What problem does it solve? When several studies measure the same effect and disagree, a single pooled number without context is misleading. This Skill combines comparable study results into one precision-weighted estimate while quantifying how much the studies actually disagree and whether small studies skew the pool. ## Core Features & Use Cases - Dual pooling models: Computes both inverse-variance fixed-effect and DerSimonian–Laird random-effects estimates with 95% confidence intervals, and explains which should headline based on heterogeneity. - Heterogeneity and bias diagnostics: Reports Cochran's Q, I², and tau², plus Egger's regression test for funnel-plot asymmetry (with a low-power warning when fewer than 10 studies are pooled). - Deterministic companion tool: The stdlib-only scripts/metapool.py reads JSON or CSV study data and prints a forest table, pooled estimates, and all statistics, with a hand-verified selftest. - Use Case: A systematic review yields twelve comparable RCTs with odds ratios and confidence intervals; the Skill converts them to a common scale, pools them, reports I², and interprets Egger's test for publication bias. ## Quick Start Use the meta-analysis skill to pool these six trial effect sizes and standard errors, report fixed and random-effects estimates, and tell me how heterogeneous the studies are.
