A Multiple Indicator × Multiple Trait × Multiple Source × Multiple Occasion design was used to evaluate invariance, convergent and discriminant validity of ADHD-inattention (IN), ADHD-hyperactivity/impulsivity (HI), and academic impairment scores from the Child and Adolescent Disruptive Behavior Inventory (CADBI) using confirmatory factor analysis. Mothers, fathers, teachers, and aides completed the CADBI on 811 Spanish 1st-grade children (54% boys) twice (6-week separation). For mothers and fathers, like loadings, thresholds/intercepts, factor means, factor variances, and factor covariances/correlations were invariant across sources and occasions. All 3 factors also showed convergent (convergent correlations from .69 to .83) and significant discriminant validity. For teachers and aides, there was also invariance of parameters along with convergent and discriminant validity over sources and occasions (convergent correlations from .67 to .87). With construct validity established for home and school, it was meaningful to test construct validity between home and school. Like-item loadings and thresholds/intercepts were invariant between home and school, with the ADHD-HI factor mean being lower at school. Convergent validity of ADHD-IN, ADHD-HI, and academic impairment factors, especially ADHD-IN and ADHD-HI, was much weaker between home and school (convergent correlations from .36 to .47 for IN and HI). The strong convergent validity of ADHD-IN and ADHD-HI scores within home and school in conjunction with weak convergent validity across home and school has implications for the assessment and diagnosis of ADHD (i.e., the diagnostic criteria of symptom occurrence in 2 or more settings).
Available at: http://works.bepress.com/christian-geiser/44/