Statistics Done Wrong: The Woefully Complete Guide

by Alex Reinhart

When you need to measure an effect with precision, rather than simply testing for significance, use assurance instead of power: design your experiment to measure the hypothesized effect to your desired level of precision.