Violating Assumptions in Parametric Statistical Testing