Revisiting an âurban legendâ: an experimental assessment of common method varianceâs impact on relationships in self-reported data
Despite the ubiquity of self-reported data in social science and public administration research, widespread concerns persist regarding common method variance (CMV) and its potential to distort observed correlations. In this article, we estimate CMVâs biasing effects through five preregistered studies (including eight survey experiments) with UK and Chinese civil servants (N = 3,159), focusing on the relationship between public service motivation (PSM) and job performanceâa proposition of PSM theory often subject to CMV concerns. Our findings indicate that procedures widely advocated by methodological scholars to mitigate CMV did not substantially attenuate the PSM-performance relationship. A single-paper meta-analysis integrating these survey experiments reinforced this result, revealing a negligible overall moderating effect (mean effect size = -0.018, 95% CI [-0.08, 0.04]). Our results offer insights into the quality of self-reported measures, call into question the notion that CMV uniformly biases self-reported correlations, and strengthen the PSM theory by providing evidence for the validity of its core theoretical relationships against the CMVâs biasing effect.