| Abstract: | McIntyre and Farr (1979), Hanser, Mendel, and Wolins (1979), and Lissitz, Mendoza, Huberty, and Markos (1979) comment on the repeated measures Analysis of Variance design suggested by Arvey and Mossholder (1977) to detect job differences and similarities. These authors propose alternative procedures to determine job differences. We, in turn, point out here that the problems specified by these critics may not be as severe as they indicate, and that some problems are even nonexistent. Moreover, the alternative solutions they suggest also have their own limitations. Finally, we propose an additional procedure–a multivariate approach to repeated measures data–which might be useful in the context of detecting job differences. It appears as if there are assumptions and limitations to both the univariate and multivariate approaches to the problem; these assumptions and limitations are delineated more precisely in the present paper. |
