A Reexamination of Child Problem Behaviors as Measured by ECBI: Factor Structure and Measurement Invariance Across Two Parent Training Interventions
The Eyberg Child Behavior Inventory (ECBI) is a widely used instrument that assesses disruptive problem behaviors via parent report. The present study compares three different conceptualizations of problem behaviors using traditional confirmatory factor analysis (CFA), a bifactor–CFA, and a bifactor exploratory structural equation model, whereof the latter two represent novel conceptualizations of disruptive problem behaviors. Data were derived from 353 Norwegian parents assigned to parent management training—the Oregon model (PMTO; N = 137) or brief parent training (BPT; N = 216), who rated their child’s problem behavior (age 3-12 years). The factor models were estimated within the PMTO and BPT samples, and in the total sample. Results showed that the bifactor–CFA, with a general problem behavior factor and three specific factors representing oppositional defiant, conduct problem, and inattentive behavior provided the best model fit. Furthermore, factorial invariance across control/intervention groups and across time (pre/post intervention) was estimated within the PMTO and BPT samples. Results revealed properties of metric and intercept invariance across intervention/control groups and over time, within both interventions. Implications of the results are discussed.