The Family First Prevention Services Act monetarily incentivizes child welfare agencies to invest in programs that prevent foster care placements. It constitutes the largest commitment to that effort in the history of federal funding.
However, there are two key flaws that greatly limit the value of the Family First Act. One pertains to what services are permitted under this law, and the other pertains to how the effectiveness of those services is assessed.
First, the Family First Act only approves services along four solution areas: mental health, substance abuse, in-home parent skill based and kinship navigator.
This narrow scope means that issues such as lack of access to housing, which one study showed threatens out-of-home placement of kids in one of six families investigated for maltreatment, will not be approved for funding. Similarly, neglect that leads to 61 percent of out-of-home placements, encompasses everything from educational neglect, to failure to provide adequate clothing or supervision. While some of these underlying factors of neglect could be addressed by approved solutions like in-home parenting, many other underlying factors, especially those related to poverty, cannot.
The top contributing factors toward foster care placements also vary from county to county and state to state. Mandating a small scope of solutions does not give counties the incentive to diagnose and create solutions to the causal factors in their own regions.
These factors also change when we look at racially disaggregated data. According to a report by the Center for Advanced Studies in Child Welfare, American Indian children are 15.5 times more likely and African American and multiracial children four times more likely to be placed in out-of-home care, as compared to white children. A racial lens needs to be applied to the specifics of the law. Enforcing solutions without diagnosing the underlying factors to foster care placement, specifically for kids of color, would increase institutionalized racial inequities.
The second flaw in the Family First Act is that the tapered scope of services permitted has been further narrowed by a deliberate push to prioritize impact evaluations of these services using randomized controlled trials, or RCT.
The Prevention Services clearinghouse, which conducts a review of prevention programs for approval and funding, gives an undue preference to RCTs. Only studies designed using RCTs can earn a “high support of causal evidence” status, according to the clearinghouse. But quasi-experimental methods are considered to show only a “moderate support of causal evidence.”
RCTs originated from clinical trials and work best in public health. They rely on randomized allocation of individuals to a control and a treatment group and estimate the difference in outcome between both of these groups.
Unfortunately, the results of RCTs are difficult to generalize to a larger population. This means that a program that has shown positive results in a county through an RCT can show very different results when scaled up across a country as large as the United States. Development sector researchers have repeatedly argued that RCTs have issues when generalized to a larger population and this makes reliance on them problematic for the child welfare sector.
RCTs also pose ethical dilemmas, especially for agencies that have been implementing a prevention program, even before the Family First Act. In RCTs, a researcher must decide which families can receive the prevention service and place them into the treatment group, and which families cannot receive services and place them into the control (placebo) group. How does one decide which child, potentially reported for maltreatment, is eligible to receive services, and which is not?
Impact evaluators cannot play God with the lives of children.
Quasi-experimental methods do not pose such problems. Here, a researcher artificially constructs a control group from the general population without restricting what services a family can have access to. While there is an understandable fear that this could lead to bias in the selection and allocation of groups, several researchers have found that quasi-experimental methods closely mimic RCT results, when conducted with high scientific rigor.
According to one report, an average quasi-experiment costs one-fourth of that of an RCT. Quasi-experiments are particularly advantageous because they can include people who would normally be excluded from RCTs. Quasi-experiments are analyzed in real-world settings instead of artificially created settings that RCTs require. Hence, researchers agree that the results of quasi-experiments are more reliable while scaling up to a larger population, as compared to those of RCTs.
Granted, policymakers want to adopt the “gold standard” and look for the best evidence available while allocating federal dollars to prevention programs. They do not want to encourage agencies to invest in programs that do not show evidence of positive impact on children and families. But by defining evidence so narrowly, they force agencies to cut down investment in programs that are in fact, relevant and necessary. And by pushing agencies to attempt RCTs, they are setting a questionable standard to define impact and are incorrectly assuming that what works for one agency will work for everyone else across the United States.
Policymakers should create the space for agencies to deep-dive and diagnose the top contributing factors to foster care placement, in their own regions. They should keep agencies accountable and ask for evidence for impact created in the lives of families and children.
Creating incentives to preserve families is a step in the right direction. Mandating agencies to apply irrelevant solutions and promoting contentious research methodologies as the pillar of evidence, are not.
Mathangi Swaminathan is an alumna of the Harvard Kennedy School of Government and a Foster America fellow. She is currently placed with the Olmsted County Department of Health, Housing and Human Services in Minnesota.