The power of wishful thinking: the case of “race-blind removals” in child welfare

Wishful thinking is a very human pattern of thought that can even be functional at times. Thanks to wishful thinking, a placebo can actually cure an illness. Great ideas can gain support even if we don’t know they will succeed. But when wishful thinking is used to distort available data to support a given theory or policy, it becomes a problem. Such is the case with “race-blind” removals. The story of how this simple concept became viewed as a solution to the disproportional share of Black children in foster care, in relationship to their share of the general population, is a case study in the misuse of data to promote a particular viewpoint.

As reported in The Imprint and the Los Angeles Times, the Los Angeles Board of Supervisors has voted to support a project testing “race-blind removal” or “blind removal” of children into foster care. Blind removal was pioneered in Nassau County, New York in 2011 and “discovered” (as they describe it) by a team of researchers headed by Jessica Pryce of Florida State University, who were investigating the practices of two counties that were credited by New York’s Office of Child and Family Services (OCFS) with reducing racial disparities. Before the inception of blind removal, county investigative workers were presenting cases to a committee made up of supervisors, managers, and an attorney before a child could be removed and placed in foster care. Under blind removals, the members of this committee were no longer given information that might give a clue as to the race of the child and family. According to an email from a county official, the information that is withheld includes race, ethnicity, first and last names, addresses, the location of the reporter if that reflects the community where the child lives, and any other information (such as socioeconomic status or receipt of government benefits) that is not deemed to affect safety or risk.

Nassau County adopted the blind removal policy as a way to address its high rate of racial disproportionality in foster care, with Black children being much more likely to be removed and placed in foster care than White children. According to data provided by New York State, Black children were over 14 times more likely than White children to be placed in foster care in Nassau County in 2010. The blind removals policy is based on the belief that implicit racial biases affect the decision to remove a child and that removing this information from the process will remove the bias.

Unless there is strong evidence in support of such a program, one might worry about a practice that relies on people who know so little about a family. One might wonder if such a meeting is the best use of time for overburdened social workers and supervisors. Perhaps it would be better to make sure investigators have enough time to interview everyone who might be able to give them information about the family under consideration rather than burden them with another meeting. And what about emergency situations, where a child cannot be safely left in the home? An article in Children’s Bureau Express documents concerns from social workers who fear that blind removals would make it harder to do their jobs for these and other reasons. Another concern is whether race-blind removals might provide more of an opportunity for investigative workers to express any racial bias they have, since they control the information that is presented to the committee.

But if blind removal truly does cause a significant reduction in racial disparities, perhaps it is worth implementing despite the costs. And if one can believe a TED Talk by Jessica Pryce that has been viewed 1.3 million times, the practice has been spectacularly successful. According to Pryce, “In 2011 57 percent of the kids going into foster care were black, but after five years of blind removals, that is down to 21 percent.”  (At which point the audience broke into applause). Such a simple idea and such a huge impact! Casey Family Programs, the nation’s most influential child welfare funder, highlighted this program in an article on its website, stating without providing numbers that “within five years, the number of Black children removed from their families was reduced considerably, representing the most significant decrease in racial disproportionality within the county system ever.” New York State was so excited that it required all counties to develop a blind removal process effective October 14, 2020, offering a strikingly vague and yet broad description of what information must be kept from the committee: “all demographic and identifiable information (race, gender, language needs, zip code, etc. sic)).” Several other jurisdictions have expressed interest, including Los Angeles County, which is proceeding with its pilot.

Such a great result should be documented and the data made available to the public and researchers, preferably online, so it is surprising that Pryce was unwilling or unable to provide the source of the percentages at the heart of her popular talk. Instead, she referred me to the data team at Nassau County, who did not respond, nor did did the Commissioner’s Office. Nor was Casey Family Programs able or willing to provide the document referenced in their footnote to their statement about the program’s stellar results. Happily, I was able to obtain data from the New York State Office of Child and Family Services showing the percentage of children removed into foster care who were Black every year from 2009 to 2020. Those percentages are shown in Chart One.

