(PS3-A10) Factor Structure of the Difficulties in Emotion Regulation Scale (DERS) Among Early Adolescents: Results from the Adolescent Brain Cognitive Development Study

Introduction: Early adolescence is a developmental period characterized by biological and psychosocial transitions, as well as increases in stress and emotional reactivity. The development of effective emotion regulation (ER) skills during early adolescence can support adaptive responding while deficits in ER have been linked to negative health outcomes during subsequent developmental stages. Despite the relevance of ER on early adolescent health, limited work has validated existing measures of ER ability among early adolescent samples. Addressing this methodological gap can help inform theoretical models of emotional development as well as early-stage clinical interventions that buffer against later adolescent health consequences.

Method: This study examined the factor structure of the 36-item Difficulties in Emotion Regulation Scale (DERS) in a subset of early adolescents (N=2300) recruited from the Adolescent Brain Cognitive Development Study. Prior to analysis, the initial cohort was randomly split into two subsets (n=1150). We implemented an a-priori three-pronged approach: 1) A confirmatory factor analysis (CFA) assessed the fit of the previously validated six-factor DERS in Sample 1; 2) An exploratory factor analysis (EFA) identified an alternative factor structure in Sample 1; and 3) A second CFA assessed the new model in Sample 2. Lastly, a bi-factor model assessed the global structure of the DERS total and subscale items.

Results: The original six-factor structure yielded poor-to-adequate fit. EFA results supported an alternative five-factor model, with different item mappings and exclusion of ten items. CFA results further supported the new five-factor solution, yielding adequate-to-good fit. A bi-factor model, estimating an overall DERS general factor along with the five subscales, also demonstrated good fit.

Discussion: A five-factor structure of the DERS appears justified in a large community sample of early adolescents. Consistent with prior work, items from the former Awareness and Clarity subscales of the DERS were combined into a single factor. All but one item from the former Strategies subscale were omitted, suggesting there may be developmental considerations rendering those items less relevant. While a bi-factor model including a total score was supported, the use of the distinctive five subscales is recommended.