ABSTRACT
The Valence Instrumentality Expectancy (VIE) model (Vroom, 1964) has been one of the most inuential theories in work motivation. However, the reviews that appeared over the years (Mitchell, 1974, 1982; Pritchard & Campbell, 1976; Schwab, Olian-Gottlieb, & Heneman, 1979; Wanous, Keon, & Latack, 1983) raised several questions concerning its validity. In a meta-analysis (Van Eerde & Thierry, 1996) we gave an overview of effect sizes of nonexperimental research. The results of the meta-analysis show different average correlations across the criterion measure categories, whereas the average correlations referring to models, such as the Expectancy-valence (EV) or VIE model, did not differ from those referring to the single constructs valence, instrumentality, and expectancy, or across different operationalizations. Within-subjects correlations were not found to be clearly superior to between-subjects correlations, although this may have been partly due to the fact that relatively few within-subjects correlations were available. We concluded that the issue of which criterion variable is to be predicted is more important than which form and operationalization of expectancy theory is used.
