The psychometric properties and measurement invariance of the Burnout Assessment Tool (BAT-23) in South Africa

Abstract

Background: Burnout is an increasing public health concern that afflicts employees globally. The measurement of burnout is not without criticism, specifically in the context of its operational definition as a syndrome, also recently designated as such by the World Health Organisation. The Burnout Assessment Tool (BAT-23) is a new measure for burnout that addresses many of the criticisms surrounding burnout scales. The aim of this study is to determine the validity, reliability, and measurement invariance of the BAT-23 in South Africa. Method: A quantitative, cross-sectional survey, approach was taken (n = 1048). Latent variable modelling was implemented to investigate the construct-relevant multidimensionality that is present in the BAT. For measurement invariance, the configural, metric, scalar, and strict models were tested. Results: The analyses showed that the hierarchical operationalisation of BAT-assessed burnout was the most appropriate model for the data. Specifically, a bifactor ESEM solution. Composite reliability estimates were all well above the cut-off criteria for both the global burnout factor and the specific factors. The measurement invariance tests showed that gender achieved not only strong invariance, but also strict invariance. However, ethnicity initially only showed strong invariance, but a test of partial strict invariance did show that the mean scores could be fairly compared between the groups when releasing certain constraints. Conclusions: The BAT-23 is a valid and reliable measure to investigate burnout within the Southern African context.