Measure for measure: A multitrait-multimethod analysis of guilt and shame

There is ongoing discussion in the literature regarding the validity of various checklist and scenario-based measures of guilt-and shame-proneness. Confirmatory factor analysis was used to perform multitrait-multimethod assessments of trait, method, and error components in the top three contending instruments in this area. College students (85 men, 197 women) completed the Test of Self-Conscious Affect-Modified (TOSCA-M), Personal Feelings Questionnaire-2 (PFQ-2), and the Guilt Inventory. A model in which shame and guilt each represented a separate latent trait construct provided a good fit to the data. Scores for shame from both the TOSCA-M and PFQ-2 loaded highly on the latent shame factor. With the exception of scores for nonruminative guilt derived from the TOSCA-M, all remaining scores for guilt had substantial loadings on the latent trait construct of guilt. Discussion focuses on the need to pay serious attention to methods effects in this area, provides explanations for why two qualitatively distinct forms of guilt can be expected using these methods, and urges specialists in this area to refrain from using single measurement instruments to assess multiple construct.