Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Feb 1.
Published in final edited form as: Child Maltreat. 2022 Feb 2;28(1):42–54. doi: 10.1177/10775595211070248

Inequities in CPS contact between Black and White children

Margaret M C Thomas a,*, Jane Waldfogel b, Ovita F Williams b
PMCID: PMC9325927  NIHMSID: NIHMS1796178  PMID: 35081781

Abstract

Child protective service (CPS) contact occurs at substantially higher rates among Black than White families. The present study considers systemic racism as a central driver of this disparity and emphasizes racialized poverty as a possible mechanism. We used data from the Fragile Families and Child Wellbeing Study and logistic regression analyses to assess the associations between income poverty, a racialized experience, and CPS contact, separately among Black and White families. Results indicated that income poverty was a significant predictor of CPS contact among White families, who were protected by higher income. In contrast, income per se was not a significant predictor of CPS contact among Black families, who were instead impacted by racialized family regulation and consequences of poverty, such as poor health and depression. Refundable state Earned Income Tax Credit (EITC) policies were protective for Black families, and more expansive Temporary Assistance for Needy Families (TANF) programs decreased CPS contact for Black and White families. Implications include centering systemic racism and specifically racialized poverty as causes of racial inequities in CPS contact and rethinking the role of CPS in protecting children.

Extensive research has documented consistent inequities in child protective services (CPS) involvement between Black and White children in the US, such that Black children make up a substantially larger proportion of children in CPS cases as compared to their share of the child population (Kim et al., 2017; Lanier et al., 2014). We consider that, at their root, these inequities are the consequence of systemic racism: there is no inherent relationship between race and child maltreatment. Rather, race is a proxy for the societal and institutional privileges and oppressions people experience because of their membership in a racialized group1 (Zuberi, 2001). Increasingly, researchers are using conceptual frames that emphasize the role of systemic racism in racial inequities in health and wellbeing (Jones, 2000). We approached this study in that tradition, defining systemic racism as the ways in which institutions enact practices and policies that distribute resources and power with inequitable benefits and burdens by racialized group. We premise our empirical work on the theoretical position that aggregate racial inequities reflect systematic differences in the ways Black and White families experience contact with CPS systems. From this perspective, we argue that the central, outstanding, empirical questions about racial inequities in CPS contact pertain to identifying proximate and distal sources of racism which drive these inequities, ultimately allowing us to target interventions to eliminate racial inequity in CPS involvement.

One central way in which racism may drive inequities in CPS contact is through racialized poverty, such that racism creates circumstances of economic deprivation among Black families which result in CPS contact (Drake et al., 2011; Font et al., 2012; Jonson-Reid et al., 2009; Lanier et al., 2014). We define racialized poverty as the institutional policies and practices that have historically and continue today to harm Black families’ economic wellbeing due to their racialized group membership. A rich body of empirical evidence has investigated specific policies and practices which have led to lower incomes, less wealth, more debt, and more material hardship for Black Americans in the aggregate. These policies and practices include racial segregation (Andrews et al., 2017); exclusion of occupations predominantly employing Black Americans from coverage under Social Security and minimum wage laws (Derenoncourt & Montialoux, 2020; DeWitt, 2010); mass incarceration of Black Americans and subsequent barriers to employment (Alexander, 2010); federal housing policies, including redlining and racially restrictive covenants (Rothstein, 2017); reliance on local property taxes to fund public education (Reardon, 2016); and the privileging of heteronormative, single-generation, two-parent family structures via tax incentives, public assistance requirements, and CPS protocols (Lee, 2016; Parke, 2004; Tax Policy Center, 2020). These examples highlight some of the racialized causes of disparate exposure to poverty for Black Americans. Ultimately, the racialization of poverty is rooted in slavery and has been reinforced by US law and policy throughout the history of the nation (Parolin, 2019; Rothstein, 2017).

In the present study, we examine one mechanism of systemic racism, racialized poverty, and its possible consequences for inequities in CPS contact between Black and White children. To do this, we bring together rich, individual-level data about young children and their families with city- and state-level policy data to examine racial inequities in CPS contact, the differential associations of income poverty with CPS contact for Black and White families, the role of family demographic and risk factors which may reflect racialized correlates of CPS contact, and the potential for local and state policy to exacerbate or mitigate racial inequities in CPS contact.

Background

Racial Inequities in CPS Contact

The inequitable exposure of Black children and families to CPS involvement is clear and persistent. National estimates suggest that 53% of Black children will experience CPS contact by age 18, as compared to 28% of White children (Kim et al., 2017). In 2018, Black children made up 21% of children with substantiated CPS cases but just 14% of the child population, whereas White children made up 44% of CPS cases but 50% of the child population (U.S. Department of Health & Human Services, 2020). In addition to inequitable rates of contact with CPS, Black families face more punitive regulation across numerous stages of CPS involvement, from initial maltreatment report (Putnam-Hornstein et al., 2013), to the decisions to investigate (Harris & Hackett, 2008) and to substantiate (Dettlaff et al., 2011), to placement in out of home care (Putnam-Hornstein et al., 2013), and ultimately to family reunification (Lu et al., 2004). Each of these represents a stage where institutionalized racism drives policing of Black families (Dettlaff et al., 2020). In this study, we examine CPS contact broadly, with the aim of understanding how one manifestation of systemic racism – racialized poverty – drives inequities. We conceptualize racialized poverty as a form of systemic racism which can be redressed through policy and which, if eliminated, could prevent CPS contact entirely for substantial numbers of Black families, thereby also decreasing families’ exposure to subsequent stages of CPS involvement.

