Psychometric characteristics of the adult scale of hostility and aggression: Reactive/proactive (A-SHARP) and relation to psychiatric features of adults with developmental disabilities

Recently, we described the development of the Adult Scale of Hostility and Aggression: Reactive/Proactive (A-SHARP) (Matlock & Aman, 2011). The A-SHARP was derived by factor analysis of ratings of 512 adults with intellectual and developmental disabilities (IDD), and its resulting five subscales were designated as (1) Verbal Aggression, (2) Physical Aggression, (3) Hostile Affect, (4) Covert Aggression, and (5) Bullying. The items on each subscale are rated first for severity (the Problem scale) and second for “origin” (i.e., to reflect extent to which behaviors are planned or reactive; “Provocation scale”). This study evaluated psychometric characteristics of the A-SHARP in the developmental sample of 512 adults. Mean item-whole subscale correlations ranged from .67 (Physical Aggression) to .78 (Verbal Aggression) on the Problem scale. Interrater reliability (n = 39) ranged from .59 to .78 on the Problem subscales and from .54 to .78 on the Provocation subscales. For the entire sample, the correspondence between the Problem and Provocation subscales was low (−0.04 to 0.28), indicating independence between the scales. The A-SHARP Physical subscale was strongly correlated with Behavior Problems Inventory (BPI) Aggression frequency ratings (n = 512, r = 0.79), and strongly with BPI severity ratings (r = .86). We examined congruent validity between A-SHARP subscale scores on the one hand and four DSM-IV categories and Down syndrome on the other. A number of significant associations were observed between the A-SHARP subscales and diagnostic group, supporting the A-SHARP's congruent validity. Likewise, we examined the correspondence between use of psychoactive medicines and A-SHARP scores and found a number of associations between medication use and higher subscale scores. Overall, these results support the reliability and validity of the A-SHARP, and, as intended, the problem and provocation subscales appear to assess different constructs. However, much more work is needed to determine fully how well each of the subscales performs psychometrically.