While the latest case if more about very LARGE effects, it may reflect potential publication bias against null results.

Since 2000s, the share of claims that are null have declined in almost all fields.
Post image

Comments