Two prominent theoretical explanations for the inequitable exposure of Black children to CPS involvement attempt to address racism and risks which are nominally race-neutral. These models, the “bias” and “risk” models, both emphasize individual-level factors which, in aggregate, produce racial inequities in CPS involvement (Barth et al., 2001). In brief, bias models suggest that individual racial bias among actors in the CPS system results in disparate involvement of Black children in CPS (Webb et al., 2002). Such models draw on evidence such as findings that some case workers assess comparable risks as more dangerous within Black families than White families (Dettlaff et al., 2011; Font et al., 2012). In contrast, risk models posit that sources of risk which are associated with CPS involvement are more common among Black than White families and explain racial inequities in CPS involvement (Drake et al., 2011; Lanier et al., 2014). Such models suggest that, absent differences in underlying risks, Black and White families would be similarly represented in CPS systems (Drake et al., 2011). Some applications of risk models emphasize structural roots of disparities in risk (Font et al., 2012; Maguire-Jack et al., 2015) while others do not examine systemic causes (Drake et al., 2011; Lanier et al., 2014).

Numerous studies have tested these two models, often against one another. On the whole there is little evidence to support the bias model as an explanation for consistent racial inequities in CPS involvement (Drake et al., 2011; Font et al., 2012; Maguire-Jack et al., 2015), whereas studies do suggest that increased exposure among Black families to key risks for CPS involvement, particularly income poverty, explains a substantial portion of racial inequities (Bartholet et al., 2011; Drake et al., 2011; Font et al., 2012; Lanier et al., 2014; Maguire-Jack et al., 2015). However, the frequent implication that bias and risk models are opposing explanations for racial inequity in CPS involvement masks the critical role of systemic racism in driving precisely the types of increased risks Black children face as articulated in risk models. The success of risk models in identifying proximate causes of racial inequities in CPS involvement reinforces the importance of recognizing that historic and contemporary institutionalized racism creates inequities in poverty and other risk factors (Alexander, 2010; Andrews et al., 2017). Some prior research has acknowledged the systemic racism underlying inequities in poverty and other risk factors (Font et al., 2012; Maguire-Jack et al., 2015), yet the framing of most prior work has emphasized disparate risks and higher incidence of maltreatment in Black families without consistently attributing these findings to systemic racism (Bartholet et al., 2011; Drake et al., 2011). While it remains difficult to measure systemic racism, it is essential to consistently identify this structural force which may drive racial inequities in CPS contact (Dettlaff et al., 2020). With this in mind, we considered systemic causes of higher exposure to a range of risks for Black families versus White families, including poor physical health, depression, and criminal justice involvement. Such risks are correlates of CPS contact and poverty (Berger et al., 2016; Heflin & Iceland, 2009; Slack et al., 2011) and also occur at higher rates among Black families.

Racialized Poverty and CPS Contact

Income poverty stands out as a key risk for CPS involvement, affecting a substantial proportion of Black families, in large part due to past and present systemic racism. Because of the racialization of poverty, 31% of Black children (but 10% of White) lived in families with income below the federal poverty level (FPL) in 2019 (Annie E. Casey Foundation, 2021). Black children also experience deep poverty (income below 50% FPL) at three times the rate of White children (15% versus 5%) (Annie E. Casey Foundation, 2021). Moreover, 75% of Black children spend at least one year in poverty, compared to 30% of White children (Ratcliffe, 2015).

Racialized poverty has significant consequences for racial inequities in CPS involvement as low-income families are disproportionately subject to CPS involvement, for three primary reasons. First, a majority of CPS cases are substantiated on the basis of neglect, which refers to a caregiver’s failure to meet a child’s basic safety and wellbeing needs (U.S. Department of Health & Human Services, 2020). While most states hold that neglect is caused by more than poverty alone, there is little guidance to distinguish poverty from neglect (Dale, 2014), and the basic definition of neglect reflects resource deprivation which is more common among low-income than higher-income families (Karpman et al., 2018). Second, many low-income families utilize one or more public assistance programs. These programs introduce explicit surveillance of families’ behavior, resources, and relationships, scrutiny which may increase the risk that families are reported to CPS (Fong, 2019). Third, poverty can heighten the risk of interpersonal and parenting stress, harsh parenting, and mental health and child behavior challenges, which may result in increased risk of maltreatment (Jonson-Reid et al., 2009). Because poverty is associated with CPS contact, the systematic ways in which poverty is racialized represent systematic ways in which racial inequities in CPS contact have been created and are maintained.

While a large body of research has identified associations between racialized group membership, poverty, and CPS involvement, many studies do not include household measures of income but instead rely on other indicators of poverty, such as cash assistance receipt (Jonson-Reid et al., 2009), or examine aggregate measures of poverty and racial group membership (Drake & Rank, 2009). Further, very little work has considered differences in the associations between poverty and CPS contact by racial group, and such work has focused on neighborhood (Drake & Rank, 2009; Klein & Merritt, 2014; Zhang et al., 2021), county (Maguire-Jack et al., 2015), or state (Lanier et al., 2014) poverty rather than household poverty. This study addresses important gaps in the literature by examining associations between household poverty, racialized group membership, and CPS contact and by considering differences between Black and White families in these associations. Additionally, we considered the role of key anti-poverty policies in exacerbating or mitigating racialized poverty.

Research Questions

Using longitudinal data on a cohort of children born in large US cities, this study addressed four research questions: 1) What is the cumulative prevalence of CPS contact by age 5 and is there inequity between Black and White children? 2) Are there disparities by poverty status in CPS contact and are there differences in such disparities between Black and White children? 3) Do racial inequities and disparities by poverty status in CPS contact persist when we account for differences in family characteristics? 4) What is the role of city and state factors, specifically racialized public policies, in exacerbating or mitigating racial inequities in CPS contact?

