Relative Importance Analysis: A Useful Supplement to Regression Analysis Scott Tonidandel • James M. LeBreton Published online: 7 January 2011 Ó Springer Science+Business Media, LLC 2011 Abstract This article advocates for the wider use of rela- tive importance indices as a supplement to multiple regres- sion analyses. The goal of such analyses is to partition explained variance among multiple predictors to better understand the role played by each predictor in a regression equation. Unfortunately, when predictors are correlated, typically relied upon metrics are flawed indicators of vari- able importance. To that end, we highlight the key benefits of two relative importance analyses, dominance analysis and relative weight analysis, over estimates produced by multi- ple regression analysis. We also describe numerous situa- tions where relative importance weights should be used, while simultaneously cautioning readers about the limita- tions and misconceptions regarding the use of these weights. Finally, we present step-by-step recommendations for researchers interested in incorporating these analyses in their own work and point them to available web resources to assist them in producing these weights. Keywords Relative importance Á Predictor importance Á Relative weight analysis Á Dominance analysis Á Multiple regression Relative Importance: A Useful Supplement to Regression Analyses Multiple regression is perhaps the most frequently used statistical tool for the analysis of data in the organizational sciences. The information provided by such analyses is particularly useful for addressing issues related to predic- tion such as identifying a set of predictors that will maxi- mize the amount of variance explained in the criterion. However, most researchers and practitioners are simulta- neously interested in multiple regression for theory testing or explanation purposes. Here, the question of interest becomes understanding the extent to which each variable drives the prediction. Essentially, one wishes to understand the contribution each predictor makes towards explaining variance in the criterion. Past research has documented how indices commonly produced by multiple regression analyses fail to appropriately partition variance to the various predictors when they are correlated (Darlington 1968). In response, two alternative approaches, dominance analysis (Budescu 1993) and relative weight analysis (Fabbris 1980; Johnson 2000), have been developed that allow for more accurate variance partitioning among cor- related predictors. The purpose of this article is to call greater attention to these estimates of relative importance by describing how they can be a useful supplement to traditional regression analysis. In what follows, we will discuss the types of information these analyses provide, describe how this information differs from more commonly used indices of importance, present some of the limitations of these analyses, and provide some brief examples of their potential uses. We will not focus on the intricacies of performing dominance analysis or relative weight analysis. Instead, we would refer interested parties to the many papers referenced throughout this article that provide detailed information concerning computing these impor- tance weights. Our goal is to provide a practical, user- friendly guide for those wanting to supplement their regression analysis with relative importance analysis. Information is also provided regarding where one can S. Tonidandel (&) Davidson College, Box 7061, Davidson, NC 28035, USA e-mail: sctonidandel@davidson.edu J. M. LeBreton Purdue University, West Lafayette, IN, USA 123 J Bus Psychol (2011) 26:1–9 DOI 10.1007/s10869-010-9204-3