Confounding in the Analysis of Variance

Unintentional and intentional confounding in the analysis of variance is discussed. Some different kinds of unintentional confounding that can occur in between-subject designs, within-subject designs, and mixed designs are considered. The usefulness of losing information to conserve resources by intentionally confounding variables gives rise to split-plot, confounded factorial, and fractional replication designs. An example from infancy research is provided. Keywords: analysis of variance; confounded variables; confounding, aliasing