20. In Problem 14 in Chapter 11, we described a study
showing that students are likely to improve their test
scores if they go back and change answers after
reconsidering some of the questions on the exam
(Johnston, 1975). In the study, one group of students
was encouraged to reconsider each question and
to change answers whenever they felt it was appropri-
ate. The students were asked to record their original
answers as well as the changes. For each student, the
exam was graded based on the original answers and on
the changed answers. For a group of n = 40 students,
suppose that 26 had higher scores for the changed-
answer version and only 14 had higher scores for
the original-answer version. Is this result significantly
different from chance? Use a two-tailed test with
α = .01.