Measurement invariance of employee engagement across race groups

Abstract

Purpose The testing of measurement invariance is important in cross-cultural research to establish whether the psychometric properties of an instrument remain valid and reliable across different sample groups as these assumptions are rarely tested statistically. The purpose of this paper is to determine the factorial invariance of the employee engagement questionnaire across the various race groups by means of structural equation modelling. Design/methodology/approach Cross-sectional and descriptive research designs were followed in this study in the form of non-probability, convenience sampling to attract a sample of 1,175 employees in financial institutions. The employee engagement instrument (EEI) was electronically administered to 285,000 people who form part of a research database. Findings The results confirmed the reliability and validity of the instrument as determined by the exploratory and confirmatory factor analysis. Lastly, the results indicated that invariance can be assumed across race groups for financial institutions. Practical implications It is important for organisations to take cognisance of how specific socio-demographic variables influence the measurement of employee engagement, in this case race. The conclusion reached was that the EEI can be used with confidence in the financial sector for future employee engagement assessments. Originality/value These findings add to the current body of literature that exists on employee engagement and race in the South African work context and addresses one of the complexities assessment practitioners might have to comply with regarding questionnaire validity across race groups.

The influence of time on employee engagement in the SA business environment

“…Cronbach's α for the six dimensions ranged between 0.813 and 0.942. A recent study indicated that the fit indices of the base model (GFI: 0.809; RMSEA: 0.048; NFI: 0.860; IFI: 0.947; TLI: 0.942; CFI: 0.947) were all at levels recommended by the relevant scholars in the field (Gallant, 2017). The six dimensions are discussed in the section Measuring instruments.…”

Exploratory study

“…Considering that most of these employee engagement measurement instruments have only been around since 2010, they are relatively new, and it is consequently too early to reach conclusions about their universal shareability. As mentioned, the instrument of Nienaber and Martins (2015) is accompanied by a range of contextual, validation studies (Gallant and Martins, 2018; Martins, 2016; Martins and Nienaber, 2018; Nienaber and Martins, 2015) and has therefore overcome the limitations evident in similar instruments (Purcell, 2014).…”

“…The authors use the latest survey results of the validated employee engagement instrument of Nienaber and Martins (with the permission of the developers of the instrument) to demonstrate empirically how employee engagement can act as a mechanism to facilitate successful strategy implementation, if properly applied. The instrument of Nienaber and Martins (2015) is accompanied by a range of contextual validation studies (Gallant and Martins, 2018; Martins, 2016; Martins and Nienaber, 2018; Nienaber and Martins, 2015), overcoming the limitations of similar instruments (Purcell, 2014). The next section extends the literature provided in the introduction to make the links between employee engagement, strategy implementation and organisational effectiveness clearer.…”

Determinants of Tier 2 Indian consumer’s online shopping attitude: a SEM approach

“…The results do not support full metric invariance. Byrne et al (1989) introduced the concept of partial measurement invariance, in which only a subset of parameters in a model is constrained to be invariant while another subset of parameters is allowed to vary across groups (Teo et al, 2009;Gallant and Martins, 2018). Thus, subsequent analysis provides support for partial metric invariance.…”