How could behavioral design improve frontline child welfare practice?

Behavioral design and child welfare

Behavioral design is a methodology that combines the thoughtful person-first approaches of human-centered design with rigorous scientific insights and iterative testing through impact evaluation. In 2015, ideas42 introduced a set of three behaviorally informed principles for human service professionals to use in service, program, and policy design: Cut the Costs; Create Slack; Reframe and Empower. As articulated in the ideas42 research report Poverty Interrupted, these three core design principles offer a roadmap for evidence-informed practice improvement for systems serving people in poverty. Program administrators, policymakers, and other social impact professionals can use these principles to design for the context of poverty (which can be understood as the context of chronic scarcity), and address systems-level barriers like racism and xenophobia.

After the publication of Poverty Interrupted, ideas42’s Economic Justice team began operationalizing the design principles through field-tested interventions (many of which incorporated random-assignment impact evaluation) on an array of programs including WIC (Special Supplemental Nutrition Program for Women, Infants, and Children),² SNAP (Supplemental Nutrition Assistance Program),³ and public housing.⁴ Recently the team began working on extending the findings from Poverty Interrupted to the context of child welfare. Since early 2020, ideas42 has worked with Casey Family Programs and a set of state child welfare jurisdictions to uncover insights about where behavioral science may advance day-to-day child welfare work, and create practical design ideas using behavioral insights and the Poverty Interrupted lens. This brief shares an overview of some of key reflections and considerations. 

Scarcity and bias in child welfare

Scarcity and its effect on cognition

The primary construct explored in Poverty Interrupted is the condition of scarcity, which can be understood as limitations on any key resource such as time, energy, or money. Resource scarcity affects people’s cognitive capacity because their minds “tunnel in” (focus intensely) on acquiring the things they urgently need. As a consequence, their minds also simultaneously tune out other needs that may be less urgent, but still important. This intense focus can be helpful to meet goals related to those resources in the short term. However, it becomes problematic when people are addressing emergencies constantly and must continually make tradeoffs between emergencies and other important needs. Anyone experiencing scarcity of any kind faces this “tunnel” dilemma and its consequences when those important (but less urgent) needs go unaddressed.

While Poverty Interrupted includes a variety of recommendations for programs and services to help families contend with the effects of scarcity, ideas42’s newer work in child welfare has allowed for an examination of the parallel dynamic of scarcity’s effect on staff. When child welfare staff face scarcity and its associated consequences, the families they serve suffer as a result. This brief will focus on opportunities to design for the context of scarcity affecting staff as a pathway toward transforming the child welfare system and the ways families experience the system.

Staff’s experience of time scarcity

One of the biggest sources of scarcity for child welfare staff is time. Given the need for child welfare staff to provide both unbiased decision-making and proactive holistic support for children, parents, and caregivers involved with the child welfare system, the cognitive effects of time scarcity are worrisome. Consider the following scenario: A hotline worker receives a call about alleged neglect. The scenario sounds similar to a report from earlier that day where the case was assigned to investigation. The worker feels uncertain about the current case, but there is a queue of 15 other calls waiting. To compound the situation, the worker is nearing the end of the work shift and already has worked through lunch and break. 

This scenario is far from unusual in child welfare. What is the correct choice for the screener to make under these conditions of time scarcity and stress? Using evidence from an earlier case to assess this family’s case is neither fair nor ethical. Yet spending additional time may run the risk of exacerbating the call queue, or even delaying the response to a serious case of maltreatment. Accurately assessing each child’s safety and risk individually is important but people tend to be more risk averse and susceptible to bias when they are task-overloaded.⁵ And in the face of looming ambiguity, screening in feels safer than screening out. While avoiding risk might sound favorable on its face, it often results in the over-surveillance and over-investigation of families. That increased intervention leads to trauma, family separation, and worse outcomes in the long-run — the opposite of the aims of a child welfare intervention.⁶ These persistent challenges in trying to do what’s “right” in a resource-constrained environment are driven by the effects of scarcity. Being overworked and having to weigh complex risks and tradeoffs constantly lead not just to suboptimal decisions, but also burnout and turnover.⁷ Ultimately, this ends up hurting those on the receiving end of these decisions: children and families, particularly those disproportionately impacted by child protective services, which include children and families of color and those in a lower socioeconomic stratum.

