Informant Discrepancies in Externalizing and Internalizing Symptoms and Adaptive Skills of High-Functioning Children with ASD

Background: Assessment of clinical symptoms requires information from multiple informants.  However, informant discrepancies may have significant repercussions when performing evaluations (i.e., access to services, treatment planning, progress monitoring; De Los Reyes & Kazdin, 2004).  Understanding informant discrepancies may be more important for children with ASD, who often have difficulty generalizing across settings (Stratis & Lecavalier, 2015).  Despite potential effects on access to services and intervention planning, there is a paucity of research that has comprehensively examined informant discrepancies for scales assessing externalizing, internalizing, and adaptive behaviors for children with HFASD.  

Objectives:  This study examined parent-teacher discrepancies for ratings of externalizing and internalizing symptoms and adaptive skills of children with HFASD.

Methods:  This study included a total of 236 BASC-2 ratings of children with HFASD from two informant groups (parents and teachers).  Each parent-teacher pair (n = 118 parents/caregivers and n = 118 teachers) rated the same child.  The children with HFASD, ages 6 to 11 years, were recruited from clinical trials and had a prior clinical diagnosis of ASD, WISC-IV short-form IQ > 70, and VCI or PRI factor score >80.  All diagnoses were confirmed using the ADI-R.  The dependent measure used was the Behavior Assessment System for Children-2 (BASC-2), Parent (PRS) and Teacher (TRS) Rating Scales.  Protocols were scored using the BASC-2 ASSIST Plus computer scoring software.  Scores on the Internalizing Problems, Externalizing Problems, and Adaptive Skills Composites were examined for mean differences, level of agreement, linear relationships, and moderators of discrepancies.

Results:  Mean difference comparisons were non-significant for the Externalizing Composite and its Aggression and Conduct Problems scales (negligible-to-small effect size ds, .02 - .25).  Only the Hyperactivity scale within this composite yielded a significant between-groups difference (d = .45).  For the Internalizing Composite and its scales, all between-groups comparisons were non-significant and the effect sizes were small-to-negligible (ds < .17).  Parent-teacher ratings on these Composites and scales were significantly correlated (generally moderate).  In contrast, teacher ratings were significantly higher than parents for the Adaptive Skills Composite and several of its constituent scales.  Effect sizes on these significant comparisons ranged from small (d = .48) to large (d = 1.06), with the Adaptive Skills Composite difference reflecting a medium effect size (d = .73).  Between-groups comparisons for non-significant scales resulted in small effect sizes (ds = .20-.30).  Correlations between informants on the Adaptive Skills Composite were significant (low-to-moderate), with variability in the correlations among its constituent scales.  Parent-teacher discrepancies were not moderated by child or parent variables.

Conclusions:  This study suggests a reduced likelihood of informant discrepancies for externalizing and internalizing symptoms, with larger discrepancies expected when assessing adaptive skills.  Reasons for the inconsistent pattern of findings are unknown but may involve more diversity in skills making up the adaptive skills construct or greater salience of behavioral symptoms. Poor agreement may have treatment and progress monitoring implications (e.g., difficulties identifying targets, establishing baselines, assessing intervention efficacy).