Psychometric Properties of the Difficulties in Emotion Regulation Scale (DERS) and Its Short Forms in Adults With Emotional Disorders

Objective: The Difficulties in Emotion Regulation Scale (DERS) is a widely used self- report measure of subjective emotion ability, as defined by a prominent clinically derived model of emotion regulation (Gratz and Roemer, 2004). Although the DERS is often used in treatment and research settings for adults with emotional (i.e., anxiety, mood, obsessive-compulsive, or trauma-related) disorders, its psychometric properties are not well-characterized in this population. Method: We examined the psychometric properties of the DERS and three popular short forms (DERS-16; DERS-18; and DERS-SF) in a large (N = 427) sample of treatment-seeking adults with one or more DSM-5 emotional disorders. Results: For the original DERS, internal consistency was strong for all subscales except Awareness. A bifactor structure consisting of one general emotion dysregulation factor and five uncorrelated specific factors corresponding to the original DERS subscales (excluding Awareness) provided the best fit. A series of structural equation models (SEMs) demonstrated unique incremental contributions of the general factor and several specific factors to explaining concurrent clinical severity. The general factor and one specific factor (Goals) also prospectively predicted treatment outcome following a naturalistic course of outpatient cognitive-behavioral therapy (CBT) in a subset of participants (n = 202) for whom discharge data were available. Specifically, more severe emotion dysregulation at intake predicted better CBT response, while more severe impairment in goal-directed activity when distressed predicted worse CBT response. All three short forms showed a robust bifactor structure and good internal consistency and convergent validity vis-à-vis the original measure, albeit with a slight decrement in incremental utility (1–3% less variance explained in clinical severity). Conclusion: With the Awareness items excluded, the DERS showed good internal consistency and a robust bifactor latent structure. The general factor and several specific factors incrementally and prospectively predicted clinical severity and treatment outcome, which suggests that the DERS may have clinical and predictive utility in treatment-seeking adults with emotional disorders. Additional research is needed to establish convergent and discriminant validity in this population. The use of a short form in lieu of the full DERS may be sufficient for many general clinical and research purposes, particularly when participant burden is a concern