How can the collection, analysis, and reporting of demographic data support children and families?
For child welfare systems to design effective and equitable solutions that support families upstream, accurate data about who the system impacts most are essential. Collecting, analyzing, and reporting demographic data by race and ethnicity, gender, sexual orientation, and other variables help systems better understand any disparate outcomes from their programs, and then use what they have learned to better serve the needs of children and families moving forward.1
National data demonstrate disproportionality within the current child welfare system for both Black and American Indian/Alaska Native children beginning at the first point of contact. This overrepresentation occurs even when controlling for other characteristics. Similarly, youth who identify as lesbian, gay, or bisexual are nearly 2.5 times more likely to be placed into foster care than heterosexual youth. Child protection agencies have a responsibility to understand data trends at every decision point and determine how existing and proposed practices and policies may contribute to disparities within the child welfare system.
Commitment of staff at all levels, involvement of the community, funding for the implementation and monitoring of effective interventions, and collaboration within and across systems also are necessary to understand the current landscape and develop targeted solutions to address disparities.2
Guidelines for collecting meaningful data about people and programs
Communities should inform the methods and categories used to track demographic data. Categories need to be nuanced enough to accurately describe people. Caseworkers require guidance and clear expectations around accountability, and data systems must be nimble enough to record, revise, and report nuanced categories.
Move beyond standard categorization
Longstanding and widely used federal Office of Management and Budget (OMB) standards include five race categories (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Other Pacific Islander, and White), and one ethnicity category (Hispanic/Latino). The OMB classifications, however, often are too broad to be meaningful. “A lot of people don’t know how to identify because they don’t fit in current OMB categories,” said Ali Jawetz, senior policy analyst at the Center for the Study of Social Policy. “For example, people from Jordan, Turkey, and Lebanon are categorized as white, and Hispanic/Latino people need to identify a race in addition to their ethnicity, but they may not identify as one of the five broad race categories.”
Important distinctions within subgroups are not made, such as within Asian populations or among American Indian tribes. Details about the specific backgrounds of multiracial people also are not included, and often only a primary race or ethnicity is used in any analysis of disproportionality and disparities. All data collection systems should allow more than one race/ethnicity to be recorded. Whether a child qualifies for support under the Indian Child Welfare Act also should be carefully tracked.
The Putting Pride into Practice project recommends tracking gender identity for children 3 and older, including options for male, female, transgender, non-binary, or another gender. Sexual orientation also should be collected and tracked for children 10 and older who are able to understand and discuss the concept.
Provide staff training
Caseworkers should be trained to ask children and families about their demographic background rather than assume designations (often incorrectly) based on appearance, language spoken, or other characteristics. “We came up with new policies basically saying that the individual gets to decide how they identify,” said Rachel Thorburn, assistant program administrator at the Hawaii Department of Human Services. “It’s not something for someone else to define.”
In focus groups with caseworkers, the Center for the Study of Social Policy found that some caseworkers were comfortable asking about people’s racial and ethnic backgrounds, while others said they could benefit from some training. In Hawaii, caseworkers appreciated receiving scripts they could use when talking to families about race. New York has created guidance for respectfully asking questions related to gender identity and sexual orientation, and the Human Rights Campaign Foundation has published a comprehensive guidance applicable to all jurisdictions.
In Washington, the Department of Children Youth & Families (DCYF) has taken steps to align guidelines around race and ethnicity data collection across agency program areas (i.e. early learning, child welfare, and juvenile rehabilitation). Historically, guidance had been inconsistent both between and within these three program areas. For example, not all programs were providing specific guidance on what to do if a client’s racial or ethnic background was unknown. To avoid field staff being put in a position of having to guess when filling out required race and ethnicity fields for clients, DCYF issued guidance to include an “unknown” category in all data collection systems.
Refine data collection and reporting systems
Data systems need to be updated to more accurately capture race and ethnicity. Even if agencies are required to report using the OMB categories, they can collect nuanced data to use internally and aggregate categories as needed for external reporting. Data systems also should be created such that racial and ethnic identity can be updated over time. “Race/ethnic identity is fluid,” said Lisa Mishraky-Javier, senior associate at the Center for the Study of Social Policy. “It changes over time. Systems should put mechanisms in place for that information to be updated.” Data broken out by race/ethnicity and age also should be publicly available and accessible, as Minnesota has done with its dashboard of children in foster care, as should data broken out by gender identity and sexual orientation.
The categories that we have aren’t that meaningful, even if they are accurate. They’re too broad. We want to be specific in order to figure out what people’s needs are.
