Factor Structure and Psychometric Properties of the Children’s Scale for Hostility and Aggression: Reactive/Proactive (C-SHARP)

Although the problem of aggression in children with developmental disabilities is pervasive, there are currently no scales adequate for its measurement. This study explored the factor structure and psychometric characteristics of the Children’s Scale for Hostility and Aggression, Reactive/Proactive (C-SHARP). The C-SHARP comprises 58 items with two Likert scales: (a) the Problem Scale, where frequency and severity are rated, and (b) the Provocation Scale, where the degree of proactivity or reactivity is evaluated. The ratings of 365 children with developmental disability (mean age, 12.5 years; 60.5% male) were subjected to exploratory factor analysis. The most appropriate factor structure, judged by face validity and measures of fit and reliability, left 48 items on five factors: (I) Verbal Aggression, (II) Bullying, (III) Covert Aggression, (IV) Hostility, and (V) Physical Aggression. The fit of the model was adequate (RMSEA = 0.071), and the internal consistency of the subscales was high (average α = 0.87). The validity of the C-SHARP was explored using demographic, diagnostic, and medication variables. Several validation groups differed in expected ways on C-SHARP subscales scores, supporting the convergent and divergent validity of the instrument. Interrater reliability was extremely high on the Problem Scale (ICCs ranged from 0.67 for Hostility, to 0.91 for Bullying); although interrater reliability was considerably lower for the Provocation Scale. Overall, the C* Department of Psychology, The Ohio State University, Columbus, OH 43201, E-Mail: [email protected]. I would like to thank Dr. Michael Aman for his generous guidance. SHARP appears to have good reliability and validity, and has promise as a useful tool in the assessment of children with developmental disabilities. Introduction and Method Aggression is a common and often devastating condition of human existence. Because its impact is so great, it comes as no surprise that there is a wealth of literature discussing the etiology, prognosis, and treatment of aggression. There are several scales that measure aggression in typically-developing populations. However, it is not appropriate simply to use a scale designed for this group when assessing children with developmental delays (DD), as we have no assurance that the structures of these instruments are valid or appropriate. Although the problem of aggression and behavioral disturbance is consistently shown to be more common in DD populations than in the general population (e.g., Einfeld & Aman, 1995; Hill & Bruininks, 1984; Stark, Menolascino, Albarelli, & Gray, 1988), there are relatively few instruments appropriate for DD individuals (Aman, 1991). Aggression is commonly addressed with serious pharmacological and behavioral treatments, so a scale that can adequately measure it is imperative. With the creation of the C-SHARP, we hope to fill this void. Aggression is a heterogeneous construct, and there is reason to believe that types of aggression have varying responses to treatment. Reactive (“hot”) aggression is characterized by hostile attributions (i.e., the tendency to assume, erroneously or otherwise, that others intend to hurt the subject) and subsequent impulsive retaliatory behavior. Proactive (“cold”) aggression, however, is associated with the learned connections between desired consequences and aggressive behavior. Both subtypes are consistently seen in children (e.g. Dodge & Coie, 1987; Hinshaw & Lee, 2003; Pulkkinen, 1996). The delineation is important because optimal outcome is probably best achieved by matching the subtype of aggression to a treatment which addresses the specific type of behavior found in each (Dodge, 1991; Kempes et al. 2005; Kingsbury, Lambert, & Hendrickse, 1997; Smithmyer, Hubbard, & Simons, 2000). To this end, the CSHARP was developed to measure the extent to which the child’s behavior is reactive or