statistic-based chi-square test the Cochrane Handbook specifies — not an approximation.
Leave-one-out re-pooling shows whether any single study is driving your overall effect, before a reviewer asks the question for you.
Per-study traffic light plus an aggregate risk-of-bias graph — the two figures every Cochrane-style review needs, generated from your extraction data.
The statistical methods are the ones Cochrane reviews are built on — the difference is that your studies, extracted effect sizes, and pooled results live in the same place as your screening decisions, so nothing has to be manually re-typed between screening and analysis.
See the full systematic review workflow →FAQ
Inverse-variance (fixed and DerSimonian-Laird random effects), Mantel-Haenszel, and Peto's method for dichotomous outcomes — plus mean difference and Hedges' g standardized mean difference for continuous outcomes.
Yes — tag studies with a subgroup label at extraction, and any pooled run with two or more subgroups automatically shows subtotal diamonds and a test for subgroup differences.
Yes — GWAS variant aggregation and RNA-seq gene-level synthesis are supported alongside clinical meta-analysis.
Yes, with no subscription required.
Cochran's Q, I², and τ² are computed for every pooled run, with an automatic low/moderate/high interpretation alongside the numbers.
Extract a handful of studies and run a pooled analysis to see the forest plot, heterogeneity statistics, and interpretation generated automatically.