ADHD and method variance: a latent variable approach applied to a nationally representative sample of college...

This study employed a correlated trait-correlated method application of confirmatory factor analysis to disentangle trait and method variance from measures of attention-deficit/hyperactivity disorder obtained at the college level. The two trait factors were Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV ) Inattention and DSM-IV Hyperactivity-Impulsivity. The two source factors were self-reports and parent-reports. Data were collected for an epidemiological sample (N = 1,079) of college freshmen stratified for race/ethnicity, gender, and ability level according to national targets for the U.S. college population. Results revealed (a) parents' ratings were better measures of internalizing behavioral dimensions and that students' ratings were better measures of externalizing dimensions of behavior, (b) informants have a greater impact on behavior ratings than the behavioral construct that is presumed to be the primary cause of the behavior as measured by the CARE, (c) relationships among the method factors revealed a substantial amount of unique variance among informants, and (d) relationships among trait factors were largely within expectation.