RACIAL DISPROPORTIONALITY IN CHILD WELFARE: False Logic and Dangerous Misunderstandings
Russell, Jesse, Judicature
Dependency court judges struggle to achieve the best outcomes for children and families.
Disproportionality and disparities in child welfare appear to be widely recognized, if not fully understood, phenomena.1 Dependency court judges see first-hand racial imbalances appearing in the courtroom and struggle to understand why and how to safely reduce the imbalance to achieve the best outcomes for children and families. There is often disagreement on how to interpret or find meaning in the empirical evidence that supports the existence of disproportionality and disparities; some the result of fertile and valuable discussion; some possibly stemming from honest misunderstanding or misinterpretation of the empirical evidence. Several of these potential missteps are addressed in this paper, and these might not be the only potential pitfalls in meaningfully understanding the data surrounding disproportionality and disparities. They are presented here merely to further the discourse along a productive path.
Clarifying areas of misunderstanding or misinterpretation of data is a step toward formulating policies, practices and tools for reducing disproportionality and disparities in a meaningful way. This paper introduces five areas of potential misinterpretation or misunderstanding of empirical data, and it certainly does not exhaust these topics. The five areas of potential misunderstanding or misinterpretation are based in:
* the ecological fallacy concept,
* the fallacy of hidden assumptions,
* the lessons from different measures of disproportionality,
* the difficulty in understanding how probabilities relate to each other, and
* the effect that multicollinearity can have upon statistical findings.
This paper fits into a larger discussion about evaluating research, policy and best practices on racial disproportionality and disparities in juvenile dependency. First, some definitions are in order. For this paper, disproportionality is defined as follows: disproportionality refers to one population being out of proportion with respect to the general population.2 Disparity is defined as a lack of equality:3 unequal treatment of one racial or ethnic group as compared to another racial or ethnic group. The National Council of Juvenile and Family Court Judges' Courts Catalyzing Change (CCC)4 Steering Committee expressed the judicial perspective on the challenges of communicating an understanding about disproportionality and disparity as:
The challenge is how to maximize this opportunity to do something that will reduce disparity without getting mired down in the feelings and emotions you have when you think about how this affects one personally. The challenge is to stay focused on the question: How do I communicate this in my jurisdiction in a way that will not create barriers?5
The judges of the CCC Steering Committee are right - the challenge is to maximize the opportunity. The challenge is not only to do so without being stuck in the emotional mire, but also without being stuck in any empirical/statistical swamps.
The Ecological Fallacy
It has been posited that the large representation of African American children in foster care is due to higher maltreatment rates for African American children, resulting in the disproportionate representation of African American children in the dependency system. This claim must be carefully examined to understand fully the actual units of analysis, the groups being analyzed, and the individuals making up those groups.
For example, Elizabeth Bartholet argues that: "First and foremost is that blacks [sic] are disproportionately associated with a set of characteristics that have been repeatedly found by many different child welfare experts to be accurate predictors for child maltreatment ... and there is no doubt that they are disproportionately associated with black families."6 She argues that "there is substantial evidence that black maltreatment rates are significantly higher than white, because black families are affected by poverty and other risk factors for maltreatment at significantly higher rates than whites"7 At first blush, this logic seems obvious: some characteristics are associated with higher rates of maltreatment, one group is more likely to have these characteristics, and thus, this group is more likely to have higher rates of maltreatment. …