An Assessment of Group Measurement Invariance Across Five Years of Healthcare Employee Engagement

Abstract

Assumptions of measurement invariance are infrequently tested. Establishing the psychometric properties of instruments supports validity and reliability of group research. This study aims to determine factorial invariance of the Employee Engagement Survey (EES) within and between groups through structural equation modeling. The research design is a cross-sectional analysis of archival data. The EES was administered electronically for five years (2011 - 2015) to employees of a healthcare system made of up five hospitals and a shared services group. Results confirmed the reliability of the Engagement construct within the instrument, but no other latent factors. The measurement invariance results indicate non-invariance within and between pairings of years and facilities. Although a major M&A occurred in 2013, it is not a primary source of non-invariance in the data set. Organizations relying on psychometric instruments to measure unobserved phenomena may err in results interpretation and decision-making when invariance is not established. The findings contribute to the body of literature on employee engagement and measurement invariance in professional settings.