Interpretation of an ANCOVA is more problematic when: The IV has more than two levels Assignment to groups is not random The covariate is a random variable The covariate is a pre-treatment measure of