Research That Matters (January 17 - 20, 2008)

Purpose: To what degree are child-placement professional decisions of social workers influenced by their prejudices against families from lower socioeconomic (SES) or minority statuses? Existing research indicates over-representation of low-SES families and minorities in the child-welfare system, and among those whose children were removed from home. However, minority and low SES families are more exposed to monitoring by the welfare system, which may explain the disproportional rates of intervention with such families. How can we disentangle the subjective decision making process from the intervening effect of overexposure, and verify the existence or lack thereof of biases and prejudices in child-protection decision-making?

Method: The participants were 105 social workers working for the child-welfare system in Israel. The research design was a factorial (i.e., experimental) survey, based on 12 vignettes. The vignettes were drawn from actual welfare files, and edited to fit the experimental manipulation. The manipulated factors included (1) level of objective risk to the child (High, None, or Ambiguous); (2) Child's family SES (Low-SES vs. High-SES); and ethnic origin of family (Majority=Ashkenazi vs. Minority=Sephardic). Each participant was presented with eight vignettes. Following each vignette, participants were asked (1) to assess the level of risk to the child ["subjective risk"], and (2) whether they would recommend out-of-home placement. Analyses were conducted at the vignette level, controlling for the within-subject-clustering using GEE models of regression and of logistic-regression in STATA.

Results: The subjective assessment of risk and the recommendation for out-of-home placement were both influenced by the manipulated child's family SES and minority status. Those results were especially prominent in the set of ambiguous objective-risk vignettes. Families of low-SES and families of minority status were more in danger of being assessed as risking their children and, after statistically controlling for the subjective risk assessment, were more in danger of having their children recommended for removal from home. Furthermore, even for the vignettes of high objective risk, low-SES cases were still subjectively evaluated as at risk more than high-SES families; and even after statistically controlling for the subjective risk assessment, were more in danger of having their children recommended for removal from home.

Implications: The results indicate the need for higher attention given to the role prejudice plays in the decision making process of professionals in child-protection and placement decisions, both at the assessment stage and at the placement recommendation stage. Further, the results indicate that social workers tend to perceive low-SES as a risk factor in and by itself that may interact with other risk factors even in cases of clear-cut high risk. This finding aggravates the implications of the previous one, and indicates that an intervention directed at helping poor families out of poverty may decrease their suffering from undue pervasive interventions. The results further indicate the need for stronger focus in SW education on issues of multicultural sensitivity and for disentangling SES from child-risk factors. Further, much more research is indicated in order to verify if those results are specific to the Jewish-Israeli context only, or may be universal.