We want to test whether tutoring helps improve exam scores by at least ten points on average. We give a class a midterm exam, then get them tutoring in the material, and give a similar exam one month later. We do a matched pairs t-test on the mean gain in score, with HO :u=10 and Ha :> 10, looking evidence at the 0.05 level. We get a P-value of 0.061. We throw out the scores of several students whose scores on the second exam were lower than on the first exam, and get a P -value of 0.020. Which of the following is true. Since we can reject the null hypothesis in the second case, we say the study is inconclusive. Throwing out the outliers gave a more accurate P-value, since it was significant at the 0.05 level after doing SO. O Since we got a P-value of 0.061 for the first test, the null hypothesis is probably true. We failed to account for the lurking variable that, having taken a similar exam once already, scores might improve without tutoring. Since we got a P-value of 0.020 for the second test, the null hypothesis is definitely false. O A matched pairs test was inappropriate; we should have used a two-sample t-test