Difference Scores versus Direct Effects in Service Quality Measurement

Abstract

Difference scores have been widely used in many areas of marketing research. There have been numerous criticisms of difference scores, and the authors add to these criticisms by testing an implicit assumption of difference scores that has not been tested in consumer research. Specifically, difference scores assume that the components have equal and opposite effects on the dependent variable. This assumption is tested in a services setting in which difference scores are still widely used and in a laboratory experiment using a product. An alternative operationalization, the Direct Effects Model, is shown to be superior. In addition, as a result of treating the antecedents as separate predictors, it is shown that performance is the much stronger predictor of satisfaction compared to expectations. Such a finding could not be uncovered using difference scores.