Measuring Disparities in Child Welfare Systems: Five Lessons

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

Blog measuringdisparitiesinchildwelfare 2017

Ensur­ing that chil­dren are safe from abuse and neglect is not an 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

Popular Posts

View all blog posts   |   Browse Topics