The Multifaceted Nature of Measurement Artifacts and Its Implications for Estimating Construct-Level Relationships

Abstract

Measurement artifacts, including measurement errors and scale-specific factors, distort observed correlations between measures of psychological and organizational constructs. The authors discuss two alternative procedures, one using the generalized coefficient of equivalence and stability (GCES) and one based on structural equation modeling, to correct for the biasing effect of measurement artifacts in order to estimate construct-level relationships. Assumptions underlying the procedures are discussed and the degrees of biases resulting from violating the assumptions are examined by means of Monte Carlo simulation. They then propose an approach using cumulative knowledge in the literature about properties of measures of a construct to estimate the GCES. That approach can allow researchers to estimate relationships between constructs in most research situations. The authors apply the approach to estimate the GCES for overall job satisfaction, an important organizational construct.