IBM Support

Unexpected sign in regression coefficients

Troubleshooting


Problem

I have run a linear regression with 2 predictors and noticed two surprising results. 1. Although the correlations among the 2 predictors (X and W, say) and the dependent variable are positive, the sign of the regression coefficient for X is negative and its standardized coefficient is much larger than the standardized coefficient when X is the only predictor. 2. The squared multiple correlation (R^2) for the two-predictor models is larger than the sum of the two R^2s for the simple regressions with each of these predictors as the sole predictor. The presence of W seems to enhance the predictive power of X. I thought that the presence of 2 correlated predictors would reduce the predictive value of each of them.

[{"Product":{"code":"SSLVMB","label":"IBM SPSS Statistics"},"Business Unit":{"code":"BU048","label":"IBM Software"},"Component":"Not Applicable","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"Not Applicable","Edition":"","Line of Business":{"code":"LOB76","label":"Data Platform"}}]

Log InLog in to view more of this document

This document has the abstract of a technical article that is available to authorized users once you have logged on. Please use Log in button above to access the full document. After log in, if you do not have the right authorization for this document, there will be instructions on what to do next.

Historical Number

42165

Document Information

Modified date:
16 April 2020

UID

swg21479993