Abstract
Barth et al. (2021) published an article in this journal identifying ten topics in the field of child welfare that are frequently discussed among professionals, advocates, and researchers in an effort to shape discussions of practice and policy reform. Concerned that these discussions are often poorly informed by the research evidence, Barth et al. intended to offer a corrective to these common, erroneous narratives. The Editor-in-Chief, Bruce Thyer, asked for suggestions for commentators and then invited some number of respondents to offer their perspectives on the original article. Here, we respond to each of the submitted papers, highlighting areas of agreement, and addressing other topics where we—sometimes sharply—disagree. We welcome an ongoing, fact-based, respectful dialogue to help shape child welfare reform. Efforts to improve the child welfare system are urgently needed; we stand by our view that large-scale practice and policy reform, in particular, must be guided by the best available research evidence.
Introduction
We welcome and appreciate the time and thought of other scholars and practitioners regarding our critiques of ten too-common misconceptions in understanding child welfare. We also appreciate Bruce Thyer’s willingness to give us space to share our commentary, and his industry in reaching out to many respondents. There are too few of these opportunities for discussion in social work.
We endeavored to learn from all the papers, though most offered more general critiques of our approach, focusing on our privileging common scientific methods, than the content of our summary conclusions. The responses to our commentary capture some of the key messages that we sought to offer originally, as well as a number of critiques.
Our commentary drew on research involving many sources and methods including administrative records, interviews with people involved with child welfare as clients and providers, standardized assessments of children, and employed appropriately rigorous analyses. We want to affirm that, yes, as child welfare researchers, we do privilege research as a means of informing child welfare reform from problem formulation, to intervention development, to outcome evaluation. While our commentary focused on quantitative findings using a variety of data sources, we also affirm that research includes both rigorous qualitative and quantitative methods. Contrary to the implications of several responses to our commentary, each one of us has significant experience combining solid quantitative research, participant voices, and the interests of policy makers to advance child welfare policy. We would argue that the best path forward is to utilize the strengths of both approaches.
We wrote the Barth et al. (2021) commentary because we see a great deal of thought and energy flowing into child welfare reform and we wanted to contribute from our areas of expertise. We acknowledge, however, that some of the fields we looked at have such thin or inconsistent data that we could have been more careful in drawing our conclusions. We also wish that we had been clearer in communicating the decisions we made as we prepared the original commentary to clarify that it was neither a comprehensive or systematic review of child welfare services, nor was it an assault on particular reforms or abolitionist movements. In selecting our ten topic areas, we relied mainly on primary social science studies of child welfare, applying decades of research and/or practice experience. We drew upon experiences in working with child welfare leaders, in our teaching, at conferences, and listening to the more general child welfare chatter online. The goal was to offer a corrective to “ten commonly held misconceptions which we assert are inconsistent with the best available contemporary evidence” (Barth et al., 2021, p. 1).
Now, with seven published responses, we have endeavored to not only respond to each author, but also emphasize some overall themes. We do so with all due respect, though on some themes, we may sharply disagree. In the spirit of healthy debate, we have chosen a tone of restraint even if some of the commentaries appeared to digress from helpful conversation that can move the field forward. Disagreements and the exploration of alternative hypotheses are the hallmark of good science. We hope this debate can be allied with the dedication we all share to create a society where all families are supported and children are safe.
Wakefield and Wildeman (2022) Response
We begin with Wakefield and Wildeman’s conclusion that we failed to adequately recognize the ongoing nature of debates surrounding the relationship between foster care placement and child outcomes. We agree and acknowledge that the evidence on whether foster care is harmful is not settled, particularly for some understudied populations. In hindsight, we may have focused too much on pushing back on the unsupported conclusion of those allied with the UpEND movement and the abolitionists who had resolved that foster care has been shown unequivocally to be harmful, especially to Black children (Dettlaff et al., 2020). We stand firmly by our interpretation that the data cannot support that claim. On the other hand, there have been some studies that find some adverse effects for some groups and some benefits for others. Wakefield and Wildeman are correct, then, to “problematize the idea that near-zero average effects equate to unimportant effects” (p. 1). We appreciate their keen argument about this and encourage others to read their response. As well, we heartily agree with their appeal for further research to address the significant heterogeneity of the child welfare population, and the various outcomes that might obtain for any number of subgroups of children.
