Researchers often claim that the Student`s t-test are robust to violations of the assumptions of normality and homoscedasticity, negating the need to use alternative procedures, but considerable researches indicates that this happens only in a limited number of circumstances. The recent findings about this are examined here.