– Ali Jawetz, Senior Policy Analyst, Center for the Study of Social Policy
Strategies for creating and using disaggregated data
Addressing disproportionality and disparities also requires transforming the way knowledge is generated and applied. This includes: involving community members and people with lived experience in designing evaluations, interpreting results, and developing policies to address disparities; conducting decision-point analyses and examining root causes; using data to promote continuous quality improvement; and examining which interventions are successful for different groups of people. Collecting information on clients’ race and ethnicity, gender, sexual orientation, and other demographic variables is important not only to monitor for disproportionality and disparities, but to develop and implement services that are tailored and relevant.
Including community voices in the collection of data, research design, and interpretation of outcomes through participatory action research can improve the quality and representativeness of information and help maintain the continuity and sustainability of work. In addition, it can help address the lack of demographic diversity among people who typically collect, analyze, interpret, and publicize data on children, families, and communities. When data analysts report on cultures or community norms unfamiliar to them, they may overlook key contexts that influence outcomes. This can lead to incorrect conclusions that blame families and further exacerbate disparities.
Involving community stakeholders requires time and money, and community partners should be credited and compensated for their work. Findings gleaned through data analysis need to be presented in ways that are accessible for community members, yet not overly simplified.
Resources from the federal Office of Planning, Research & Evaluation can help inform jurisdictions on how to engage community members.
Ensure stakeholders have shared language
To move this work forward, common understanding of terms is important. “Part of this work is making sure we have shared language and that we’re all on the same page with what this work is about,” said Stephanie Jones Peguero, administrator for the Performance Accountability and Quality Improvement Administration of the Washington, D.C., Child and Family Services Agency. “Are we talking about the same things when we talk about disparity and disproportionality?” To promote common understanding across stakeholders, Washington, D.C.’s race equity steering committee has a subcommittee dedicated to shared language.
Consider partnering with local research organizations and universities
Jurisdictions vary in their internal capacity for data analysis and reporting. Local research organizations and universities can contribute analytic expertise and may be positioned to solicit feedback from community members and people with lived experience who may be reluctant to share their experiences openly with staff from child welfare agencies. In Utah, an external research organization has been brought on board to ask clients specifically about their experiences involving caseworkers inquiring about their racial/ethnic background. Allegheny County, Pa., worked with an external research firm to conduct an institutional analysis of the experiences of youth who identify as LGBTQIA+ who were involved with the child welfare system.
Identify and address decision points where disparities occur
Child welfare agencies should conduct root cause analyses to assess for disproportionality at each decision point, and identify and address policies and practices that contribute to those trends. Washington, D.C., identified that substantiation rates are higher for Black children than white children, and its race equity steering committee is exploring causes and solutions. A study in New York City found that LGBTQIA+ youth are overrepresented in foster care and more frequently youth of color. The study also found they were more likely to experience group or residential placements, challenges in relationships with family members, and higher risk factors for depression. The New York City Administration for Children’s Services subsequently created an action plan outlining its approach and commitment to addressing these disparities, providing targeted and quality services, and improving outcomes for LGBTQIA+ youth in foster care.
We need to be able to present data in a way that promotes conversations around equity. If we present it in a way that is overly complex, it becomes unpalatable to the folks who are ultimately being tasked with implementing changes to eliminate these differences between populations. If we present it in a way that is overly simplified, it becomes misleading and ultimately can be misinterpreted.
– Kevin Cummings, Senior Researcher for Early Learning, Office of Innovation, Alignment and Accountability, Washington State Department of Children, Youth, & Families
Examine data at the local level
Just as broad demographic categories mask important differences in subgroups, examining outcomes at broad geographical levels can mask important differences by location. Well-being indicators in Washington, D.C., show that children and families are doing well overall, but a closer examination shows that children in three neighborhoods are struggling more than their peers from other parts of the district. “When you’re looking at indicators on a broad scale for D.C., the success masks what’s really happening for our families that are most impacted by child welfare,” Jones Peguero said. When examining data from small localities, it may be necessary to oversample some populations (such as Asian American subgroups, Native Hawaiians, LGBTQIA+) to ensure sufficient representation.
Use data to improve services
Robust demographic data also can be used to tailor services to specific groups of children and families. For example, some Native Hawaiian people may benefit from social services in Hawaii that are rooted in their values and culture, such as pounding taro root and making poi together., “You don’t give everyone the same thing,” said Rachel Thorburn, assistant program administrator for the Hawaii Department of Human Services. “You give them what they need, tailoring services based on what you find out about their ethnic background, culture, and practices. The more you know, the better you understand their background, the more you’ll be able to tailor your services.”
Jawetz recounts that when working on the Center for the Study of Social Policy’s Data for Equity project, “we were talking to people in focus groups about the importance of collecting this data, and a lot of people were talking about the importance of caseworkers getting to know them — not just asking these questions to check a box, but actually getting to know a family’s cultural background for the purpose of better serving them and their case.”
