Clinician's Guide to Understanding Effect Size, Alpha Level, Power, and Sample Size

Abstract

Effect size, α level, power, and sample size are misunderstood concepts that play a major role in the design and interpretation of studies. Effect size represents the magnitude of a change in an outcome or the strength of a relationship. Often, the effect size may be more important than just relying on the α level when interpreting results from a study because it informs a researcher of the actual magnitude of the difference or relationship. Confidence intervals can also assist in making this assessment. Power represents the probability of rejecting the null hypothesis when it is false. Effect size, power, and α level are all important in the calculation of sample size needed to conduct a study. Calculating the sample size a priori and including the required number of participants is essential. Studies with small sample sizes, relative to the needed number provided from a power analysis, may lead to false negative results. Studies with grossly large sample sizes may yield statistically significant findings with small effect sizes that may not be clinically relevant. It is beneficial when all four components are clearly presented in analytic studies. Failure to include these elements may limit the ability of other researchers to replicate the study's findings and lead to difficulty when interpreting the study's results.

Keywords: effect size; power analysis; research design; sample size; statistics; α level.