Methods

Data

This study drew on the restricted access version of the Fragile Families and Child Wellbeing Study (FFCWS) data (approved by institutional IRB). The FFCWS is a longitudinal birth cohort study of a stratified, multistate probability sample which follows nearly 5,000 children born in 20 large US cities2 between 1998 and 2000 (Reichman et al., 2001). When weighted, the data are representative of all births in cities with populations of 200,000 or more in 1998. The present study draws on the first four waves of data, collected when children were newborn and 1, 3, and 5 years old. At each of those waves, the child’s mother was interviewed, and we rely on those interviews for data on the focal children and their families. The design of the FFCWS sample is particularly useful for this study, as the data intentionally oversampled unmarried parents (by a 3:1 ratio) as a proxy for families facing precarious circumstances (Reichman et al., 2001). As a result of the study design, including the urban focus, the families in the FFCWS data include unusually substantial populations of low-income families, Black families, and unmarried and single-parent families. As such, we have sufficient samples among racialized and income groups to test important subgroup differences. Additionally, capitalizing on variation across the 20 cities and 15 states in which the children were born, we merged data from several sources to the FFCWS data to measure city- and state-level policy factors.3

Measures

Race.

Our measure of child race is the only measure which draws on FFCWS data beyond the 5-year wave. We measured race as a composite based on child self-reported race at the 15-year wave and mother and father race as reported at baseline. We constructed a binary measure of race distinguishing White, non-Latinx children from Black children of any ethnicity. We identified as Black those children who self-identified as Black or had a Black parent of any ethnicity. We identified as White those children who had a White mother, did not have a Black father, and did not self-identify as Black, excluding those children who had a Latinx mother. Because of our central focus on inequities between Black and White children, our analyses excluded children who were not either Black or White and non-Latinx. Additionally, we conducted sensitivity tests using mother race only, which produced comparable results.

CPS contact.

The 5-year wave is the first in which the FFCWS collected information about CPS contact. Mothers were asked whether they had ever been contacted by CPS since the child’s birth. We used this information to construct an indicator for any CPS contact. We measured CPS contact at age 5 in order to focus our analysis on a defined developmental and social stage (early childhood) and to include data from the greatest number of families possible.

Income poverty.

Drawing on mothers’ reports at the 5-year wave, we constructed a categorical measure of household income-to-poverty ratio (<50% federal poverty line (FPL), 50–99% FPL, 100–199% FPL, 200–299% FPL, 300%+ FPL).

Family demographic characteristics.

At the baseline wave, we measured child sex, child low birth weight, and mother US-born status. At the 5-year wave4, we measured child age (in months); mother age (in years); mother education level; mother marital and cohabitation status, capturing both marital status and whether the mother was living with the child’s father, a new partner, or no partner; mother housing status, capturing ownership and government assistance; whether father did not have regular contact with the child; mother total number of children; and whether mother had more children at the 5-year wave than at baseline.

Mother experiences of risk.

At the 5-year wave, we constructed measures of mother experiences which may increase risk of CPS contact. These include a binary measure of mother self-reported health (good, fair, or poor versus very good or excellent); an indicator for mother meeting the threshold for probable major depression on a standardized scale5; an indicator for mother use of any recreational drug; an indicator for mother criminal justice involvement (any incarceration, criminal charges, or convictions in prior two years); and an indicator for mother experience of any intimate partner violence in her romantic relationship at the 5-year wave.

City policy factors.

For each city-level measure, we used data from 2005, when most of the FFCWS families completed their 5-year survey.6 We constructed city- and race-specific rates of CPS contact for children ages 0 to 6 using restricted administrative data from the National Child Abuse and Neglect Data System (NCANDS) Child File (approved by institutional IRB) in combination with population data (Centers for Disease Control and Prevention, 2020).7 We used publicly available data from the Incarceration Trends dataset (Vera Institute of Justice, 2020) to construct city- and race-specific measures of the adult jail population. Finally, we included a publicly available index of city political ideology, an established measure of aggregate liberalism/conservatism at the city level (Tausanovitch & Warshaw, 2016).8 These measures all reflect the policy context of the cities in which the focal children were born.

State policy factors.

We relied on data from 2005 from the University of Kentucky Center for Poverty Research (2020) for four state-level policy measures: the absence of a refundable Earned Income Tax Credit (EITC); minimum wage; proportion of residents with income under 100% FPL who received Temporary Assistance for Needy Families (TANF; the “TANF-to-poverty ratio”), and the percentage of low-income children without health insurance.9

Sample

To construct the study sample, we first excluded children who were not Black or White as defined in the measures section (n=1,529). Next, we excluded children whose mothers were not interviewed at the 5-year wave (n=509). Finally, for our main analyses, we excluded children with missing covariates (n=236), examining a complete case sample of 2,624 children, 2,007 of whom were Black and 617 of whom were White. Table 1 summarizes the family demographic, risk experience, and city and state policy characteristics of the sample. These descriptive statistics are weighted to be representative of the focal children’s birth cities. We conducted adjusted Wald tests to determine whether the difference in means between the Black and White samples were statistically significant for each factor (p-values for Wald tests are presented below).

Table 1.

