(WITHDRAWN) Concordance of Youth and Key Adult Ratings of Peer and Sibling Relationships for Youth in Care: Impact of Respondent Characteristics and Associations to Youth Resilience

Background

Estimating the level of agreement across youth and key adults on ratings of peer and sibling relationships can be valuable to assessment and intervention with youth in out-of-home care. To inform assessment and case planning efforts this study examined the role of respondent characteristics on rater concordance for sibling and peer relationships and their associations to youth ratings of perceived resilience. The research questions were: (1) what are the levels of concordance between youth and key adults on ratings of sibling and peer relationships? (2) are respondent characteristics associated with rater concordance? (3) are respondent characteristics and rater concordance associated with youth resilience?

Methods

A survey designed to assess the relational well-being of youth in out-of-home care was deployed to 316 dyads (youth, caseworker/caregiver). Constructs included a three-item measure of sibling relationship quality (alpha: youth=.76, adult=.80) and a two-item measure of peer relationship quality (alpha: youth=.75, adult=.74). Youth independently responded to an 11-item measure of perceived resilience (alpha=.86). Constructs were valued on a four-point scale and summed. Covariates included youth age, youth gender (male/female), length of relationship, and adult type (caregiver/caseworker).

Analysis

An independent samples t-test, Pearson correlation coefficient, Cronbach’s kappa, and double-entry intraclass correlation coefficient were calculated to determine rater concordance for sibling and peer relationship constructs. To determine if respondent characteristics were associated with rater concordance the absolute difference in dyadic ratings for sibling and peer constructs were standardized and entered into separate linear regression models as dependent variables with covariates youth gender (male/female), relationship type (caregiver/caseworker), and standardized values for youth age and relationship length. To assess the effects of rater concordance on youth resilience, rater characteristics, sibling concordance, and peer concordance were regressed on youth resilience.

Results

Sibling Relationships. Youth report of sibling relationships (mean=7.4, SD=3.1) were lower than adult reports (mean=10.7, SD=3.0). An independent samples t-test confirmed differences were statistically significant (meandiff=3.23, SD=4.53; t(315)=12.70, p<.01). The bivariate correlation (r=-.10, p=.08), kappa coefficient (κ=-.01, SE=.01), and two-way ICC (ICC=-.22, CI=-.52-.03) suggested low to no agreement between raters. In multiple regression analysis older youth age (β=.21, SE=.09) and greater length of time the adult knew the youth (β=.21, SE=.10) were associated with greater discrepancy. If the adult was a caregiver, this was associated with less discrepancy (β=-.19, SE=.10).

Peer Relationships. Youth report of peer relationships (mean=6.3, SD=1.7) were similar to adult report (mean=6.21, SD=1.34). An independent samples t-test did not find statistically significant difference between raters (meandiff=-.04, SD=2.01; t(315)=-.36, p=.72). No respondent characteristics were associated with peer concordance.

Resilience. Higher sibling relationship discrepancy (β =.20, p<.01), older youth age (β=.13, p<.05), and caregiving relationship (β=.25, p<.05), and were associated with greater youth resilience. Higher peer relationship discrepancy was negatively associated youth resilience (β=-.16, p<.01).

Conclusion

Results suggest older youth age, greater relationship length, and caseworker relationships were associated with greater discrepancy for sibling but not peer relationship constructs. Rater discrepancy for sibling and peer constructs were inversely associated with youth resilience, suggesting resilient processes differ when considering dyadic perspectives of siblings and peer relationships.