Publication Bias in 3D | mysimulator.uk

Study Field

Studies per second 18

Significance threshold α 0.05

True anomaly rate 2%

Literature so far

Studies run

File drawer

Published rate

–

True rate

–

False discovery rate

–

Published count

How it works

Each glowing sphere is one completed study drifting up out of the spawn cloud. A small slice (the true anomaly rate) are looking at something genuinely real and, with typical statistical power, usually come back significant. Every other sphere is testing pure noise — and by construction a noise study still reads "significant" (p < α) purely by chance at a rate of exactly α, no matter how many are run.

P(false positive per null study) = α
E[false positives] = α · (1 − true rate) · N
E[true positives]  = power · true rate · N
apparent rate = published / N  ≈  α(1−p) + power·p

Spheres that clear the bar rise to the glowing published shelf; everything else drops into the dim file drawer bin below, out of sight. Because false positives scale with the total number of studies while true positives scale only with the (small) true rate, running more studies fills the shelf mostly with noise dressed up as discovery — the same multiple-comparisons dynamic behind file-drawer bias and the reproducibility crisis in real meta-science.

Significance threshold α — the p-value cutoff; lower it and fewer noise studies clear the bar, but never to zero.
True anomaly rate — how often a study is actually looking at something real.
Publish only significant — toggle off and every sphere, null or not, rises to a neutral open shelf instead of a bin — watch the published rate collapse toward the true rate.

True anomaly, published False positive, published Null, filed away

⚠ Couldn't load the 3D engineThree.js failed to load from the CDN. Check your connection and reload.