Cheung, C., Goodman, D., Leckie, G., & Jenkins, J.M. (2011). Understanding contextual effects on externalizing behaviors in children in out-of-home care: Influence of workers and foster families. Children and Youth Services Review, 33(10), 2050-2060.
Some research suggests that children residing in out-of-home care are more likely to exhibit externalizing behaviours (e.g., destruction of property, violation of societal norms, harm towards others) than children living with their biological parents. There is considerable individual variation in levels of externalizing behaviour. The expression of externalizing behaviour is likely impacted by many factors, including cognitive ability, placement disruptions, and maltreatment experiences. In addition to child specific features, some research shows that foster parent characteristics influence the externalizing behaviour of children in their care.
Using the Assessment and Action Record (AAR) data from the Ontario Looking After Children (OnLAC) project (2007-2008), the authors conducted secondary cross-sectional data analysis to explore the relationship between worker and foster families characteristics and externalizing behaviours of children in-care (i.e., foster, n=865, 81%; kinship, n=91, 9%; and group care, n=107; 10%) aged 10-17 years (average age of 13 years, 7 months). Results reveal two important findings: (1) child protection workers with less formal education tend to work more with children that have higher levels of externalizing behaviours (24% of the variance seen at the worker level); and (2) foster parent negativity (such as disfavouring an individual foster child) tends to be associated with higher levels of externalizing behaviour in the disfavoured child. These findings imply that child-level, worker-level, and foster family-level characteristics should be considered when developing policy and practice approaches.
The authors examined three sources of data: (a) AAR data from 2007-2008 for 1,063 children in-care who were randomly selected from 12 Children’s Aid Societies (9 urban agencies, 3 rural agencies); (b) Information on 528 workers, who were identified through cross-referencing worker-specific variables (of which 221, 42%, worked with multiple children in-care); and (c) Information on 962 foster parents, who were identified through cross-referencing foster parent specific variables. Five models were created to test the participant specific variables. The authors did acknowledge limitations to the study, including its cross-sectional design, which was unable to address the long-term relationships between the variables. There was also a low participant rate from child welfare agencies, and a lack of information on genetic influences.