Exploring Joint Variance Between Independent Variables and a Criterion

Abstract

Current methods used in the analysis and interpretation of behavioral data tend to ignore a potentially important explanatory component. That component is the joint variance shared between predictors in explaining variance in the outcome variable. The authors provide an example of joint variance and how it could be interpreted. The authors believe ignoring this component has inhibited development of explanatory theories. The authors discuss a method developed by Mood for calculating joint explanatory variance. This method was initially developed to better interpret the unique effects of predictors on a criterion but can also be used to gain a better understanding of joint effects as well. They reanalyze published data to demonstrate the contribution of this approach in analyzing and interpreting behavioral data. They also provide a method for calculating the significance of joint variance components.