Sample Descriptive Statistics (Weighted to Represent 20 Cities)

Black Children (n=2,007) White Children (n=617) Difference in Means (Wald test), Black vs. White
Mean SD Mean SD p
Child Protective Service (CPS) Contact Outcome
CPS contact rate by age 5 0.12 0.03 0.03 0.01 .016
Mother’s Household Income Poverty
Poverty ratio, continuous 1.44 0.08 3.74 0.25 <.001
Poverty ratio
 <50% federal poverty level (FPL) 0.30 0.03 0.03 0.01 <.001
 50–99% FPL 0.22 0.02 0.10 0.07 .078
 100–199% FPL 0.21 0.03 0.14 0.03 .153
 200–299% FPL 0.15 0.05 0.12 0.04 .368
 300%+ FPL 0.13 0.02 0.62 0.05 <.001
Family Demographic Characteristics
Child sex (female) 0.47 0.05 0.43 0.06 .577
Child age at year 5 wave (Y5) (months) 61.50 0.10 60.50 0.26 .011
Child born at low birth weight (yes) 0.12 0.02 0.04 0.01 .017
Mother age at Y5 (years) 30.06 0.41 35.21 0.33 <.001
Mother not US-born 0.13 0.05 0.06 0.03 .214
Mother education level at Y5
 Less than high school 0.22 0.03 0.07 0.03 .014
 High school or equivalent 0.34 0.07 0.20 0.08 .296
 Some college or technical school 0.35 0.03 0.19 0.05 .046
 College or advanced degree 0.09 0.03 0.54 0.05 <.001
Mother housing status at Y5
 Own 0.16 0.04 0.60 0.04 <.001
 Rent, no assistance 0.41 0.04 0.28 0.05 .056
 Rent, assistance 0.17 0.02 0.03 0.03 .009
 Public housing 0.21 0.04 0.07 0.07 .085
 Other 0.05 0.01 0.03 0.01 .019
Mother marital and living status at Y5
 Married, lives with bio father 0.25 0.03 0.85 0.02 <.001
 Married, lives with new partner 0.04 0.01 0.02 0.01 .357
 Not married, lives with bio father 0.14 0.02 0.02 0.01 <.001
 Not married, lives with new partner 0.14 0.02 0.03 0.01 <.001
 Not married, not living with partner 0.44 0.02 0.08 0.03 <.001
Father does not have regular contact at Y5 0.31 0.03 0.05 0.01 <.001
Mother number of children at Y5 2.80 0.11 2.72 0.21 .781
Mother had new child since baseline 0.41 0.03 0.60 0.05 .004
Mother Experiences of Risk
Mother poor health status at Y5 (good/fair/poor vs. very good/excellent) 0.42 0.04 0.21 0.03 .002
Mother depression at Y5 0.15 0.01 0.12 0.03 .437
Mother drug use at Y5 0.05 0.01 0.06 0.02 .490
Mother criminal justice involved at Y5 0.06 0.02 0.01 0.00 .081
Mother intimate partner violence at Y5 0.16 0.04 0.17 0.03 .875
City Policy Factors
City political ideology (mean; higher=more conservative) −0.46 0.01 −0.41 0.01 .003
Race-specific city jail population (per 100,000) 996.67 17.87 218.16 3.26 <.001
Race-specific city CPS rate (percentage; 0-to-6-year-olds) 6.21 0.11 2.49 0.06 <.001
State Policy Factors
State does not have refundable Earned Income Tax Credit (EITC) 0.36 0.02 0.43 0.02 .035
State minimum wage (dollars) 5.57 0.03 5.65 0.03 .112
State Temporary Assistance for Needy Families (TANF)-to-poverty ratio (percentage) 13.26 0.19 12.02 0.22 .002
State rate uninsured low-income children (percentage) 5.11 0.13 6.10 0.16 .002
Open in a new tab

The Black and White samples demonstrated notably different aggregate demographic and risk characteristics, likely reflecting ways in which Black families face marginalization. CPS contact by age 5 was substantially more common among Black than White children (12.4% vs. 2.6%, p=.016), and Black families were significantly more likely to experience deep poverty (income below 50% FPL; 30% vs. 3%, p<.001) while White families were much more likely to have high income (above 300% FPL; 62% vs. 13%, p<.001). Similar patterns of racial marginalization existed in education level and housing status: the majority of White mothers had a college degree or more (54% vs. 9%, p<.001) and owned their home (60% vs. 16%, p<.001), while the majority of Black mothers had a high school degree or less and rented their home. Black mothers were significantly more likely to be unmarried and not to live with the focal child’s father or any partner (44% vs. 3%, p<.001), while the majority of White mothers were married to and living with the child’s father (85% vs. 25%, p<.001). Black mothers also faced significantly higher rates of poor health (42% vs. 21%, p=.002). Overall, Black families lived in more ideologically liberal cities than did White families (p=.003). City-level rates of the race-specific adult jail populations were highly inequitable, with more than four-fold higher rates of incarceration for Black than white adults (997 per 100,000 Black adults vs. 218 per 100,000 White adults, p<.001). Similarly, race-specific rates of city-level CPS contact for young children demonstrated significantly higher CPS contact among Black than White children (6.2% vs. 2.5%, p<.001). At the state level, Black families lived in states with more generous social safety net policies, including of a refundable state EITC (not available to 36% of Black families vs. 43% of White families, p=.035), higher state TANF-to-poverty ratio (13% vs. 12%, p=.002), and lower rate of uninsured low-income children (5% vs. 5%, p=.002).

Analysis

As reported in Table 1, we first estimated the prevalence of CPS contact among the Black and White child samples. Then, we employed logistic regression to test the association between child race and CPS contact, the association between household income-to-poverty ratio and CPS contact, and the joint associations of child race and household income-to-poverty ratio with CPS contact. Given the significant differences in the distribution of poverty by race, we also conducted logistic regressions assessing the association of household income-to-poverty ratio and CPS contact separately in the Black and White child samples. We then used multiple logistic regression to examine whether racial inequities and disparities by income-to-poverty ratio in CPS contact persisted when we accounted for family demographic and mother risk characteristics, again separately in the Black and White samples. Finally, we extended the multiple logistic regression models to examine the role of city and state policy variables in exacerbating or mitigating racial inequities. We also conducted a series of sensitivity tests, replicating our main analyses with a) family demographic and risk factors measured in earlier waves; b) alternative city and state policy variables, described in the measures section; c) city fixed effects models; and d) multiply imputed data, to address the missing covariates. Our findings were robust to these tests (Supplementary Materials, Appendix A). Further, to understand the relative importance of differences between the Black and White samples in group-level characteristics (“endowments”) and differences in consequences of those characteristics (“returns”), we conducted an Oaxaca decomposition analysis (Supplementary Materials, Appendix B).