Wakefield and Wildeman also suggest that a suite of customized services that were proportionate to family needs might be a better service design than what exists in the U.S. We agree and have also argued—along with others—that we need to address adverse community environments if we are to prevent engagement with child welfare services (e.g., Barth et al., 2020; Drake & Zuravin, 1998; Font & Maguire-Jack, 2015; Garcia & DeNard, 2017; Jones Harden, Simons, Johnson-Motoyama, & Barth, 2020). We agree and have also argued about the import of addressing family needs related to poverty, greater access to evidence-based parenting services, and other supports like adult behavioral health services (Drake, Jonson-Reid, & Dvalishvili, 2022; Garcia, DeNard, Morones, & Eldeeb, 2019; Jones Harden et al., 2020; Libby et al., 2007). Wakefield and Wildeman further argue that such a suite of services might better serve as the counterfactual to foster care in efforts to assess the effects of out-of-home care. Again, we concur, though we are not aware of a robust, large-scale service design in the U.S. that researchers might turn to as such a research site. The field is lacking a rigorous test of such a suite of effective services that are implemented in accordance with a sound evidence-based process. This process should attend equally to the best available research, client preference and needs, professional expertise, and current context (Haynes, Devereaux, & Guyatt, 2002). We agree with Wells and colleagues. (2009) that this requires investment in research with diverse populations.
Wald (2022) Response
Michael Wald’s commentary offers a thoughtful response to the questions we pose relating to child neglect and—importantly—extends our work. In particular, we welcome his critique that our work falls short in offering specific implications for practice and policy. We call the reader’s attention to his suggestions as they merit serious consideration.
Two of the misconceptions we address in the Barth et al. (2021) paper relate to child neglect, its definitional boundaries, its meaning, and its implications for children. We felt compelled to raise these questions due to the rising chorus of voices falsely claiming that “poverty” is too often conflated with “neglect.” Indeed, Richard Wexler, founder of the National Coalition for Child Protection Reform, credits himself with promulgating this narrative. “We put phrases such as ‘foster-care panic,’ and ‘confusing poverty with neglect’ into the child welfare lexicon,” he writes (Wexler, 2020, para. 11). Such a misleading view has seeped into mainstream media (see, for example, Gottlieb, 2021; Schwartz, 2020). The chain reaction from “neglect is only poverty,” to “we just punish parents because they’re poor,” to a call to “abolish child neglect” (Riley, 2022) is real. We view that narrative as ultimately perilous for children and families for at least two reasons. First, it ignores the vast literature suggesting the very real harm that can occur due to neglect, offering no remedy for children with such experiences. Second, the position that neglect is merely poverty draws attention to the abolition of CPS as a solution, rather than focusing attention on ending societal and racial economic inequities.
Wald reminds us that efforts to reduce family poverty in the U.S. could have a substantial impact on child neglect and he highlights a number of studies that point to this important effect. We enthusiastically agree and as he notes, we also made this point with substantial emphasis in our recent commentary (Barth et al., 2021), as well as in a prior paper (Barth, et al., 2020). Two of our co-authors also advance these ideas in a book chapter on poverty and child maltreatment (Drake et al., 2022). Recent policy shifts related to COVID-19 and direct financial supports to families may allow some natural experiments that can help us understand how we may substantially reduce the burden of neglect related to the direct or indirect impacts of poverty.
Though enormously beneficial, would increases in family income completely eliminate child neglect? Probably not. Poverty is a powerful predictor of maltreatment, particularly neglect, but it is only one such predictor typically accompanied by a range of other parenting challenges (e.g., Clément, Bérubé, & Chamberland, 2016; Mulder, Kuiper, van der Put, Stams, & Assink, 2018). Barth and Xu (2022) studied what happens in neglecting families for 3 years after investigation and found that their self-reported adverse circumstances were far more predictive of a child’s placement into foster care than was their experience of poverty. Drawing our attention to the parental behaviors that result in harm to children, some parents with adequate income still engage in behaviors that are dangerous, whether that means voluntarily leaving a young child unattended for long periods of time, or knowingly exposing a child to other adults who sexually or physically abuse a child.
Where we focus on data which convincingly show that child neglect is not synonymous with poverty, Wald calls on the research community to clearly articulate the boundaries of child neglect. In other words, given widely discrepant state statutes, what is child neglect, if it is not “just” poverty? Child neglect definitions that increase specificity and are uniformly applied likely have promise for reducing inappropriate maltreatment referrals, reducing downstream child welfare involvement, and increasing public confidence in child protection. As we note in our commentary, “Many states try to differentiate between what may be termed, involuntary neglect—conditions associated with poverty alone—from cases that should be determined as maltreatment because parental behaviors or inaction stem from more than lack of resources” (Barth et al., 2021, p. 6; Rebbe, 2018). We argue that these policy approaches should be extended nationwide. Recent work by Feely (2022) highlights the importance of considering “modifiable” parental behaviors in child neglect assessments. Rather than focus on parental characteristics (e.g., substance-involved, experiencing a mental health condition, domestic violence, or housing instability), neglect statutes must shift toward indicators of parental behaviors that pose risks to children’s safety, taking into account parents’ available resources. Importantly, policy efforts to address the available resources part of the equation are key to alleviating the circumstances that might lead to or exaggerate conditions of risk.
