TESTS AND CONTRASTS FOR COMPARING CHANGE PARAMETERS FOR A MULTIPLE SAMPLE MCNEMAR DATA MODEL

The McNemar test of dichotomous change is extended to the case in which a researcher has two or more independent groups that are to be compared for identical change. Planned and post hoc comparison methods using contrasts for testing identity of change parameters are presented. The proposed procedures are illustrated for an attitude change study and are compared to other tests designed to examine no change across two or more groups. The methods illustrated permit a direct identification of significant pairwise differences, while the competing procedures are shown to terminate in an omnibus test of statistical significance. 1979 The British Psychological Society