A Bifactor Model of Disgust Proneness

Abstract

The current research evaluated a bifactor model for the Disgust Emotion Scale (DES) in three samples: N = 1,318 nonclinical participants, N = 152 clinic-referred patients, and N = 352 nonclinical participants. The primary goals were to (a) use bifactor modeling to examine the latent structure of the DES and in turn (b) evaluate whether the DES should be scored as a unidimensional scale or whether subscales should also be interpreted. Results suggested that a bifactor model fit the DES data well and that all DES items were strongly influenced by a general disgust proneness dimension and by five content dimensions. Moreover, model-based reliability analyses suggested that scoring a general disgust dimension is justified despite the confirmed multidimensional structure. However, subscales were found to be unreliable after controlling for the general disgust factor with the potential exception of the Mutilation/Death and Animals subscale. Subsequent analysis also showed that only the general disgust factor robustly predicted an obsessive-compulsive disorder symptom latent factor—a clinical condition closely related to disgust proneness; latent variables representing DES domains displayed weak relations with an obsessive-compulsive disorder factor above and beyond the general disgust factor. Implications for better understanding the structure of DES responses and its use in clinical research are discussed.