Reply
Sun 6 Sep, 2015 11:31 am
Pre-election polls have differed in their predictions. Poll A suggests that your candidate will get 50 percent of the vote, and poll B that the candidate will get 55 percent. You wish to take a third poll to test whether A or B was correct. Taking 50 percent as the null hypothesis, how big a sample size would you need to achieve a power of test equal to 0.90 with a significance level of 0.10.