Turning to the harms that result from neglect, Wald next addresses our assessment of the research literature. In line with his concerns—which we share—that current neglect statutes may be too broad, the range of harms resulting from different forms of neglect may point to different service responses. In particular, Wald draws the reader’s attention to the developmental harms that may result from child neglect, indicating that this broad category shifts child welfare from a safety orientation to more diffuse concerns about well-being. While many European nations use their child welfare system to respond to children’s compromised well-being, that has not been the typical American response (Berrick, Gilbert, & Skivenes, in press). Wald would suggest that narrowing current definitions of child neglect would more uniformly screen out reports of neglect that result in developmental harm, so that CPS could instead focus its resources on cases of neglect that pose a danger to children’s safety.
Wald offers suggestions for a range of preventive services that hold promise for supporting children’s developmental well-being, with a focus on home visitation, WIC, Early Start, and Head Start, coupled with income supports. We do not disagree, but point again to science to help assess the degree to which these family supports, alone or in combination with other services, are the most beneficial approach. For example, home visitation as a standalone intervention, despite status as an evidence-based intervention, has not shown a large effect in maltreatment prevention (Chen & Chan, 2016). As Merkel-Holguin and Bross (2015) remind us, we have a tendency to abandon innovations before we sufficiently build the science to test results alone or in combination with other efforts. Given the assertion in research that families involved in maltreating behaviors generally, and neglect specifically, may face a number of barriers to healthy functioning (Clément et al., 2016; Mulder et al., 2018), it should not surprise us that families may need a combination of preventive interventions.
What are the implications of (a) narrowing our definition of child neglect to voluntary parental behavior that compromises child safety, and (b) differentiating our concerns about children’s outcomes between physical harm and developmental harm? As Wald indicates, many circumstances that currently come to the attention of CPS suggest a public that is anxious about the well-being of the nation’s children. Too often, neighbors, teachers, and other professionals make referrals to child welfare because they are deeply concerned about a child’s well-being, and because—importantly—there is no one else to call. We concur with Wald that any approach to addressing child neglect must attend to narrowing the definition for intervention while also attending to the provision of economic and parenting supports outside the current reactive system. In concert with Wald, we also want to advance the importance of developing a broad platform of financial security, coupled with a service system that helps parents struggling with parenting, substance abuse, domestic violence, mental health, and housing insecurity. This would likely address a large proportion of the families who currently come to the attention of CPS, and would respond to the developmental opportunities that children need and deserve. Going further than Wald, however, effective interventions that promote child well-being in the face of adversity also deserve attention. In other words, we argue that children and families must receive effective interventions following maltreatment while we also work toward advancing prevention science. Location of such approaches either external or internal to a re-visioned child welfare system is paramount to effective approaches to child neglect.
Font (2022) Response
Sarah Font (2022) underscores the need for critical examination of core assumptions and awareness of the characteristics and outcomes of today’s child welfare services. We appreciate her caution against mischaracterizing and overgeneralizing existing research findings. She also showcases how challenging it can be for practitioners and policymakers to interpret and process the rigor of research, and ultimately to use research to reform the child welfare system. In particular, she showcases examples where the nuances of child welfare research findings have been lost to over-simplified narratives that distort and sometimes mislead (e.g., some findings relating to kinship care and to material supports for families). Based on Font’s response, our summary of ten common misconceptions may have been too limited.
Importantly, Font emphasizes the difference between research and meaning-making, with a particular concern about the “abolitionist” movement’s misuse of data and mis-interpretation of findings on neglect and poverty. She posits that the standard we assert to simply “consider the evidence” is insufficient for evidence-based reform. Instead, she challenges us to conduct the types of analyses that make convincing meaning from research. We heartily endorse her arguments, especially that policy and programs based on the misuse of research are no better than policymaking that considers no evidence at all.
We appreciate her critique of the path we took: To address ten rather complicated child welfare issues was not without the risk of oversimplification. To that end, we heartily endorse her conclusion that “now the scientific community must forge a path forward that emphasizes clear and cautious communication with the public” (Font, 2022; p. 2).