Scarcity activates implicit bias

Unfortunately, scarcity also activates other biases. Implicit bias — or unconscious prejudicial attitudes or stereotypes toward specific social groups — is troublesome for child welfare, particularly along the lines of race/ethinicity and socioeconomic position. The literature indicates that several conditions make us vulnerable to implicit bias: having to make quick decisions, being angry or anxious, or being confronted with highly salient social categories like race or socioeconomic position.⁸ Of course, these are the exact conditions faced by child welfare staff every day. 

Let’s return to the earlier scenario with the hotline screener. While the staff person is feeling the time pressure to make a choice about screening the family in or out, the worker may begin to judge the family by characteristics that should not be relevant but feel intuitive or come to mind quickly. Often these characteristics are indictors of race or socioeconomic position. For example, the address the report came from was in a middle class neighborhood, or the worker has a “gut feeling” about the family because the family’s surname is familiar from a previous contact with the system and the worker figures that the “apple doesn’t fall far from the tree.” Without even consciously recognizing it, these implicit signals may become the deciding factor for whether a family is subject to investigation or consequent actions. 

It is critical to understand how much room child welfare staff have to make these judgements in order to understand how much bias may exist. While hotlines were the focus of the earlier scenario, implicit bias may affect a variety of contexts, including screening, treatment, removal, and reunification. Left unmitigated, bias leads to disproportionate outcomes for non-white families and people living with low incomes.⁹ While the evidence is alarming, the call to action is clear. Policymakers and program designers who care about addressing disproportionate outcomes and avoiding downstream harms to families have an important opportunity to apply behavioral design to reduce the effects of scarcity and address bias in the child welfare system. 

Poverty Interrupted design principles adapted to child welfare staff 

What might behaviorally informed child welfare system improvements look like? The three design principles of Poverty Interrupted can be applied to child welfare to address the time scarcity faced by staff and the resulting exacerbation of bias. The first two principles, Cut the Costs and Create Slack, directly address the context of scarcity from different ends. Cutting costs for staff means making needed resources easier to get, thus reducing the burden on their scarce time and attention. Creating slack for staff is the flip side; they must be given more of what they lack, and when the issues can’t fully be addressed, it is critical to ensure the right supports are in place to reduce the negative consequences. For child welfare staff, creating slack largely involves finding ways to give them time back or reduce stress. The third principle, Reframe and Empower, speaks to how to mitigate the various biases that creep into child welfare. Below are a handful of examples that suggest how practitioners and policymakers might apply these three principles to assist child welfare staff and transform the system overall.

Cut the (administrative) costs faced by child welfare staff. 

Child welfare staff make complex decisions and take actions with limited time and ample caseloads. These decisions have high stakes. Cutting the costs involves finding ways to reduce the hassles, complexities, and associated stressors that child welfare staff face in making decisions to ensure better outcomes for children and families. Below are a few design considerations for reducing hassles for child welfare staff: 

• Streamline data entry processes. By integrating and automating data systems, staff will be required to spend less time on important yet cognitively demanding administrative tasks. Instead, those attentional resources can be allocated toward helping families meet their needs and keeping children safe. 
• Make robust referral systems and/or prevention pathways easy for staff to navigate and use. Related to the above recommendation, it can be taxing on staff to have to navigate complex systems of referrals and prevention pathways. Jurisdictions with differential response mechanisms have implemented different approaches toward this problem, including establishing partnerships with community-based organizations specializing around a particular need (such as housing or food insecurity) or re-directing families to a vetted set of accessible family resource centers that are close to them. These systems should be designed in such a way that minimize costs for families and staff. 
• Set up easily accessible prompts to guide cognitively consuming tasks. For example, hotline workers must recall all the different data points they need to gather from a call to determine whether an allegation fits the definition of a specific maltreatment criterion. They need to do this while listening to often emotionally weighty information, writing down all of that information, and keeping the conversation focused on the allegation at hand. Simple on-call elicitation prompts or probes can guide the important task of getting the information workers need from a call. Measures like this to cut cognitive costs can impact both the quality of the data gathered as well as subsequent decision-making based on it.  