By the Numbers: Using Disaggregated Data to Inform Policies, Practices, and Decision-Making (Annie E. Casey Foundation, 2016)
- As part of a Race for Results case study series, this brief provides examples of using disaggregated county and neighborhood data to better understand the context of poverty, healthcare, and employment.
Focusing on Race Equity Throughout Change and Implementation (Capacity Building Center for States, 2021)
- As part of a Change and Implementation in Practice series, this brief includes questions for teams to discuss as they address race equity, from exploration of root causes to selecting, evaluating, and sustaining solutions.
Centering Racial Equity Throughout Data Integration (Actionable Intelligence for Social Policy, University of Pennsylvania, 2020)
- This toolkit includes best practices and strategies to promote racial equity in planning, data collection, data access, use of algorithms and statistical tools, data analysis, and reporting/dissemination.
A Guide to Anti-Racist Data Collection for Case Workers and Other Frontline Staff and A Guide to Anti-Racist Data Collection for System Leaders and Data Administrators (Center for the Study of Social Policy, 2021)
- These tools provide information about best practices for collecting data on race and ethnicity from youth and families, including guidelines for data collection, analysis, and reporting. The Center for the Study of Social Policy’s Data for Equity and Action project also has available a literature review, poster, and know-your-rights guide for youth.
Racial Bias in Data Assessment Tool (Chapin Hall at the University of Chicago, 2021)
- This tool can be used to assess whether survey datasets used for secondary data analysis may have racial/ethnic biases. It includes recommendations for addressing biases, and the questions included in the tool can also inform child welfare agencies and their partners in developing unbiased datasets.
- This paper provides guidance for ensuring that racial/ethnic equity and community input are incorporated throughout the research process, from planning for and designing research, to dissemination.
Doing Evaluation in Service of Racial Equity (W.K. Kellogg Foundation, 2021)
- This set of three practice guides demonstrates how to integrate racial equity into all evaluation practices, including debunking myths, deepening community engagement, and diagnosing biases and systems.
ACS Office of LGBTQ Policy & Practice (New York City Administration for Children’s Services)
- New York City’s ACS has created several guides for practitioners, including how to respectfully ask about sexual orientation and gender identity and best practices for working with transgender and gender non-conforming youth.
- This publication, created in conjunction with the Putting Pride into Practice Project, provides guidelines for collecting information on sexual orientation, gender identity, and gender expression for individual case planning and agency planning and assessment, as well as for recording information, sharing information, and embedding such practices.
Sexual Orientation, Gender Identity, and Gender Expression Data Collection (Human Rights Campaign Foundation, 2017)
- This guide provides information on collecting data on sexual orientation, gender identity, and gender expression, including scripts for staff to use when asking questions and recommended questions.
1 Content of this brief was informed through consultation with members of the Knowledge Management Lived Experience Advisory Team on April 18 and May 6, 2022. This team includes youth, parents, kinship caregivers, and foster parents with lived experience of the child welfare system who serve as strategic partners with Family Voices United, a collaboration between FosterClub, Generations United, the Children’s Trust Fund Alliance, and Casey Family Programs. Members who contributed to this brief include: Aleks Talsky, Roberto Partida, and Keith Lowhorne.
2 This brief also is based on interviews with Lesley Lundeberg, Service Continuum Administrator, Utah Division of Child and Family Services; Dr. Brandynicole Brooks, Administrator, Child Welfare Training Academy, and Stephanie Jones Peguero, Administrator, Performance Accountability and Quality Improvement Administration, Child and Family Services Agency, Washington, DC, August 24, 2021; Lisa Mishraky-Javier, Senior Associate, and Ali Jawetz, Senior Policy Analyst, Center for the Study of Social Policy, September 23, 2021; Rachel Thorburn, Assistant Program Administrator, Hawaii Department of Human Services, Social Services Division, Child Welfare Services Branch, October 4, 2021; and Kevin Cummings, Senior Researcher for Early Learning, Office of Innovation, Alignment and Accountability, Washington State Department of Children, Youth, & Families, November 3, 2021.
3 When examining decision points throughout a case, calculating a relative risk index (rather than a disparity index) makes it easier to determine when in the life of each case disparities occur. This is because disparity indices always use the general population to calculate rate per 1,000 children. While disparity indices may be appropriate for front-end decision points (such as cases accepted for investigation), they are not appropriate for later points in the decision-making continuums (such as number of exits from foster care) because only children who have entered care are able to exit care. The relative risk index adjusts the denominator at each decision point to reflect the smaller population.