A series of 25 wet/new concrete samples were mixed in the lab and then each split in half (like twins!), with the first set of "twins" receiving admixture A and the second set of 25 sub-samples receiving admixture B. Once hardened and strength tested, compressive strength at failure of the first set averaged 50.2 ksi, and the second set averaged 53.4 ksi. The sample standard deviation of the first set of strengths was 8.2 ksi, the second set's S value was 9.8 ksi, and the S value for the 25 differences in strengths was 3.3 ksi. What test statistic will you use to determine whether you can accept a hypothesis that Admixture B creates a stronger concrete than Admixture A?