“…Despite substantial evidence supporting the use of the bifactor model for ADHD symptoms, existing studies that used IRT set forward either a single unidimensional model for all ADHD symptoms, or separate unidimensional models for inattention and hyperactivity/impulsivity (Garcia-Rosales et al, 2015; Gomez, 2008; Li, Reise, Chronis-Tuscano, Mikami, & Lee, 2015; Purpura, Wilson, & Lonigan, 2010). While assumptions of unidimensionality of ADHD symptoms may be appropriate for nonclinical samples (Arias et al, 2016), the multidimensionality that is observed in clinical samples and associated unmodeled residual dependence between items may result in biased parameter estimates (Chen & Thissen, 1997), and overestimation of IRT item information and test reliability (Sireci, Thissen, & Wainer, 1991).…”