Tajima, Day, Kanuha, Rodriguez-Jenkins, and Pryce (2022) Response
The Tajima et al. (2022) paper departed significantly from the prior three responses. Of concern, there are a range of positions which they claim were present in our original article that were not. Also of concern was the apparent misrepresentation of scientific data and reliance on opinion pieces. While researchers are human and may sometimes omit an article through oversight, this is different than choosing only those pieces that agree with a particular point of view. Perhaps most troubling, was the tenor that research should be disregarded in favor of personal or political opinion. Evidence-based approaches to addressing significant social and public health problems do take into account both the preferences of populations served and clinical expertise, but equally value the best available science to guide the application of the most effective approaches (Evidence-Based Medicine Working Group, 1992; Haynes et al., 2002).
Tajima et al. state that “Despite token statements of internal critique, the authors are clearly comfortable with the child welfare system as it stands” (p. 2). We found this assertion to be deeply troubling as every one of us has worked tirelessly, along with many other researchers, to improve child welfare services. We have published, between us, scores of books and hundreds of articles and most or all have concluded with suggestions about what we can do to make child welfare services markedly better. We have proposed many innovations in services and policies, some fairly radical and relying on quantitative, qualitative and mixed methods approaches (e.g., removing “substantiation” as a finding category to promote services linked to need; limiting group care for young children; shared family care to allow parents and children to live together in foster care; optimizing organizational agency contexts to support caseworkers so they, in turn, can deliver culturally relevant, evidence-informed child welfare interventions; improving medical and mental health care for youth in foster care; removing barriers to community services like home visitation; developing curricula to strengthen natural mentoring, eliminating child support obligations for child welfare-involved families, and more). We would argue that our children and families deserve healthy debates and a serious review of the available evidence rather than divisive rhetoric.
Tajima et al. further state that we claim that “most culturally based child welfare interventions have poor or neutral outcomes” (p. 4). Their statement is simply untrue as we never make this claim. Instead, we focus on the importance of offering evidence-based services to diverse families and reference several research-supported practices that have been found to be effective for and acceptable by families of color (Parent-Child Interaction Therapy, Trauma-Focused Cognitive Behavioral Therapy, Level 4 Triple P [Pathways Positive Parenting Program], and Multi-Systemic Therapy). We allude to increasing attention to both cultural adaptations and novel interventions that will undoubtably offer new opportunities to better serve diverse communities and are encouraged by the emerging body of work in this area.
Tajima et al. go on to assert that “By suggesting that the passage of the Indian Child Welfare Act (ICWA) has been successful for American Indian and Alaska Native (AIAN) children, Barth et al. fail to acknowledge that Indigenous children continue to enter care at a higher rate than any other racial/ethnic group…. and remain at high risk of being separated from their families” (2022, p. 3). We are deeply dismayed by the authors’ misleading portrayal of our writing. Nowhere in Barth et al. (2021) do we suggest that ICWA has been “successful for American Indian and Alaska Native (AIAN) children” (Tajima et al., 2022, p. 3). To be clear, in offering readers examples when research “contributed significantly to ending negative practices and building new supports” we wrote: “For example, data from studies conducted by the Association of American Indian Affairs on egregious removal and adoption practices laid the foundation for the Indian Child Welfare Act in 1978 which established the rights of federal tribes to be engaged in all aspects of placement decision making for eligible native children (Mannes, 1996)” (Barth et al., 2021, p. 1). We do believe that ICWA offered a substantial improvement in federal child welfare policy vis-à-vis American Indian and Alaska Native children and families, but we hardly characterize the policy as a success in addressing the multiple challenges tribes and tribal families face in the context of contemporary child welfare policy and practice.
Tajima et al. also state that we “describe the child welfare system as ‘comparatively young’” (p. 4). In reality, we wrote “Clearly, there is much more work required, but the
Tajima et al. write that, “These studies lead us to question the assertions by Barth et al. which conclude that studies conducted utilizing state administrative data are the ‘gold standard’ in research practice” (p. 2). Although the quotation marks used in this sentence suggest that these were our words, we never used the words “gold” or “standard” in our paper. We would not have made such an assertion, in any case, because there is no single standard for what is best in research—it depends upon the question being addressed. Science is improved when researchers engage with data from multiple and diverse sources, and lead with facts.
Tajima et al. allege that we believe “Western-based and defined empirical, scientific ‘evidence’ is the best and only valid research” (p.1). We wonder, why are there quotation marks around the word “evidence?” We are also concerned with the use of the (seemingly) pejorative epithet, Western, which apparently is synonymous with quantitative work, despite the long history of Western researchers who have been in the forefront of qualitative work. We have a more inclusive view of research methods and their historical traditions. Qualitative and mixed methods research are an important part of the scientific literature with standards of rigor and training just like quantitative work (e.g., Gill, 2020; Harrison, Reilly, & Creswell, 2020; Mackieson et al., 2019; McFarlane, Occa, Peng, Awonuga, & Morgan, 2021).
