Method Effects, Measurement Error, and Substantive Conclusions

Abstract

Common method variance is routinely viewed as a pervasive problem in organizational research, one that undermines good science and biases empirical conclusions. The authors review research that has used multitrait-multimethod (MTMM) designs to estimate the magnitude of common method variance in organizational research. The results of this study show that method variance accounts for less variance (18%) than has been suggested by previous reviews. The authors also consider simultaneously the attenuating effect of measurement error with the inflationary effect of common method variance on observed relationships. Results indicate that although common method variance does have an inflationary effect on observed relationships, this effect is almost completely offset by the attenuating effect of measurement error.