An ideal point account of the JDI Work satisfaction scale

Abstract

Applied psychologists have begun to show interest in an ideal point Item Response Theory (IRT) model, the Generalized Graded Unfolding Model (GGUM). As opposed to the traditional dominance models based on the Likert method, such as the Graded Response Model (GRM) and the Nominal Response Model (NRM), the GGUM does not assume that responses have a monotonically increasing relationship with the latent attribute. Currently, it is somewhat unclear as to when an ideal point process should be assumed. Here, the GRM and the NRM, two dominance IRT models, are evaluated against the GGUM, an ideal point IRT model, in fitting scales of the JDI. Only the Work scale possesses a large number of items that unfold. Fit was better under the GGUM than the GRM or NRM. Attribute estimates for the dominance models correlate .99, whereas GGUM estimates correlated .85 with GRM and NRM estimates. Unfolding items in the Work scale showed marked differences at the response-option level. It is proposed that ideal point processes may better theoretically and empirically explain appraisals of Work satisfaction.