This issue circles back to the title of their response paper: “What counts as evidence in child welfare research?” In general, we subscribe to the position that research methods should match the particular question of interest. So, if our question is “Are young children entering foster care likely to grow up in foster care?,” then this is a question that administrative records can best answer. Our commentary addressed this question. If the question is, “What are the experiences of parents whose children are removed from home and placed into foster care?” then there are no administrative data or representative survey data on this question. The most representative survey ever done on child welfare [NSCAW] did not include this question and we did not try to answer this question in our commentary. Qualitative studies might be well suited to answer such a question and some of our authors have previously attempted to offer partial answers (Berrick, 2009; Garcia, Ohene, DeNard, Morones, & Connaughton, 2018).
Moreover, we do not claim that research is a standalone metric upon which policy should be based. Our paper is, in fact, replete with statements such as “reforms…should be informed by a wide array of stakeholders together with reliable data” (p. 1), and “Reform can be especially effective when empirical data are coupled with the views of a variety of stakeholders. This includes acknowledging the pain of parents who were, or who have felt mistreated, and especially those who were unfairly separated from their children” (p. 2, emphasis added). Tajima et al. write: “To suggest that the only legitimate ‘evidence’ emanates from standard Western measures and methods selectively cited by Barth et al., discounts the wealth of deep and expansive data generated by community studies, collaborative and participatory methods, case studies, field work, storytelling and story collecting, archival and historical sources, and participant observer/observation in the qualitative tradition” (p. 2). In our view, multiple methods bring important value to the research enterprise. Storytelling is important, as is science, but they serve different purposes. Storytelling helped to inspire our dreams to go to the moon, but science launched the first rocket. The COVID-19 vaccine was developed because of science; storytelling undermined its universal delivery. In our paper, we argue for a both/and approach, suggesting the value of quantitative and qualitative data to shape our understandings of micro- and macro-child welfare policies and practices.
This brings us to our concern that Tajima and associates misunderstand or misrepresent data and sometimes rely upon sources that are not scientifically grounded. Tajima et al. write that “there is overwhelming evidence that structural and institutional racism is implicated in the entry and treatment of BIPOC children and families at every stage of the child welfare system” (p. 4). They proceed to cite articles that purportedly demonstrate that there is raw disproportionality in child welfare (e.g., Crofoot & Harris, 2012)—an obvious fact that we do not contest in our article. They also cite literature reviews and legal reviews that are deeply flawed and frequently confuse correlation and causality (Dettlaff, 2020; Dettlaff & Boyd, 2020; Hill, 2004; Merritt, 2021; Raz, 2020; Roberts, 2002). Exemplifying their tendency toward catch phrases and away from precise scientific critique, the first article Tajima et al. cite (beyond ours) actually refers to music education (Hess, 2018), not child welfare. These issues extend to some of their data presentation. In their Figure 1, we see that both Hispanic and Asian children (who currently comprise 62% of all BIPOC children in the United States (Federal Interagency Forum, 2021)) enter foster care at rates lower than White children, precisely the opposite of their “every stage” claim (pp. 3-4). Figure 1 is also remarkable in that it is administrative data unleavened by lived experience. This, despite their desire that “researchers and those with lived experience share equal power in the research process” (p. 3). We agree that there are many times when a careful analysis of such data can be instructive.
We share Tajima and associates’ concerns about racial/ethnic disproportionality in child welfare, but we come at the question from a position of concern and curiosity. We fully accept that understanding the process and impact of child welfare services requires accounting for the structural racism that pervades society and we noted this plainly. Nevertheless, we reject the idea that we can, therefore, simply claim that all child welfare interventions are unfair or harmful to BIPOC children. Decades of research fails to support this claim (Courtney et al., 1996; Barth et al., 2021). This is a dangerous over-generalization that belies the specific evidence of how children of different ages, races/ethnicities, and genders are actually doing.
Tajima and associates’ argument about racism and child welfare includes a lengthy discussion of how decisions are made and how they reflect biases and stereotypes. We do not contest this point—we acknowledge that everyone, regardless of identity and positionality has biases, that these biases are varied, and that most professionals do their best to engage in reflexivity to recognize and address them. Indeed, we know that a common and serious problem is that children of color often do not get access to the services they are entitled to in health, education, and early childhood care. We must both optimize the societal and organizational contexts and capacities to deliver services fairly and assure that professionals are well-trained and capable of engaging in reflexive practice. These indeed are empirical questions—what are the social and organizational conditions necessary to deliver services in a way that give fair access and opportunities for success, and how might reflexivity play a role in achieving this goal? Science cannot progress without fact checking of ideological claims or grand theories. The national COVID-19 experience provides timely evidence of what happens when sound science is ignored.
