a set of comparisons made between samples to identify significant differences among their mean values. Multiple comparisons are generally done in a post hoc manner (i.e., are unplanned) in order to keep the Type I error rate controlled at a prespecified level. Also called multiple contrasts.