Results

Table 2 presents results of uncontrolled logistic regression models assessing across-race associations between household poverty and CPS contact by age 5. In Panel A, results indicate that Black children were at greater risk than White children for CPS contact by age 5 (odds ratio (OR)=1.6, 95% confidence interval (CI)=[1.13, 2.26], p-value (p)=.001). Panel B demonstrates that each income poverty category from deep poverty (income <50% FPL) to moderate income (200–299% FPL) was associated with greater odds of CPS contact than the highest income category (300%+ FPL). Results presented in Panel C indicate that the association of race with CPS contact disappeared (0.92 [0.63, 1.34], p=.662) when the effects of race and poverty were modeled jointly, while the association of poverty with CPS remained significant and of consistent magnitude. Finally, Panel D presents the interaction of child racial group with household income poverty. These results suggest that there was a statistically significant interaction between racial group membership and income poverty for deep poverty (<50% FPL), other income poverty (50–99% FPL), and near-poverty income (100–199% FPL). Accounting for this interaction, the main effects on CPS contact of child race (3.35 [1.04, 10.81], p=.043) and income poverty ratio were statistically significant (p<.001 for all income poverty categories below 200% FPL).

Table 2.

Across-Racial Group Associations of Income Poverty with Child Protective Service (CPS) Contact by Age 5

Black and White Children (N=2,624)
OR 95% CI p
Panel A: Association of Child Race Alone
OR
 95% CI
 p
Child is Black (vs. White) 1.60 1.13, 2.26 .007
Panel B: Association of Household Income Poverty Alone
OR
 95% CI
 p
Poverty ratio (vs. 300%+ federal poverty level (FPL))
 <50% FPL 5.83 3.26, 10.43 <.001
 50–99% FPL 6.49 3.61, 11.68 <.001
 100–199% FPL 3.74 2.07, 6.78 <.001
 200–299% FPL 3.16 1.64, 6.08 .001
Panel C: Association of Child Race and Household Income Poverty Jointly
OR
 95% CI
 p
Child is Black (vs. White) 0.92 0.63, 1.34 .662
Poverty ratio (vs. 300%+ FPL)
 <50% FPL 6.08 3.29, 11.21 <.001
 50–99% FPL 6.75 3.66, 12.44 <.001
 100–199% FPL 3.87 2.10, 7.13 <.001
 200–299% FPL 3.23 1.66, 6.28 .001
Panel D: Association of Interacted Child Race and Household Income Poverty
OR
 95% CI
 p
Child is Black (vs. White) 3.35 1.04, 10.81 .043
Poverty ratio (vs. 300%+ FPL)
 <50% FPL 34.62 10.42, 114.98 <.001
 50–99% FPL 17.45 5.12, 56.70 <.001
 100–199% FPL 9.38 2.99, 29.39 <.001
 200–299% FPL 2.23 0.49, 10.12 .300
Interaction: Child is Black*Poverty ratio
(vs. Child is White*300%+ (FPL)
 Child is Black*<50% FPL 0.09 0.02, 0.36 .001
 Child is Black*50–99% FPL 0.22 0.55, 0.88 .032
 Child is Black*100–199% FPL 0.23 0.06, 0.86 .030
 Child is Black*200–299% FPL 1.09 0.20, 5.91 .922
Open in a new tab

Given the substantial differences in the distribution of poverty within Black and White families (see Table 1), Table 3 presents within-race models. While poverty was significantly associated with CPS contact for both Black and White children, the trend and magnitude of those associations differed notably. For Black children, compared to highest-income families (300%+ FPL), those in each lower income category were between 2.12 (CI=[1.05,4.28], p=.035) and 3.74 times (CI=[1.88, 7.45], p<.001) more likely to have CPS contact. In contrast, for White children, compared to highest-income families, those in deep poverty (<50% FPL) were 34.6 times (CI=[10.42, 114.98], p<.001) more likely to have CPS contact, other families in poverty (50–99% FPL) were 17 times (CI=[5.12, 56.70], p<.001) more likely to have CPS contact, and those with near-poverty income (100–199% FPL) were 9.4 times (CI=[2.99, 29.39], p<.001) more likely to have CPS contact, while families with moderate income (200–299% FPL) were no more likely to have CPS contact than families with highest income (OR=2.23, CI=[0.49,10.12], p=.300). While differences in sample size and the small White sample may certainly explain the high odds and large confidence intervals in predicting CPS contact for White families, the overall findings suggest comparatively higher odds of CPS contact for White families than for Black families, even at the low end of the estimate range.

Table 3.

