Logo image
Interaction Effects May Indeed be Artifacts of Common Method Variance
Abstract   Open access

Interaction Effects May Indeed be Artifacts of Common Method Variance

Arturs T. Kalnins
Academy of Management Annual Meeting Proceedings, Vol.2026(1)
07/2026
DOI: 10.5465/AMPROC.2026.20954abstract
url
https://doi.org/10.5465/AMPROC.2026.20954abstractView
Published (Version of record) Open Access

Abstract

Conventional wisdom in organizational sciences holds that effects of quadratic and interaction terms cannot be artifacts of common method variance (CMV) in ordinary least squares regression. CMV is believed only to attenuate such effects. As a result, researchers often conclude that they can ignore CMV when testing moderated relationships. In this paper, we disprove the universality of the conventional wisdom by deriving counterexamples that reveal overlooked boundary conditions. A common scenario lies outside these boundaries: when a regression includes both a quadratic and an interaction term involving the same primary term, and at least one of them influences the dependent variable (DV), one of the estimated effects may be artefactual. We conclude that the conventional wisdom holds only if (1) the interaction term affects the DV and researchers include it in their regressions, but quadratic terms do not and studies exclude them, and (2) neither term affects the DV, regardless of inclusion. We recommend that researchers apply standard CMV precautions in all moderated regression analyses and present multiple model specifications to assess robustness.

Details

Metrics

1 Record Views
Logo image