Alternative procedures for testing fixed effects in repeated measures designs when assumptions are violated