Within-Racial Group Association of Income Poverty with Child Protective Service (CPS) Contact by Age 5

Black Children (n=2,007)
 White Children (n=617)
OR 95% CI p OR 95% CI p
Poverty ratio (vs. 300%+ federal poverty level (FPL))
 <50% FPL 3.11 1.57, 6.17 .001 34.62 10.42, 114.98 <.001
 50–99% FPL 3.74 1.88, 7.45 <.001 17.05 5.12, 56.70 <.001
 100–199% FPL 2.12 1.05, 4.28 .035 9.38 2.99, 29.39 <.001
 200–299% FPL 2.42 1.14, 5.16 .022 2.23 0.49, 10.12 .300
Open in a new tab

The stark contrast in these within-race results motivated our decision to stratify our main analyses by race rather than pool the Black and White samples. That analysis used multiple logistic regression models incorporating family demographic and mother risk characteristics, presented in Table 4. For Black children, these results reflect the importance of specific social positioning and risk factors over income poverty in predicting CPS contact. Families with moderate income (200–299% FPL) faced greater odds of CPS contact than highest-income families (300%+ FPL) (OR=2.10, CI=[1.01, 4.36], p=0.46), but there were no significant associations between other income poverty categories and CPS contact. Instead, significant correlates of CPS contact included single mother status (vs. mother married and living with child’s father; OR=2.12, CI=[1.22, 3.70], p=.008); greater number of children (OR=1.29, CI=[1.16, 1.43], p<.001); mother poor health status (OR=1.42, CI=[1.00, 2.01], p=.049); mother depression (OR=1.73, CI=[1.04, 2.89], p=.036); and mother criminal justice system involvement (OR=1.99, CI=[1.29, 3.08], p=.002). Other factors were associated with CPS contact for White children. First, poverty remained an important correlate of CPS contact: families in deep poverty (<50% FPL) and poverty (50–99% FPL) had higher odds of CPS contact than highest-income (300%+ FPL) families (respectively, OR=8.89, CI=[1.74, 45.24], p=.009, and OR=3.88, CI=[1.13, 13.27], p=.031), while there was no difference for near-poverty (100–199% FPL) and moderate income (200–299% FPL) families. Other significant correlates for White children were mother non-US-born status (OR=5.84, CI=[1.27, 26.79], p=.023); mother living in public housing (vs. owning home; OR=3.44, CI=1.12, 10.54], p=.031); greater number of children (OR=1.72, CI=[1.17, 2.51], p=.005); mother poor health status (OR=2.84, CI=[1.29, 6.26], p=.010); and mother criminal justice system involvement (OR=3.50, CI=[1.38, 8.87], p=.008).

Table 4.

Associations of Income Poverty with Child Protective Service (CPS) Contact by Age 5, Adjusted

Black Children (n=2,007)
 White Children (n=617)
OR 95% CI p OR 95% CI p
Mother’s Household Income Poverty
Poverty ratio (vs. 300%+ federal poverty level (FPL))
 <50% FPL 1.32 0.50, 3.50 .574 8.88 1.74, 45.24 .009
 50–99% FPL 1.84 0.91, 3.72 .087 3.88 1.13, 13.27 .031
 100–199% FPL 1.39 0.68, 2.85 .362 2.82 0.7, 11.28 .144
 200–299% FPL 2.10 1.01, 4.36 .046 1.04 0.17, 6.39 .965
Family Demographic Characteristics
Child sex (female) 0.97 0.82, 1.16 .746 0.68 0.39, 1.20 .187
Child age at year 5 wave (Y5; months) 0.93 0.87, 1.01 .070 0.84 0.67, 1.06 .138
Child born at low birth weight (yes) 1.25 0.93, 1.67 .142 1.08 0.27, 4.26 .914
Mother age at Y5 (years) 0.98 0.96, 1.01 .225 0.97 0.90, 1.05 .504
Mother not US-born 0.65 0.15, 2.78 .560 5.84 1.27, 26.79 .023
Mother education level at Y5 (vs. college+)
 Less than high school 0.95 0.46, 1.97 .893 3.62 0.64, 20.40 .145
 High school or equivalent 0.72 0.35, 1.47 .362 1.59 0.28, 8.99 .601
 Some college or technical school 1.01 0.45, 2.24 .985 2.63 0.60, 11.57 .201
Mother housing status at Y5 (vs. own)
 Rent, no assistance 0.76 0.46, 1.25 .277 0.89 0.41, 1.93 .760
 Rent, assistance 1.22 0.61, 2.42 .570 0.47 0.09, 2.37 .363
 Public housing 1.05 0.54, 2.06 .884 3.44 1.12, 10.54 .031
 Other 0.80 0.33, 1.94 .619 1.41 0.58, 3.38 .446
Mother marital and living status at Y5 (vs. married, lives with father)
 Married, lives with new partner 1.67 0.85, 3.30 .139 2.77 0.71, 10.73 .141
 Not married, lives with bio father 1.15 0.56, 2.37 .706 1.57 0.45, 5.44 .476
 Not married, lives with new partner 1.51 0.69, 3.34 .306 1.12 0.19, 6.67 .900
 Not married, not living with partner 2.12 1.22, 3.70 .008 1.77 0.42, 7.35 .435
Father does not have regular contact at Y5 1.24 0.84, 1.85 .280 1.23 0.54, 2.76 .624
Mother number of children at Y5 1.29 1.16, 1.43 <.001 1.72 1.17, 2.51 .005
Mother had new child since baseline 1.11 0.83, 1.48 .496 0.76 0.3, 1.96 .577
Mother Experiences of Risk
Mother poor health status at Y5 (good/fair/poor vs. very good/excellent) 1.42 1.00, 2.01 .049 2.84 1.29, 6.26 .010
Mother depression at Y5 1.73 1.04, 2.89 .036 1.01 0.46, 2.19 .984
Mother drug use at Y5 1.13 0.64, 2.03 .669 0.92 0.18, 4.67 .915
Mother criminal justice involved at Y5 1.99 1.29, 3.08 .002 3.50 1.38, 8.87 .008
Mother intimate partner violence at Y5 1.18 0.71, 1.97 .530 0.74 0.23, 2.40 .614
Open in a new tab

