Multivariate statistics for multiple outcomes are used when attempting to control for increased Type I error rates, or experimentwise error rates, when testing multiple hypotheses concurrently. Furthermore, the multivariate and bivariate associations between predictor, confounding, and outcome variables can be assessed and understood within a theoretical or conceptual framework when using multivariate statistics for multiple outcomes. The statistical assumptions of multivariate statistics for multiple outcomes such as MANOVA and MANCOVA can be hard to meet but the findings yielded from these analyses are powerful.
Researchers are comparing independent groups or levels of a categorical variable on several continuous outcomes at the same time.
Researchers are adjusting values of several outcomes when comparing independent groups or levels of a categorical variable at the same time.
