MOMUS · the peer · critique that earns its keep · kept by MOM

THE META-ANALYSIS ◧ 2D · ◍ 3D · ◆ 4D · ◐ shadow · 👶 TAP

One study is an anecdote with error bars. Pool many and the noise averages down — but not equally: a precise study should count for more than a sloppy one. Slide (or tap) to pour the studies into one inverse-variance-weighted estimate and watch the confidence interval tighten.

◆ LIT▲ AMBER
◧ THE MEASURE · 2D
◍ THE POOL · 3D · many studies → one
◆ THE FOURTH · 4D · a tesseract turns
◐ THE SHADOW · one dimension down
👶 THE TODDLER CORNER — tap to bring them together
studies
7
pooled effect
95% CI half-width
most weight

◆ LIT — exact / checkable

Fixed-effect meta-analysis, computed live: each study i has an effect yᵢ and standard error SEᵢ; its weight is wᵢ = 1/SEᵢ². The pooled estimate is the inverse-variance weighted mean μ = Σwᵢyᵢ / Σwᵢ, with SE(μ) = 1/√Σwᵢ and 95% CI = μ ± 1.96·SE(μ). The precise (small-SE) study genuinely dominates the pool — a fail-loud self-check throws unless the pooled estimate sits nearer the lowest-SE study than the plain average does.

▲ AMBER — the figure

The 7 studies are a fixed synthetic set (seeded yᵢ, SEᵢ); real meta-analysis adds heterogeneity (random-effects τ², I²) and worries about which studies are comparable at all. The inverse-variance pooling arithmetic shown is exact.

MOMUS: praise is cheap; a fault named precisely is the gift.  — MOM
David Lee Wise / ROOT0 / TriPod LLC  ·  kept by MOM, with AVAN