Our final set of results built on the multivariate models described above and incorporated city and state policy variables which reflect both broad political ideology with potential consequences for the racialization of poverty and specific policy landscapes which directly reflect the racialization of criminal justice and CPS systems involvement. Table 5 presents the results for these policy variables (family controls were also included but not shown). Our measures of city-level adult jail population and CPS rates did not have meaningful consequences for CPS contact. Neither factor was statistically significant for White children, and the odds of CPS contact for Black children exposed to higher jail (OR=1.00, CI=[1.00, 1.00], p<.001) or CPS rates (OR=0.94, CI=[0.90, 0.99], p=.024) were functionally even. Black children born in states without refundable state EITC’s had higher odds of CPS contact (OR=2.02, CI=[1.34, 3.07], p=.001) while those born in states with a higher TANF-to-poverty ratio had significant but modestly lower odds of CPS contact (OR=0.92, CI=[0.88, 0.97], p=.001). For White children, only state-level TANF-to-poverty ratio emerged as a significant policy variable: higher TANF-to-poverty ratio predicted decreased odds of CPS contact (OR=0.86, CI=[0.75, 1.00], p=.050).

Table 5.

The Role of City and State Policy Variables in Predicting Child Protective Service Contact

Black children (n=2,007)
 White children (n=617)
OR 95% CI p OR 95% CI p
City Policy Factors
City political ideology (mean; higher=more conservative) 0.80 0.37, 1.75 .579 2.73 0.38, 19.39 .317
Race-specific city jail population (per 100,000) 1.00 1.00, 1.00 .001 1.00 1.00, 1.01 .571
Race-specific city CPS rate (percentage; 0-to-6-year-olds) 0.94 0.90, 0.99 .024 .91 0.70, 1.19 .506
State Policy Factors
State does not have refundable Earned Income Tax Credit (EITC) 2.02 1.34, 3.07 .001 2.87 0.37, 22.45 .316
State minimum wage (dollars) 1.16 0.84, 1.59 .374 1.28 0.56, 2.94 .563
State Temporary Assistance for Needy Families (TANF)-to-poverty ratio (percentage) 0.92 0.88, 0.97 .001 .86 0.75, 1.00 .050
State rate uninsured low-income children (percentage) 0.95 0.87, 1.03 .216 1.01 0.79, 1.30 .926
Open in a new tab

Note. Models include full set of family demographic and risk characteristic controls.

Discussion

The results of this study suggest that racialized poverty, one mechanism of systemic racism, may explain a substantial portion of the inequitable exposure of Black children to CPS contact, a finding that is broadly consistent with prior research that suggests differences in systemic risks (e.g., poverty) by race, rather than individual racial animus, likely explain racial inequities in CPS contact (Drake et al., 2011; Font et al., 2012). Critically, findings also illustrate very different associations between poverty and CPS contact for Black and White families, an important extension of our knowledge. While differences in poverty accounted for much of the difference in CPS contact between Black and White children, poverty did not differentiate between Black families with and without CPS contact, whereas poverty played a substantial role in predicting whether White families experienced CPS contact. These findings, and the different consequences of other correlates of CPS contact within racial groups, highlight two themes related to the racialization of poverty and racial inequities in CPS contact.

First, the cross-race strength of poverty as a predictor of CPS contact was driven by the strong associations between deep poverty (<50% FPL) and other poverty (50–99% FPL) and CPS contact among White families coupled with the simple prevalence of deep poverty and other poverty among Black families, 51.5% of whom had income below 100% FPL. These results illustrate how systemic racism may operate through the racialization of poverty: in our sample, poverty was normative for urban Black families, whereas poverty represented a significant and unusual situation of potential risk for urban White families. Relatedly, moderate and high income were normative among White families (73.5% with income 200%+ FPL) and provided near total insulation from CPS contact, whereas higher income did little to protect Black families from CPS contact. These findings demonstrate that policy which criminalizes poverty, including most definitions of neglect within CPS systems (Mallon, 2020), directly criminalizes the normative economic experience of the Black families in our sample. The insignificance of poverty for Black families and the comparatively staggering significance of poverty for White families in predicting CPS contact may reflect statistical discrimination (Agan & Starr, 2016), whereby the CPS system equates race with poverty and neglect. Children in Black families may be constructed as normatively exposed to neglect and thus needing CPS contact, whereas children in White families may be constructed as protected from neglect and thus not needing CPS contact.

Second, those family demographic and mother risk factors which differentially predicted CPS contact for Black and White families highlight additional ways in which poverty has been racialized and Black mothers have been policed and marginalized. The factors predictive of CPS contact among Black children – parenting alone, poor health and mental health, and exposure to the criminal justice system – are risks that Black families generally experience higher incidences of than do White families, likely reflecting consequences of other mechanisms of systemic racism. For instance, the consequences of institutionalized racism in the form of mass incarceration of Black men (Alexander, 2010), particularly young men in urban areas as the fathers in this study were, has forced the separation of many families, whatever their family structure choices might otherwise have been (Western & Wildeman, 2009). Such forced separation of Black families through US policy has direct roots in slavery, which intentionally separated parents, and particularly mothers, from their children and Black adults from their partners and families (Pargas, 2009). Such policy effects are an example of how systemic racism manifests and has downstream consequences for racial inequities in outcomes like CPS contact.

Institutionalized separation of Black families is likely one source of the high rates of single mother families that were associated with CPS contact in this study. Critically, it is the policing of Black single mothers which exposes families to CPS contact, not the existence of Black single mothers (Roberts, 2002) or differences in family structure between Black and White families which have been documented for the past century (Morgan et al., 1993). This example serves to highlight that family regulation policies, from CPS protocols to TANF regulations, have privileged longstanding White family norms and simultaneously marginalized Black family norms (Lee, 2016). Finally, the experiences of risk which were associated with CPS contact among Black families in our sample reflect Black women’s inequitable access to health and mental health care coupled with the direct health and mental health consequences of racism, such as stress and chronic illness (Krieger, 2014; Omeish & Kiernan, 2020). These manifestations of systemic racism underlie the poor physical health and experiences of depression which were associated with CPS contact for Black children, perhaps through parenting behavior or differences in CPS actors’ perceptions of risk across racialized groups (Dettlaff et al., 2011).