Chart One

The first fact that emerges from the New York data is that Jessica Pryce’s percentages were not accurate. The 57 percent (56.7 percent) that she cites as the percentage of Black children removed in 2011 was actually the percentage of Black children removed in 2010. As for the 2016 data (the endpoint of the five-year-period cited by Pryce), 37.1 percent of the children removed in 2016 were Black, rather than 21 percent cited by Pryce–rather a large difference. There was a sharp increase in the Black share of children removed, from 45.2 percent in 2009 to 56.7 percent in 2010, the year before the program was implemented. With the implementation of blind removals, the percentage of children removed who were Black declined for two years to 45.5 percent in 2012, then rose for two years to 57.4 percent in 2014, fell to its all-time low of 37.1 percent in 2016, then rose to 49.7 percent in 2018, dipping slightly back to 45.1 percent in 2019, then popping back up to 49.5 percent in 2020–higher than it was in 2009 before the program was implemented. With such large fluctuations from year to year, as well as changes in direction, it is hard to imagine drawing any conclusions from the difference between any particular two years.

It is also important to note that the total number of children placed in foster care in Nassau County dropped precipitously from 429 in 2009 to 91 in 2020, as shown in Chart 2. This drastic drop in removals of Black and other children means that there was a lot more going on than the effort to make removals race-blind; cutting removals by three-quarters requires major changes in policy and/or practice. So it is hard to attribute any change with confidence to the race-blind policy. It also means that the numbers of children removed became smaller and smaller, resulting in a larger margin of error.

Chart Two

OCFS also provided data on Nassau County’s “Black Admissions Disparity Rate.” This rate, which New York State collects for all its counties, is defined as the “ratio of unique Black children admitted to foster care per 1000 Black children under 18 relative to comparable rate for White children.” According to OCFS, the disparity rate for foster care admissions went down from 14.30 (meaning Black children were 14 times more likely to be removed than White children) in 2010 to 12.60 in 2020. But it fluctuated to a surprising degree (between 24.4 and 6.16 between 2011 and 2019) that is not consistent with the percentages shown above and casts doubt on the correctness of the ratios provided. Assuming the 2020 ratio is correct, Nassau County currently has the highest disparity in foster care placement for Black children in the entire state. According to its ranking of counties based on this ratio, Nassau County was at the bottom in 2020 of all counties listed* with its disparity rate of 12.6, compared with 3.34 for the state as a whole. Hardly a role model for New York or the nation! Now this doesn’t mean we should blame Nassau County’s child welfare system for its abysmal disparity ranking. Other factors are probably behind that large disparity compared to other counties, such as the socioeconomic status of the Black and White populations in a given county. Which raises the question, how much can we expect blind removals to change racial disparities in foster care?

New York State recognizes the weaknesses of its data but focuses on the positive overall trend between 2010 and 2020. As John Craig of OCFS put it in his email to Child Welfare Monitor, “While Nassau County has seen fluctuations in the rate of Black children entering care over the past 10 years, overall, the trend has been very positive. OCFS commends Nassau County for recognizing the disproportionality of children of color in the child welfare system and implementing this innovative approach.” 

Despite OCFS’ valiant attempt to portray Nassau County’s data as “very positive,” the data do not provide a strong justification for expanding the program. While the Black percentage of children taken into foster care in 2020 was 49 percent compared to 57 percent in 2019, there were changes in both directions in the years between those two dates, and the 49 percent was actually higher than the Black percentage in 2009, two years before program implementation. There is reason to wonder whether New York, Los Angeles and others were really concerned about what the data showed. Instead, they may have proceeded in part based on the inherent logic of the approach, which addresses racial disparities directly in a way that is appealing to those who seek a relatively simple solution. Most importantly, they wanted it to work, so they decided that it did, regardless of the highly equivocal findings.

It would be wonderful if we had easy solutions to racial disparities in child welfare, but evidence suggests that higher reporting, investigation and removal rates among Black children stem from their greater needs, rather than bias among social workers. LA County would be better off studying how to make CPS decisions more accurate rather than imposing a cumbersome and unproven hurdle on social workers trying to protect endangered children.

* Certain counties were excluded because of very small number of Black children or Black children taken into foster care.