Assessing Performance of the UMLV Technique in Detecting and Controlling for Common Method Variance

Common method variance (CMV) is systematic error which is attributable to the instrumentation rather than the theoretical construct of interest (Bagozzi et al. 1991). CMV has been recognized as a threat to research validity by researchers in across many lines of inquiry. The unmeasured latent variable (UMLV) technique is one of the most commonly used statistical prescriptions to combat CMV. However, empirical assessment of the its performance is limited, with the only exception of the work by Chin et al. (2012). We perform Monte Carlo simulations to test the performance of the UMLV technique in detecting and controlling for CMV under models with varying number of indicators. Results from Monte Carlo simulations and power analysis suggest that this technique is ineffective in both detecting and/or controlling for CMV. Our results contribute to IS research by informing researchers about their choice of statistical technique when accounting the influence of CMV.