(PS11-95) The Affective Reactivity Index in Adults: A Confirmatory Factor Analysis

Irritability is described as a “persistently angry, grumpy or grouchy mood, i.e., touchy or easily annoyed, angry or resentful” (Carlson & Klein, 2018). Despite the frequent appearance of this mood state across forms of psychopathology, irritability itself is understudied. The construct has the potential to represent a transdiagnostic process of clinical importance. The Affective Reactivity Index (ARI; 2012) was developed to assess chronic irritability. The ARI contains 6 items that measure irritable symptoms and a 7^th item that assesses impairment due to irritability. Psychometric evaluations of the ARI have been promising, but few studies have used this measure with adults.

Two investigations exist into the factor structure of the ARI (Grassi et al., 2023; Mulraney et al., 2014). Both evaluated the ARI within adolescent and adult samples separately. Using exploratory factor analysis, Grassi and colleagues determined that the 6-items of the ARI represented one construct. Using confirmatory factor analysis, Mulraney and colleagues discovered a high level of covariance between two of the items in the measure (“I often lose my temper” and “I lose my temper easily”), and they found that the removal of either item resulted in an excellent fit for the new 5-item measurement model. The current paper uses confirmatory factor analysis to evaluate the previously found measurement models of the ARI within a large, diverse sample of adults.

The present sample consisted of 1,631 college students recruited from 3 diverse institutions of higher education (elite private, public university, religious affiliation). The same had a mean age of 19.9 years (SD = 2.9), and the sample was 66.6% female. Data was gathered from an undergraduate population that is roughly 75% White, 10% Latinx, 10% Asian, and 5% Black. The ARI was administered along with a larger battery of psychosocial assessment tools.

Initial analyses suggested that the means and standard deviations for the ARI items were consistent with prior literature. The internal consistency for the original 6-item ARI was α=.791.  The present confirmatory factor analysis replicated the findings of Mulraney and colleagues in that items 2 and 6 covaried to a high degree. Modification indices suggested the removal of one of these items from the measurement model. Within the present sample, removal of item 2 (“I often lost my temper”) provided the best fit (CFI = .97, RMSEA = .083). Notably, the fit within this very large sample was not as strong as the fit Mulraney and colleagues found.

The 5-item ARI was found to be significantly related to other potential indices of irritability, such as the Angry Mood subscale of the Adult Self-Report of ODD Symptoms measure (r = .685), and the ADS Total score (r = .648). Consistent with the presence of irritability across diagnostic categories, the 5-item ARI was found to be significantly, positively related to indices of depression (DASS; r = .371), anxiety (DASS; r = .304), and the hyperactivity symptoms of ADHD (Adult ADHD Self-Report Scale; r = .297). The presented analyses add to our understanding of the appropriate ARI measurement model within adult samples, and support the notion of irritability as a transdiagnostic characteristic that may serve as a productive treatment target.