Question

Question 15 (1 point) A study of the impact of caffeine consumption on reaction time was designed to correct for the impact of subjects' prior sleep deprivation by dividing the 24 subjects into 12 pairs on the basis of the average hours of sleep they had had for the previous 5 nights That is, the two with the highest average sleep were a pair, then the two with the next highest average sleep, etc. One randomly selected member of each pair was allowed to drink two cups of caffeinated coffee, and the other drank two cups of decaf. Each subject's performance on a standard reaction-time test was recorded. Which of the following is the correct check of the "Normal/Large Sample" condition for this significance test? ttp://aths.ankabut.ac.ae/d2I/lms/quizzing/user/attempt/quiz at...&impcf &pg-1&qi-309513&ai 2494894&dnb 0&cfql 0&d21 body.type-1 Page 8 of uiz ATHS G12 AP Statistics-ll - Applied Technology High School 5/29/19, 5:48 Confirm (graphically) that the scores of the caffeine drinkers and the scores of the decaf drinks in e study could have come from Normal distributions. Confirm that at least one of the sets of individual scores (from the caffeine or decaf group) and the differences in scores within each pair could have come from Normal distributions Confirm that the differences in scores within each pair of subjects could have come from a ormal distribution Because the central limit theorem applies to this situation, it is not necessary to check any of the samples for Normality. Confirm (graphically) that the scores of the caffeine drinkers and the scores of the decaf drinks in the study, and the diffèerences in scores within each pair of subjects could have come from Normal distributions

Answer 1

In a paired test where we take paired set of observations, the normality condition should be valid only on the differences in the values within each pair of subjects and not on the original values. Hence the correct option is C) Confirm that the differences in scores within each pair of subjects could have come from a normal distribution