Suppose you are going to test 3 treatment versus control. If you want your family wise error rate to be no more than 5%, what significance level do you need to set for each of the comparisons of treatment vs. control. What we are asking here is if you want the probability of having at least one Type 1 error -- where you would erroneously conclude there was a difference even if the null of no difference between any treatment and control were true -- to be no more than 5%, what p-value do you need to use as your threshold for declaring "statistical significance" for each of the individual tests? [Note: this was covered in the videos]