Testing Specific Hypotheses: a Tutorial on Contrast Analyses Using R

The variance (ANOVA) analyses are among the most widely used statistics in psychology. Yet its omnibus approach poses limitations when testing specific hypotheses, thus, limiting its effectiveness. Contrast analyses offer a solution by allowing direct testing of hypotheses, thereby avoiding unnecessary multiple comparisons and a decrease in statistical power. After emphasizing the advantages of contrast analyses over omnibus tests, this paper will focus on within-subjects and mixed designs, often less discussed, especially in open-source software such as R. This tutorial aims to detail the procedures for conducting these analyses and provides R functions to facilitate their implementation.