Parent-Adolescent Cross-Informant Agreement in Clinically Referred Samples: Findings From Seven Societies
Authors:
Rescorla, LA; Ewing, G; Ivanova, MY; Aebi, M; Bilenberg, N; Dieleman,
GC; Dopfner, M; Kajokiene, I; Leung, PWL; Pluck, J; Steinhausen, HC;
Metzke, CW; Zukauskiene, R; Verhulst, FC
Source:
JOURNAL OF CLINICAL CHILD AND ADOLESCENT PSYCHOLOGY, 46 (1):74-87; 10.1080/15374416.2016.1266642 JAN-FEB 2017
Abstract: To conduct international comparisons of parent-adolescent cross-informant agreement in clinical samples, we analyzed ratings on the Child Behavior Checklist (CBCL) and Youth Self-Report (YSR) for 6,762 clinically referred adolescents ages 11-18 from 7 societies (M=14.5years, SD=2.0years; 51% boys). Using CBCL and YSR data, we asked the following questions: (a) Do parents report more problems for their adolescent children than the adolescents report about themselves? (b) How do cross-informant correlations (rs) for scale scores differ by problem type and by society? (c) How well do parents and adolescents, on average, agree regarding which problems they rate as low, medium, or high? (d) How does within-dyad item agreement vary within and between societies? (e) How do societies vary in dichotomous cross-informant agreement with respect to the deviance status of the adolescents? CBCL and YSR scores were quite similar, with small and inconsistent informant effects across societies. Cross-informant rs averaged .47 across scales and societies. On average, parents and adolescents agreed well regarding which problem items received low, medium, or high ratings (M r=.87). Mean within-dyad item agreement was moderate across all societies, but dyadic agreement varied widely within every society. In most societies, adolescent noncorroboration of parent-reported deviance was more common than parental noncorroboration of adolescent-reported deviance. Overall, somewhat better parent-adolescent agreement and more consistency in agreement patterns across diverse societies were found in these seven clinical samples than in population samples studied using the same methods.