Statistics for the Clinician I: Categorical Variables

Abstract

Categorical variables are an integral part of clinical research. This article introduces their uses and most common analyses for clinicians seeking additional statistics exposure to more critically engage with literature and refine their own research endeavors. We describe and demonstrate the two most common tests of association for categorical variables: chi-square and Fisher’s exact tests, along with their underlying logic, result interpretations, and relative strengths and weaknesses. We also introduce and explain two of the most common measurements of effect size in analyses of categorical outcomes: relative risk (RR) and odds ratio (OR).