The dimensionality and scaling of job satisfaction: An internal validation of the Worker Opinion Survey

Abstract

The Worker Opinion Survey (WOS) is a multifaceted scale measuring job satisfaction with a simple yes/no/not‐sure mode of response. Exploratory techniques were originally used to define the scales and judgemental techniques were used to derive the scoring. In the present study, based on a broad sample of 636 government employees, three possible models for the dimensionality of the WOS were compared using confirmatory factor analysis, and variants on the scoring system were evaluated using a Rasch model. The six‐factor solution of Cross was confirmed as most appropriate (with correlated factors) but it was found that the scoring system could be improved by considering ‘not‐sure’ responses differently for different subscales.