Create slack where possible, but design for the presence of scarcity and associated bias.

On the flip side of finding places where staff time can be cut down, it’s important to also find places to offer staff greater slack around things they need, like time, and when that’s not possible, design intentional fallback supports.

• Reduce caseload volume by hiring more staff. When the caseload is spread across more staff, time scarcity will be improved for all, which also may address associated bias. It is important to recognize, however, that jurisdictions are resource-constrained and that this is not always feasible.
• Create safeguards for child welfare decisions. Safeguards can play an important role in situations where reducing caseload is not possible. For example, in the context of hotline screening, providing staff an opportunity to indicate how “certain” they are about their decision — through, for example, a question on the intake questionnaire that rates decisions from “highly certain” to “highly uncertain” — can flag places where ambiguity or scarcity may be playing a role. Staff then can indicate where they need additional support from a supervisor or where further departmental review may be required, rather than escalate the case to the irreversible process of investigation.
• Reduce or remove artificial constraints. Staff also may be working under time pressures enforced by department leadership, leading to increased cognitive demand on already scarce attention. For example, supervisors monitoring call durations and after-call case processing times (two important quantitative metrics in call center operations) may further compound suboptimal decision-making patterns by inadvertently sacrificing information quality in favor of timeliness. Replacing such supervisory monitoring with in-system prompts or reminders that convey the same information in a less pressure-inducing fashion can help create slack. 

Confront staff biases, and foster authentic partnerships with families. 

Child welfare staff want to feel good about their role in helping address instances of child maltreatment and supporting families holistically, but need the tools to actually do so. Additionally, authentic partnership with families is critical to reimagining the child welfare system from top to bottom.

• Confront staff biases. Everyone is subject to bias and it can have a powerful, harmful effect. Agencies must do the work to uncover where bias affects outcomes (such as assessing disaggregated intake data), and address it accordingly. In child welfare, interventions to address bias might include requiring that multiple decision-makers independently agree on a decision or action during the screening process, or instituting time lags for reflection before finalizing key decisions that affect children and families.
• Provide clear guidelines where there is ambiguity. When hotline staff rely on case histories to make decisions on whether to investigate, it can compound already existing racial and socioeconomic disparities. People who are living with low incomes often have deeper recorded histories in the child welfare and criminal legal systems. Providing clear guidelines for reliance on case history can therefore mitigate the biasing effects of using lopsided data. Examples of such guidelines include a ‘3 Rs’ system of ‘Relevance, Recency and Result’ that provide clear recommendations for what case history data may be used, or even be made accessible, to decision-makers. 
• Reframe the role of child welfare staff and foster family partnerships, and amplify family voice. Many staff view their role as “child protectors,” which exacerbates the biases in their work. Work must be undertaken to leverage more positive identities in the child welfare system, such as helping staff see themselves more as public servants and families more as authentic partners and true experts in promoting the well-being of their children. Staff may be primed or instructed to use language that is intentionally framed (for example, “our agency’s priority is to support and preserve families”), and should consult family members in the decision-making process (such as through a family voice council or lived experience steering committee) to advocate for the needs of families impacted by these systems.

Conclusion

While neither the diagnoses nor the design ideas presented above are exhaustive or conclusive, they represent a range of interventions that can be empirically tested in the field and refined through iteration rooted in evidence. Committing to better outcomes for children and families requires innovative practices and the evidence for which innovations work at producing equitable outcomes. Find out what works, how well, and under which real-life conditions: that is the first step toward using behavioral science to craft sound policy and practice, and toward scaling up effective practices that will help achieve the well-being that all children and families deserve.