Testing trust scale measurement invariance in project teams

Abstract

Purpose – An important assumption in testing theoretical models is measurement invariance. However, little research in construction project management investigates the issue of measurement invariance. To bridge the gap, the current paper aims at: first, testing trust measurement invariance in project teams across gender; second, delineating various measurement invariance tests and demonstrating them with Amos; and last but not least, increasing researchers' awareness about measurement invariance issue in the construction project management field. Design/methodology/approach – The measurement invariance of trust across gender is investigated using survey data collected from architects in project design teams of A-level architectural design institutes in China. Multisample confirmatory factor analysis is conducted with Amos to test configural invariance, metric invariance, scalar invariance, factor covariance and variance invariance, error variance invariance, latent mean invariance of the Chinese version of McAllister's two-dimension trust scale. Findings – All the test results of the above invariances are supportive. Goodness-of-fit indexes such as CFI deserve more empirical studies to verify. Research limitations/implications – The results imply that the trust structure of male and female architects is equivalent. In other words, male and female architect data of trust research in construction project management could be aggregated. Practical implications – It is recommended that test of measurement invariance should be conducted while new measurement scales are under construction. Originality/value – Since all the invariance tests are rarely conducted within a single study, the current research is the first paper to investigate the measurement invariance issue in the construction industry.