Measuring Disparities in Child Welfare Systems: Five Lessons

Posted June 4, 2017, By the Annie E. Casey Foundation

Ensuring that children are safe from abuse and neglect is not easy task. As a nation, we have struggled to know which kids need to be removed from their homes and which can stay safely with their families. In the United States, the task is complicated by our history of racial inequality, which has often translated into children of color facing more hurdles to safety and opportunity than white children.

For two decades, the Casey Foundation has collaborated with child welfare agencies to improve how they help families keep children safe from abuse and neglect. More recently, these efforts have included developing tools for measuring and addressing racial disparities (see resources, below).

Today we are sharing key information for those who use data to help their agency or community extend equitable opportunities to all children. Advice presented here is built on understanding how to use the Relative Rate Index (RRI), a measure of racial and ethnic disparities discussed in a previous publication, blog and webinar.

Lessons From the Work

“When Casey first published materials on the RRI, our data team had limited experience using it,” says Tracey Feild, managing director of Casey’s Child Welfare Strategy Group (CWSG). “After using the RRI intensively in collaboration with several child welfare agencies, we have learned five keys lessons about how best to use this metric.” Lessons include:

  1. Measure early decision making, before children enter foster care. Disparities in child welfare placements like foster care are strongly influenced by decisions made earlier. “In child welfare, the largest disparity by race is usually at a system’s front door. Often that means when families first come into contact with the system, when children are reported to agencies for neglect or abuse,” says Katrina Brewsaugh, CWSG senior associate.

    Brewsaugh says an advantage of the RRI is that it can examine and compare racial differences at different decision points. For example, at reporting (when an agency first receives a complaint about abuse or neglect) and substantiation (when an agency determines whether a parent is responsible for neglecting or abusing a child). An example of how this is done: “To compute the RRI for entries into a child welfare system, you compare the rate of entry for children of color to the rate of entry for white children among children who have substantiated reports,” Brewsaugh says, adding that entry is a common place to find disparities in decision making about children of different races or ethnicities.
     
  2. Measure broadly. Disparities by race or ethnicity can also be found at other decision points beyond entry. “It is important to look at many of decision points throughout your system,” says Judy Wildfire of Wildfire Consultants, longtime Casey advisers. “To the extent that race and ethnicity data are available, agencies will want to develop RRIs to examine decisions about screening, case prioritization and substantiation. Use RRIs to look at decisions coming out of Team Decision Making meetings and, for jurisdictions with differential response, decisions on how children are assigned to different tracks. Use it for all subsequent decision points, too.”

    Since the RRI calculation is based on the subset of children immediately at risk for a particular decision point, Casey’s child welfare data team suggests computing multiple RRIs to help pinpoint where disparities exist — and set priorities for improvements. “Knowing that disparities in your system can be found at entry, but not in kin placements, for example, can help you focus your problem-solving efforts on the system’s front door,” Feild says.
     
  3. One measure cannot stand alone; underlying trends matter. When sharply higher numbers of white children enter a child welfare system without similar increases in entries of children of color, “be cautious in your interpretation of the decreasing RRI. Examine the numbers that went into the calculation of the RRI to determine if changes may be due to changes in the community’s demographics versus actual changes in practice,” Brewsaugh says.
     
  4. How you define race and ethnicity matters. Before beginning disparity analyses, jurisdictions should clearly define how membership in racial and ethnic groups will be determined.

    Like many organizations, CWSG follows Census Bureau practice and condenses multiple race and ethnicity fields into a single race and ethnicity variable. This creates mutually exclusive fields of single-race non-Hispanic persons, multiracial non-Hispanic persons and Hispanic persons of any race.

    “We have learned from experience to carefully unpack multiracial and Hispanic populations into subgroups when possible to examine and address disparities in outcomes,” Brewsaugh notes. It is not always possible due to sample size issues, but you should examine any large subgroups within those larger groups as findings could provide important information for assessment and intervention.

    Two sites tell the tale. In one, “we noticed that a large subset of multiracial children had disparate outcomes from their peers,” Brewsaugh says. “Kids who were white and Native American faced worse outcomes than other multiracial children and kids who were white-only or Native American-only. By comparison, in another site, we found that race, not ethnicity, was the driver of disparity. In this site, Hispanic black children had outcomes similar to non-Hispanic black children, while Hispanic white and non-Hispanic white children had similar outcomes.
     
  5. Defining improvement is elusive. Quantitative information is key to learning about child welfare systems. But questions always remain. For example, say that in one year the disparity in entries between black and white children goes from an RRI of 2.5 to an RRI of 1.25, but this happens as more children of both races are entering. Is that improvement? What if the number of black children entering has remained stable and the change in entries is due solely to more white children entering? What if the change is due to a dramatic decrease in black child entries?

To answer questions raised by RRI measures, you may need qualitative data and a thoughtful review of your system’s values and mission. Feild says, “As your jurisdiction embarks on addressing disparities, the key is to remember your goal. And that is to understand the question at the heart of all child welfare analytics: How are our children faring when they come into contact with our child welfare systems — and what can we do to improve their experience?”

Resources for measuring racial disparity in child welfare systems