The Tukey HSD test looks at the probability of making one or more Type I errors (rejecting a true null hypothesis) in comparing population means and is sensitive to treatment errors. Could we use 2-sample t-tests? No. Multiple t-tests increase the probability of making a Type I error.
I hope this helps.
What is the purpose of a tukey test?
Why can't we compare all the posible pairs of group means using the two-sample t-test?
1 answer
0 answers