Bayesian Structural Equation Modeling of Reliability: A New Approach to Reliability Estimation

• In Event: 60.031 - Reliability: New Approaches

Abstract

Accurate reliability estimates are rarely available. This study proposed a new reliability index: Bayesian Structural Equation Modeling of Reliability (BSEMR) for accurate reliability estimation. This reliability index may overcome some shortcomings of the previous reliability coefficients. Particularly, it does not require the assumption of unidemensionality and tau-equivalence, and it performs well in small sample sizes. To evaluate this, the proposed reliability estimation method was compared with the other popular or recommended reliability coefficients by comparing the performances of these reliability coefficients in a Monte Carlo study. The results indicated that coefficient BSEMR outperformed the other reliability coefficient in terms of estimation accuracy under different simulation conditions.

Authors

• Wei Tang, University of Alberta

• Qi Guo, University of Alberta

• Ying Cui, University of Alberta