Applying the Bifactor S-1 Model to Ratings of ADHD/ODD Symptoms: A Commentary on Burns et al. (2019) and a Re-Analysis.

Abstract

To examine the construct validity of attention-deficit/hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD), the bifactor S-1 approach has been applied as an alternative to the fully symmetrical bifactor models in order to eliminate anomalous results and to allow for an unambiguous interpretation of g- and s-factors. We compared and contrasted our results with those of Burns et al. (2019) and extended their analyses by taking into account a two- vs. a three-factor structure of ADHD. Data from our previous research were reanalyzed and reinterpreted in accordance with the bifactor S-1 approach, constructing different models with hyperactivity (HY), impulsivity (IM) or hyperactivity/impulsivity (HI) as the general factor. No anomalous results were observed. All factor loadings were significant. Our results were comparable to those reported by Burns et al. (2019), although items from the specific subscales inattention (IN) and ODD accounted for more variance in our sample. Model fit for our HI model was comparable to that in Burns et al. (2019). In our sample, model fit was best when solely HY or IM was chosen as a general reference factor. However, in these cases, the remaining specific factor IM or HY was weakly defined. Overall, we were able to replicate the results found by Burns et al. 2019), although our factor loadings on the g-factor were slightly lower and our specificity regarding IN and ODD was slightly higher. Our results support a two-factor structure of ADHD/ODD in a clinical population.