ABSTRACT
When evaluating potentially costly or labor-intensive experiment proposals involving regression analysis, it is important to know what conditions are required for the analytic methods to have a chance of showing significant results. This chapter covers the topic of statistical power in hypothesis testing and how it relates to sample sizes, effect sizes and other criteria. Examples such as concurrent validity studies of selection instruments are used to illustrate how to calculate the required minimum sample size to achieve a target statistical power. The chapter moves through power analysis for basic hypothesis testing and then on to linear and logistic regression. Different effect size measurements sich as Cohen's d and f 2 are introduced. Example code is provided in both R and Python.
