Abstract

For decades, difference scores have been frequently used in person-environment fit research. Despite their widespread use, difference scores are prone to numerous methodological problems. Difference scores are inherently ambiguous because they combine measures of conceptually distinct constructs into a single score. They also impose constraints rarely tested empirically and reduce an inherently three dimensional relationship between their component measures and the outcome to two dimensions. Problems with difference scores can be avoided by using polynomial regression analysis and response surface methodology. The purpose of this study was to discuss the methodological problems of difference scores, introduce the polynomial regression analysis and the response surface methodology as the alternative to difference scores, and demonstrate an empirical study on person-person(leader-member) fit by using these new approaches. We investigated the influence of leader-member fit in Big five personality factors on member's satisfaction and trust toward the leader. Questionnaires were administered to 294 job incumbents to obtain (heir own personality rating, their leaders' personality rating, and their satisfaction and trust toward leaders. The result generally indicated the more correspondent the personality of leader and member in Big 5 personality factors were, the more satisfaction and trust the member showed toward the leader. Finally, the meaningfulness and usability of these new approaches in congruence research area and future research directions were discussed.