In Exercises 24, 26, 28, and 30, we considered data on articles in the NEJM. The original study listed 23 different statistics methods. (The list read: t-tests, contingency tables, linear regression, . . . .) Why would it not be appropriate to use a chi-square test on the 23 x 3 table with a row for each method?

Data From Exercise 26

The table in Exercise 24  shows whether NEJM medical articles during various time periods  included statistics or not. We’re planning to do a chi-square test.