Publication Bias, explained.
Publication bias occurs when availability of research results depends on their findings, distorting the evidence that readers can access.
Why it happens
One hundred independent known-variance Normal studies estimate a true zero mean. The journal reveals only absolute z scores above 1.96; opening the drawer reveals all estimates and the selection rule.
Publication bias occurs when availability of research results depends on their findings, distorting the evidence that readers can access.
Read the result
Visible signed means can remain near zero under symmetric two-sided selection. Inspect which absolute estimates became visible, not only their average sign.
A worked example
Open the drawer behind significant results
Inspect the selective journal’s visible estimates from the true-zero study generator.
Open the file drawer to reveal every one of the hundred synthetic studies.
Compare the selected extreme estimates with the complete distribution, preserving the same realised batch.
OPTIONAL DEEPER DETAILGo deeper: inside the model
Inside this model
One hundred independent known-variance Normal studies estimate a true zero mean. The journal reveals only absolute z scores above 1.96; opening the drawer reveals all estimates and the selection rule.
Where this idea is useful
A practical use
Ask what unpublished studies would change an apparent evidence base before treating visible findings as a complete sample.
A common misconception
“Publication selection must always create a positive average effect.”
Two-sided selection can retain both positive and negative extremes. Exaggerated absolute magnitude and a biased signed average are different outcomes.
What this explanation leaves out
- Two-sided selection exaggerates absolute estimates without necessarily producing a large signed average. The simulation does not measure the prevalence of bias in any actual field.
Is the hidden-study count a statistic about real journals?
No. The significance-only rule is stipulated for a teaching batch, not estimated from publication practices in any field.
What result-selection rule separates the evidence you see from the evidence collected?
Associated thinkers
Associations marked provisional are awaiting source review.
Further reading
Explore the original research or the teaching reference behind this experiment.