Disproportionality, Disparity, and Definitions: Race and Child Welfare

October 27, 2020

Barbara Chaiyachati, M.D., Ph.D., Fellow Physician, Child Abuse Pediatrics, The Children’s Hospital of Philadelphia

Cindy W. Christian, M.D., Field Center Faculty Director, Anthony A. Latini Chair in Child Abuse and Neglect Prevention, The Children’s Hospital of Philadelphia

The disproportionate impact of COVID-19 on US minority populations and the spring’s video evidence of the unjust deaths of Black Americans by police and others have, yet again, illuminated racial and social injustices that exist in our country.  As we reckon with how to address these issues and shine light into our own shadows of privilege, we should all ask ourselves how both our individual and collective work contributes to injustice. In this month’s blog, we reflect on racial disparities and disproportionality in the child welfare system and areas to improve the definition of race in child welfare research.

Race in the Child Welfare System:

As many of us already know, the racial distribution of children involved with the child welfare system does not match the racial distribution of children in the general population. For example, the percentage of Black children in foster care is persistently higher than the percentage of Black children in the general population while the percentage of white children in foster care is lower than the percentage of white children in the general population. In 2000, while 15% of children in the U.S. were Black, 39% of children in foster care were Black. In comparison, 61% of children in the U.S. were white yet only 38% of children in foster care were white. Over the subsequent two decades, that disproportionality has decreased, though it still persists. In 2018, while 14% of children in the U.S. were Black, 23% of children in foster care were Black. In comparison, 50% of children in the U.S. were white while 44% of children in foster care were white.

Bias in the child welfare system has received and continues to deserve critical attention. Academic work has developed hypotheses of potential levels of bias underpinning the differences seen in the child welfare system. First, at the case level, different interpretation of the same findings could cause increased likelihood of child welfare involvement and substantiation. This could be fed by individuals’ overt racism or more subtle implicit biases. Second, at the institutional or system level, policies and procedures for mandated reporters, caseworkers, and judges that are inherently biased could cause differential outcomes. Third, at a society level, current and historic discrimination including structural racism with associated concentration of risks may impact distribution of risk factors for maltreatment, such as poverty.

Disproportionality and Disparity:

A conceptual distinction related to the levels of bias is the separation of disproportionality and disparity. Disproportionality is the over- or underrepresentation of a group in child welfare compared to the general population. Disproportionality, when reflective of true differential distribution of risk, may be just. On the other hand, disparity recognizes that, even when risk is equal, outcomes are different. This second concept, disparities in child welfare, is more overtly linked to forces of racism. Whereas disproportionality speaks to differences in populations’ risks, disparities speaks to injustices in the system. Thus, central to the discussion of race in the child welfare system is the considered balance between distribution of risk contributing to disproportionality and forces of racism contributing to disparities.

Regarding disproportionality and distribution of risk factors for maltreatment, it is an uncomfortable reality that all children are not at equal risk for experiencing maltreatment, for complex reasons. Research has attempted to objectively identify markers of risk to guide prevention and intervention. Literature suggests that the strongest individual and family associates for maltreatment include parent anger, family conflict, and family cohesion. Significant risk has also been found related to younger parent age, substance use, and perceived child misbehavior. Additionally, poverty is a strong risk factor for maltreatment, as well as associated factors including unemployment and single parent status. While recognizing that the assessment of subjective risk is not immune from forces of racism, the uneven distribution of objective risk among racial groups in the U.S. may have justified contribution to the disproportionality in child welfare.

In fact, in assessing the contribution of differential distribution of objective risk as exculpatory to racial disproportionality, multiple studies have shown that measures of maltreatment including child welfare investigation and substantiations are more closely associated with poverty than with race at both individual and population levels, suggesting that risk of child welfare involvement is greater as poverty increases for white families than minority families.

Yet, valid concerns persist for the insidious contribution of disparities in child welfare systems beyond disproportionate distribution of risk. Medical studies, including our own, have shown racial differences in how children are evaluated for possible abuse, and how race influences the recognition of (or failure to recognize) abuse. Decades later, subsequent assessment of medical evaluations related to possible abuse after clinical decision support showed a reduction of racial disparities and a large review of child welfare outcomes studies did not find worse outcomes for Black children compared to white children. However, improvements do not excuse injustice and persistent, critical inquiry is required to ensure persistent, uniform, equitable progress.

Definitions of Race

As further inquiries evaluate racial disproportionality and disparity, there is pressing need to address how race is defined. There are many shortcomings in the current treatment of race as a study variable in academic literature – below, we include just three of them.

First, few studies include their methodology for how race was assessed. When race was collected, was it offered by the participant or assigned by the interviewer? What questions were used to elicit race and what options for self-identification were given? There is also evidence that report of race is fluid and may change with repeat assessments over time. A first step to address this shortcoming is better clarification of methods used to generate race variables for research publications. 

Second, most studies use a single race category with exclusion or collapse of multiple race categories. Rapidly changing population demographic trends are manifesting first in young children. This is already evident in child welfare data: from 2000 to 2018, the percentage of children in foster care classified as “Multiple Race” has increased from 1% to 8%. In 2018, CDC birth certificate data showed that only 67% of U.S. births had a single race and ethnicity as defined by both parents sharing both race and ethnicity per CDC standard racial (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, White) and ethnicity (Hispanic or Non-Hispanic) matrix. Put another way: in 2018, up to 1 in 3 infants born in the U.S. could have parents of different races or ethnicities. Given the well-documented concentration of maltreatment risk in the first years of life, child welfare research must push to include analyses reflective of actual demographic realities. One step may be rigorous sensitivity analyses to explore the impacts of race category collapse on primary study results.

And finally, we must remember that race is not biologic but a social construct. Some studies in child welfare, particularly those with biologic outcomes such as health, may be better served by including direct measures of experiences for which race is a proxy in order to avoid contributing to the fallacy of race as a biologic measure. For example, measures that explicitly capture experiences of racism may provide more accurate and sensitive results regarding the toxic stress of racism and avoid the pitfall of race as a monolithic proxy for experiences of racism. Such methodologic shifts will not be easy as they will require validation, acceptance, and dissemination.

In closing,

Our challenge, as physicians and others who work to serve maltreated children, is to ask how we individually and collectively contribute to perpetuation of racism and racial disparities, and how we can work as partners to help right the system. As eloquently stated by Kareem Abdul-Jabbar earlier this year, “Racism in America is like dust in the air. It seems invisible — even if you’re choking on it — until you let the sun in. Then you see it’s everywhere. As long as we keep shining that light, we have a chance of cleaning it wherever it lands. But we have to stay vigilant, because it’s always still in the air.”

< Return to Blog