We certainly believe that client/user involvement in program design, implementation and evaluation is critical, as do all social work researchers we know. In fact, researchers in notably different fields hold similar values. Medical researchers and engineers typically listen to the people they serve and incorporate their views and wishes. The basic tenets of evidence-based practice indicate that interventions should be based on the research evidence, professional expertise, and client values (Gambrill, 2019). Medical procedures should be beneficial and acceptable to clients; so too should physical infrastructure, such as bridges and roads, be safe and appealing to their users. Activists with lived experience created the invaluable impetus for accelerating research on HIV—but getting to zero on HIV also requires the most rigorous possible clinical and translational science.
The entire point of our paper was to provide an overview of the best available evidence with regard to particular key issues in our field. Classically, scientists with opposing points of view would review the evidence we presented or contrast it with new evidence. Tajima et al. do not address the evidentiary sources we use directly. Beyond broad paradigmatic claims about the utility of lived experience, they do not articulate arguments or cite research that arrives at different conclusions—instead, they offer only contradiction.
Tajima and colleagues also appear to question the credibility of the authors sharing the data. We strongly reject their judging the quality of our argument based on any inference about the race or ethnicity of the authors, or assumptions about any of the authors’ lived experiences. We wholeheartedly agree that there is a need to expand the number and diversity of child welfare scholars and all of us are engaged in various efforts to grow the pipeline of scholars of color. The implication that authenticity as a scientist stems largely from identity or “lived experience” (an altogether ill-defined term) is problematic. We wonder if “lived experience” is evolving into a kind of credential of credibility for our field. If so, is this credibility required of other scientists? Should diabetics stop taking insulin because Frederick Banting (insulin’s inventor) did not experience diabetes himself? It also inherently disregards the enormous diversity within the BIPOC and lived experience communities. Should we disregard research from a Native scholar if they are not from the same tribe as research participants? Should research on undocumented families be conducted only by a researcher who is undocumented?
We agree with Tajima et al. that collaborative child welfare research can strengthen our work. Many of us have been involved in qualitative studies that amplify the voices of those with lived experiences, or we have endeavored to reflect our work back to constituents. But the utility of lived experience depends on the question at hand, and the representativeness of those chosen to stand in for the vast population of individuals with lived experiences. An example from one of the authors may illustrate a potential pitfall that has been experienced by other authors. Having participated in many expert panels where individuals with lived experiences were included to inform the national context of child maltreatment, the composition of these panelists with lived experience has been remarkable. They were, not atypically, 100% urban, about 80% Black and about 80% of the contributors had experienced foster care. This population (Urban, Black, experience as foster youth or foster parents) represents a single-digit percentage of those with lived experience of the child welfare system. The vast majority of the people served by the child welfare system in the U.S.—who are white, living in semi-urban or rural areas, screened out, given alternative response, or received in-home services (U.S. DHHS, 2022; Wulczyn, 2020)—were effectively silenced by their lack of representation in these groups. We do not make this point frivolously but as a warning. Bedrock principles of “Western” science (e.g., generalizability, triangulation, replication) apply to all forms of knowledge generation. Finally, embracing the value of identity or lived experience of a researcher as a way to exclude voices that may offer other insights into improving outcomes risks engaging in the same practice of “othering” that may result in unjust and uneven distribution of resources or other forms of discrimination (e.g., Dionne & Turkmen, 2020). We should probably exercise caution so that we do not use the language of lived experience as an opportunity to exclude voices that may offer an alternative source of evidence or perspective.
We hope our summation of our response to Tajima and colleagues both dispels some of the false statements and misrepresentation of extant scholarly work and points to areas in which we can engage in continued dialogue. The scientific process is enhanced when scholars and advocates who hold varied epistemological standpoints and perspectives engage in thoughtful debate. However, we believe the debate must be evolve from the latest empirical evidence.
Schelbe and Geiger (2022) Response
Schelbe and Geiger offer a number of statements that correspond to Tajima et al., in their zeal for the inclusion of lived experience in shaping all aspects of child welfare research, policy and practice. We are sympathetic to the nothing-about-us-without-us perspective. As we understand, the origins of the term go back hundreds of years and suggest the seeds of what we later came to know as democracy (Charlton, 1998). We all agree that parents and children/youth should have a say in the services they receive and child welfare workers should have a say in the procedures and practices of their organization. We want foster parents to be included in case planning and to be able to determine and use prudent parenting practices. And research can be improved with the participation of stakeholders from the community who have experienced the phenomenon under study. We agree that all of these actors should have their views taken into account in developing policy, as they have and should be in many areas of child welfare policy at the local, state, and national levels.
