Suppressor effects in self-enhancement research: A critical comment on condition-based regression analysis

Self-enhancement conceived as a positive relation between happiness (H) and the difference S-O (self-rating S minus objective measure O) is inherently confounded with a positive self-view account: S alone may explain the positive relation to H. Condition-based regression analysis (CRA; Humberg et al., 2018a, 2018b) promises a solution. CRA assumes that opposite predictor weights (βS > 0; βO < 0) in regression of H on S and O rule out that H depends on S alone. However, despite the truism that two significant regression weights imply that both S and O contribute to the prediction model, they cannot rule out a positive self-view account. If only S shares variance with H, O may improve the prediction indirectly, by suppressing unpredictive S-variance in the prediction model. Granting S-O-redundancy, a classical suppressor effect (Conger & Jackson, 1972) results in a negative regression weight for O (binding S-variance that is unpredictive of H). Thus, the regression pattern that CRA presumes to rule out a positive self-view account indeed follows necessarily from a suppressor effect entailed in a positive self-view account. Computer simulations illustrate and corroborate this critique. (PsycInfo Database Record (c) 2021 APA, all rights reserved).