Statistics Done Wrong: The Woefully Complete Guide

by Alex Reinhart

Suppose, for example, that the effect size is 0.8 (on some arbitrary scale), but the review was composed of many small studies that each had a power of 0.2.