Schelbe and Geiger indicate that they agree with us; that reforms to child welfare should be “informed by a wide array of stakeholders together with reliable data and the best available research evidence” (Barth et al., 2021, p.1), but also urge any thinking about re-designing the child welfare system to include authentic representation of stakeholders and to use collaborations that do not privilege research over the knowledge of those with lived expertise. Nowhere do we suggest that these are either/or possibilities. We offered our summaries of ten research areas that we believe to be very often misunderstood so that there is a better basis for adding to the knowledge base. But Schelbe and Geiger point to the limitations of research—limitations that we, of course, acknowledge—and therefore assert, “For this reason, a research-driven agenda is not the solution” (p. 3). We would encourage readers to go back and read that last line again. This position defies everything we know as social workers regarding how to use data to drive policy change and innovation, including the ethical commitment to our profession to “monitor and evaluate policies, the implementation of programs, and practice interventions” (NASW, 2021). A “research-driven agenda” for child welfare does not a priori exclude the expertise of individuals with lived experience in developing such an agenda, nor does it exclude such individuals from driving the change process from the outset. We are disheartened by this statement in Schelbe and Geiger, given the great strides that our field (and others) has made by using evidence to inform policymaking.
We never asserted that evidence and research should be the only factors in policy decisions. The goal is to give credible evidence and data (rather than misconceptions about data) prominent seats at the table when decisions are being made, a table that also includes seats for individuals with lived expertise, practitioners, and other key stakeholders. Could researchers do more to include service-users in the design, implementation, and interpretation of research? Of course. But absent those voices, and setting aside the vital question of how one selects a representative sample of those diverse voices, do we discredit research findings altogether if a service user wasn’t involved in some way? Is that the new litmus test by which we judge whether a research finding is believed? Many of us took the vaccine for COVID because we trusted in the science behind its development. Too many who preferred to listen to the lived experience of their neighbors or pundits who did not get the vaccine, and who had a lived experience of perfect health, refused to take a vaccine in the midst of a pandemic. Many have died.
Schelbe and Geiger argue that we have “oversimplified what is known about child welfare through [a] presentation of ‘misconceptions’ that appear to be randomly selected” (p. 3). We understand that we did not sufficiently identify the basis for selecting our ten identified areas—which were identified based on our beliefs about what seems to be misunderstood in our field. We welcome other scholars calling attention to additional areas of concern and debate (e.g., Font, 2022) so that we can bring to bear the best science to address areas critical to improving policy and practice.
Schelbe and Geiger point out that knowing what “on average” is occurring does not necessarily assist local communities in addressing their specific concerns. We appreciate these views and agree that child welfare-involved families, local child welfare systems, and individual caseworkers differ in many ways. Capturing the average experience across populations does just that—it speaks to the average, but it masks important variability within populations. We agree that we could have done more to highlight this point—also made by Wakefield and Wildeman (2022). We would add that there are many levels of policymaking, from the national, to the state, to the local. Communicating an “average” experience offers a reference point that is not inconsequential and should be considered when large-scale policies are under review; highlighting the diversity of small-group and individual experiences is necessary in effectively implementing community-level practice.
On other matters, Schelbe and Geiger mischaracterize our work. They write that “In no way do we wish to convey that the answers are easy or that we have the solutions, however, we want to challenge Barth et al. (2021) and others who dispute racism within child welfare based on the fact that large quantitative studies do not consistently detect it” (p. 2). Nowhere in our article do we suggest that racism does not exist in child welfare. Further, we make it clear that many vulnerabilities experienced by Black families in the U.S. are caused by “racially based economic stratification” (Barth et al., 2021, p. 4). Elsewhere in our writing we extensively address our concerns about structural racism and the effects racism has—in particular—on Black children and families (Barth et al., 2020). In Barth et al., (2021) we write: “By no means do we mean to minimize the profound wrongness of economic disadvantage facing Black children--we instead wish to shine a light directly upon it. Racially based economic stratification is a profound and crippling societal flaw which must be addressed” (p.4). To improve outcomes, we need to use evidence to understand the leverage points for change. Both the ideas that racism may operate in a given agency and that it is not the primary driver of an outcome can be true. It can be equally true that racist structures in society continue to perpetuate disadvantage and segregation that harm families and that the forces requiring change may be largely located outside child welfare. We need to bring to bear the best science to understand how our country’s undeniably racist history and structures have impacted and continue to influence current child welfare processes.
