Hello Prof. Andy,
I have a question relating to paired and unpaired t-tests. I was interested in finding out gender differences in certain attitudes. I collected data from a sample of 111 couples and conducted a paired sample t-test to find out whether the husband and wife differed significantly in the variables. I opted for paired t-tests instead of independent sample t-tests because I believed that the males and females that I have included in my study cannot be considered as two randomly selected groups. Moreover, I found that there was significant correlation in the scores obtained by the husband-wife pairs. However, I was criticized by some experts who pointed out that the husband and wife are independent people who are only related by law and hence I should have done an independent sample t-test.
Which approach is correct here?