The t-Test and Assumption Violations

This applet allows you to dynamically explore particular simulation conditions - sample size, unequal sample sizes, heterogeneity of variance - and their influence on the Type I error and power rates of the independent samples t-test. Results shown are the proportion of samples that reject the null hypothesis (also called the empirical detection rate).
Try to find a combination of design factors that demonstrate the greatest disparity between the uncorrected independent samples t-test and the ones that have had Welch correction applied!