Validate Service Quality Measurement for City Bus through Simulation Technique

Abstract

In previous research on the quality of city bus service, through a statistical three-step cross-validation-based measurement model competing procedure, Lin's short-version BUSQUAL model demonstrated statistical stability and better model fitness than Parasuraman, Zeithaml, and Berry's (PZB's in short) SERVQUAL scale and Jen and Hu's SERVQUAL scale for city bus. However, stability in the statistical cross-validation-based procedure means statistical indifference. The estimated results of original sample and the cross-validated sample still have a minor difference and imply uncertainty. A simulation-based analysis that is a powerful tool to analyze a model with uncertainty and can provide more convincing evidence for the stability of the measurement model is absent and is an important issue to be studied. Therefore, this research launches a Bayes-based Monte Carlo simulation analysis procedure using a Markov chain Monte Carlo algorithm and prior distribution information to validate the stability of Lin's short-version BUSQUAL. The result of the study indicates that Lin's short-version BUSQUAL model with 14 items and four factors displays stability and therefore offers a stable and reliable approach for measuring the service quality of city bus.