violation of assumptions
a situation in which the theoretical assumptions associated with a particular statistical or experimental procedure are not fulfilled. Common assumptions for statistical tests include normality of the distribution, equal variances within the cells of treatment levels, homogeneity of variance, and linearity. Research designs also need to meet certain assumptions, such as random sampling and random assignment, sample representativeness, and the like. Because violation of assumptions introduces bias, the validity of assumptions must be confirmed prior to data analysis to ensure that the methods and strategies chosen are appropriate and will yield valid results.