Finally, Schelbe and Geiger argue that our commentary “feels like giving excuses for not doing better” (p. 4). As researchers, we can never control how our words will be perceived by the reader. We can make every effort, however, as we had intended, to be circumspect in our writing about what is known and not known, to be humble about the tremendous work ahead, and to acknowledge the importance of research being used in partnership with individuals and communities to shape policy.
Nowhere in our article do we argue that the status quo is acceptable. Indeed, the title of our paper speaks to a reform agenda. Nowhere in our article do we argue that research should be used in the absence of other sources of knowledge. Neither do we dispute the value of many ways of knowing. Our point was to try to select a collection of topics that are too often tossed into child welfare discussions with little regard for their veracity and about which we had deep knowledge. Clarifying what we know, what we think we know, and what we need to know will help as we all work to improve the systems that interface with families who we all care about very deeply.
Cabin (2022) Response
Our response to Cabin is brief as the majority of the response seems to reflect a misunderstanding of our work and the data underlying our paper. Cabin’s first point is that we failed to employ a systematic review, and therefore that we did not do a “professional literature review” (p.3). The suggestion that only systematic reviews can be acceptable as a means of reviewing literature struck us as peculiar. There are a number of types of reviews that address differing questions (e.g., Moher, Stewart, & Shekelle, 2015; Munn et al., 2018). As others have noted, we could have done a better job explaining our selection of topics. We did not sufficiently clarify that we were relying on primary research sources and that we are working in areas which closely match our specialty areas of research within child welfare. That being said, that does not make our chosen form of review unprofessional.
Cabin chastises us for making (clearly cited) claims that abolitionists want abolition, for not discussing the range of suggestions for change that those “in the movement and separate from the movement” (p. 4) have forwarded, and for not defining their major theses. We are taken to task for not addressing the work of Dorothy Roberts or Mical Raz. This is true, and would have been a substantial failing if the purpose of our work were to describe the abolitionist movement, but this was not the point of our paper. We attempted to reference the movement to contextualize the article in a time of reform, not to provide a critique of the abolition movement.
As a general comment, the extensive use of legal citations in the Cabin article highlights something worthy of attention—the schism between the legal and the scientific literature in our field. The commentary and consideration of legal structures from faculty in law schools, and to some degree, historians, is something which currently exists largely in parallel with, rather than in integration with, the scientific literature. It would be desirable if these two strands of literature, which both are powerful with regard to child welfare policy, could be more closely integrated.
Southall (2022) Response
Southall offers a perspective from a practicing CPS worker and helps to frame the related deliberation about our original paper. We cannot know if her response is representative of many practitioners, but this would not surprise us, if true. Southall does not directly grapple with any of the specific misconceptions initially posed and, instead, provides a general view about how practitioners may think about the applicability of research. She does posit that many cases come to CPS for reasons that are primarily about poverty although also recognizes that living in poverty also increases the likelihood of other entanglements that adversely affect parents and children. Southall offers a view of research that is adaptive and context specific. That is, her commentary suggests that research may only be useful to practitioners if it fits the experiences that they regularly encounter in their caseload. This relates to our previously offered comments that conversations about national child welfare dynamics always need to be supplemented by local analyses, but this is clearly not enough. If the dominant view of child welfare research is that there is no imperative to use it, and it is largely irrelevant—depending on your beliefs about how it fits the local context—this reduces our prospects for improving outcomes. We hope to see more conviction and talent applied to answering questions at the local level but this will require that child welfare agencies are viewed as capable of generating and using research. As LeCroy and Kaplan (2022) find, the public believes that science is needed to advance progress, but does not believe that social work is really amenable to science. Southall’s response underscores their findings, pointing to the limited adoption of science and impact on practice in the current context.
Despite this constraint, we remain hopeful that there is a common interest and concern among policymakers, researchers, and advocates about the importance of scientific knowledge as a defense against misconceptions, misunderstandings, and outright fallacies in child welfare narratives. As reforms continue apace, child welfare practitioners and—importantly—families will ultimately benefit.
Concluding Note
Active efforts to change child welfare are ongoing but take time. We are heartened by efforts led by NICHD funded training institutes and the Doris Duke Foundation that are expanding the number of researchers both in child welfare and closely allied areas of child maltreatment and child well-being. We also know the work of many of the commentators who contributed to this issue and how busy they are shaping a society that is more protective of the rights and needs of children and parents. We value the chance to exchange ideas with them. We think that some of the topics raised by these commentaries are among the most seminal topics for the future of our field. Although our response may sound a little prickly at times, we believe that some points are quite worrisome on their face or by implication and need to be sharply challenged. Nevertheless, we thank all those who stepped forward and we look forward to joining forces to improve child welfare services quality and outcomes.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
