Question

Why might you want to use a permutation test? You can't assume that the data follows a normal distribution You can't assume that the data is result of random sampling O You can't assume that the data has equal variance O All of the above [1 pt(s)] Submit Answer Tries 0/1 How many ways are there to assign 22 subjects into two equally sized groups? [1 pt(s)] Tries 0/3 Submit Answer A study was conducted comparing the test scores between native English speakers and native German speakers participating in a first year French course on a French test. Each group had 9 members. The reasearchers are interested in whether English speakers are on average better at French than the German speakers. It was found that 36808 arrangements out of the total possible arrangements resulted in higher test scores than what was actually observed. What is the p-value of the permutation test? [1 pt(s)] Tries 0/3 Submit Answer

Answer 1

1) We can not assume that the data follow a normal distribution because no information regarding the population distribution is given and no sample has been taken. The data is not a result of random sampling because no random sampling has been done here. No information is given regarding the variance. Hence, all the reasons are correct. Therefore, option (iv) is correct. 2) The number of ways of assigning 22 subjects into two equally sized groups: 22! 2211 2! 11 11 22! 22! 11! x2x1 11.x (22-11)!) 11!x (11-11)! 22! 705432 ,593,350,922, 240,000

3) As given in the question: Each group of English and German speakers had 9 members in their group. Therefore, the total number of members = 9+9= 18 The number of arrangements out of the total possible arrangements resulted in higher test scores = 36,808 36808 Total possible arrangements 36808 p-value = 36808 18! 9!x (18-9)! 36808 48620 0.7571