Paper
Agreement between foster child and foster parent in the severity of problem behaviour
- issue: Issue 2 / 2008
- authors: Johan Strijker, Simon Oijen
- keywords: The Netherlands, perception, agreement, foster care
- views: 4656
- downloaded: 0
- download pdf (66 Kb)
abstract
Background. In making a diagnosis, setting up a treatment plan, and measuring the development of the child input of multiple informants is needed. Possible informants are the child himself, parents, teachers, therapists, group workers and significant oth-ers. However, a meta-analysis performed by Achenbach, McConaughy en Howell (1987) shows that the mean correlation in agreement between informants measures up to only .28, and for a parent and her child it measures up to .25. Because these figures are in the range from small to moderate in size, the question would be relevant which informant provides reliable and valid information. Given these figures are in the range from small to moderate in size, a relevant question would be 'Which informant pro-vides reliable and valid information?' Since a valid criterion or 'Gold Standard' is miss-ing in determining whose information represents the true psychopathology, Achen-bach (1991) concludes that each informant provides partial valid information.
To evaluate the outcome of a treatment in, for example, the reduction in problem be-haviour, the input of the child as an informant should be needed. However, few stud-ies have been performed to observe the differences in the assesment of the severity of problem behavior between the child himself and the parent, and no study was found within the field of foster care. In a literature review only three studies were found which observed the agreement between foster parenst and the teacher of the foster child. Up to now, no studies have been found about the agreement between foster parents and their foster children. In this paper we report the agreement between foster parents and their foster children with the help of a longitudinal design. A community sample serves as a non- equivalent control group.
Research questions
- Are there any differences in agreement between the foster care sample and the community sample over time?
- What is the stability in agreement?
- What is the association between the level of agreement and the status of the foster care placement?
Method. A longitudinal design was set up to measure the differences in level of agree-ment between groups over time. The foster care sample consists of 60 foster parents with their foster children (age range 11 - 17 years) and the community sample of 87 parents and their birth children (age range 11 - 17 years). Within half a year after placement in the foster family the CBCL was administered to the foster parent and the YSR to the foster child. These measures were repeated after one and a half year of stay. The same questionnaires and time schedule were used for the community sample. To determine the level of agreement, cross-informant items were selected within the scales Internalizing, Externalizing, and Total Problems. For a particular scale Pearson corre-lations over items were computed for each pair of parent-child and for each pair of foster parent-foster child, meaning that correlations are computed at case level and not at group level. From now on, these case correlations are referred to as dyadic correla-tions. The statistical model of the univariate ANOVA repeated measurement split-plot was used, with factor Time nested within factor Group.
Findings. The test results of the statistical model are the following:
- no difference in means was found in dyadic correlation coefficients between the foster care sample and the community sample for Internalizing and Externalizing;
- the size of the agreement was in line with the study results of Achenbach, McConaughy en Howell (1987);
- a longer stay in the foster family didn't increase the agreement;
- there was stability in agreement for parent-child dyads over one and a half year, but no stability was found for foster parent-foster child dyads;
- there was a difference in means between continued placements and breakdown placements for Interalizing, with a significant smaller mean for breakdown placements;
- discordance from the perspective of the child was contributing to the risk at a breakdown placement and that's the reason why;
- the severity in discordance is clinically significant in a clinical setting.
Implications and recommendations
- within a half year of placement, foster parents are accurate assessors of the seri-ousness of problem behaviour (in comparison with the birth parents);
- there is no stability in agreement between foster parents and their foster children. Therefore, agreement is dynamic and may be subject of treatment;
- in reporting behaviour problems, there is need for systematic assessment with standardized questionnaires, and
- a strong need for children's contribution;
- clinicians must equally weigh foster parents' information and that of their foster children;
- if the discordance is large in size, mediation is needed to prevent the placement breakdown.
Key references
Achenbach, T.M. (1991a). Integrative guide for the 1991CBCL/4-18, YSR, and TRF profiles. Burlington, VT: University of Vermont Department of Psychiatry.
Achenbach, T.M., McConaughy, S.H., & Howell, C.T. (1987). Child/adolescent be-havioural and emotional problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin, 101, 213-232.
Contacts: Johan Strijker, State University of Groningen, Department of child and youth care, j.strijker@rug.nl.