The Practical Utility of Importance Measures in Assessing the Relative Importance of Work-Related Perceptions and Organizational Characteristics on Work-Related Outcomes

The utility of dominance analysis and other importance indices is the subject of much debate in determining the relative importance of predictors in multiple regression. The goal in conducting this research was to bring an applied perspective to this issue by comparing the conclusions one would draw regarding predictors’ relative importance, when using various indices of importance with real-world data sets. The overall results indicate that researchers would reach only minor differences in their conclusions when using dominance analysis or other importance indices as compared to simply examining the traditional standardized beta weights or squared semipartial correlations. The lack of differences in conclusions drawn is particularly apparent when comparing the rank ordering of the predictor importance produced by the different indices.