The factors that predicted CPS contact for White children – income poverty, non-US-born mothers, and public housing residency – are suggestive of the theory of differential assortment (Drake & Rank, 2009), which posits that, in the aggregate, White families who live in high-poverty areas have a multitude of risk factors, as they otherwise benefit from institutionalized White privilege and intergenerational wealth that broadly permit mobility and risk avoidance. Thus, White families living in public housing or experiencing deep poverty are likely to experience many other challenges, whereas Black families, due to institutionalized racism and consequent reduced wealth and mobility, are less clearly distinguished, in terms of risk profiles, by public housing residency or poverty status. This conceptualization aligns with findings on the differential risks of CPS involvement that Black and White children face based on neighborhood racial composition (Klein & Merritt, 2014) and economic wellbeing (Zhang et al., 2021).

Our policy findings suggest two mechanisms by which state policy may exacerbate or mitigate risk of CPS contact. First, for Black families, the absence of a state EITC was associated with increased risk of CPS contact, suggesting the importance of that income supplement, as other recent work has suggested (Kovski et al., 2021). State EITC’s are a particularly effective state-level anti-poverty policy (Pac et al., 2020), and policy benefits have been found to be concentrated for families with incomes just below and above 100% FPL (Hoynes & Patel, 2018). The prevalence of such low income among Black families in our sample supports the significance of the absence of a state EITC in increasing racialized poverty and risk of CPS for Black families. Second, for both Black and White families, more expansive TANF programs were associated with decreased CPS contact. The smaller benefit of TANF generosity for Black families may reflect their disparate residence in states which allocate TANF funding to non-cash-benefit priorities (Parolin, 2019). Finally, results for city adult jail and CPS contact rates suggested no meaningful consequences, perhaps reflecting confounding with family controls.

Limitations

This study provides new information about racial inequities in CPS contact but does not provide causal evidence. Additionally, while representative of many large cities in the US, these findings are not generalizable to the whole country and specifically do not apply to non-urban settings. Given evidence that CPS involvement and racial inequities in CPS involvement can vary by geography (Maguire-Jack et al., 2015), this is an important limitation. Additionally, while far from racially monolithic, rural and suburban communities in the US have higher proportions of White residents (80% and 68%) than do urban communities (44%; Pew Research Center, 2018), suggesting the present study could be missing facets of the relationship between racialized poverty and CPS contact in communities with more White and fewer Black residents. The original FFCWS data included a larger sample of Black than White families (respectively 50% and 25% of the sample), and both the differences in sample size and the relatively small White sample may have impacted our results. Particularly, our estimates within the White sample are likely less precise because of analytic power limitations in predicting the relatively low-incidence outcome of CPS contact. Additionally, our analysis focuses on families’ experiences between approximately 2000 and 2005, and CPS policy at the national, state, and county levels has developed in numerous ways since then. Nonetheless, inequities in CPS contact between Black and White children have persisted. While the specific policy implications of our findings are different than they might have been in 2005, the core inequity we investigated remains.

Our measures were also limited in two areas. First, our measure of CPS contact is self-reported and extends over a 5-year reference period, so it may be subject to both recall and social desirability bias. Other work suggests that the 5-year parent-reports of CPS contact in FFCWS are comparable to self-reports in similar surveys (Slack et al., 2011). Our analysis suggests that, compared to national estimates (Kim et al., 2017), FFCWS data capture less than half of officially reported CPS contact (FFCWS contact rate by age 5: 7%; national estimate of CPS investigation by age 5: 18%), with similar rates of underreporting among the Black and White families in our sample (40% and 48% of national estimates respectively). An additional limitation reflects the challenges of measuring structural racism. While we conceptualized racial inequities in poverty and in CPS contact as ultimately rooted in systemic racism, we did not measure racism directly. We included city and state covariates that we believe reflect racialized policy design as one attempt to concretize some components of systemic racism. However, we could not draw conclusions about specific associations of racism with CPS contact beyond the conceptual framing we brought to this study in articulating the perspective that systematic racial inequities in poverty and other sources of risk are arguably the consequences of structural racism, and considering them as such offers new insights into racial inequities in CPS contact.

Conclusion

Drawing on household-level data on racialized group membership and poverty and considering within-race differences in the associations between poverty and CPS contact, this study contributes new evidence on racial inequities in CPS contact. While not the first to underscore the fundamental role of racism in producing racial inequities in CPS involvement, our work aimed to reinforce the critical importance of treating race as a proxy for exposure to racism and not as a causal factor in explaining racial inequities. To avoid perpetuating racist framings of Black families as inherently risky, CPS research can consistently articulate the systemic racism at the root of inequities in exposure to risk. The broad implication of our study is the need for multi-pronged policy change to redress and eliminate structural racism, encompassing criminal justice and police reform, income and wealth inequality, access to early childhood education and care, physical and mental health care, and much more. Our results suggest the urgency of jointly addressing racialized poverty, through generously expanding family and income support policies like the child tax credit, subsidized child care, and EITC, which can particularly reduce poverty among Black children (Collyer et al., 2021), and simultaneously halting the policing of Black families, through reform of CPS (Waldfogel, 1998) or its abolition (Dettlaff, 2020). Pairing such actions could serve to reduce the racial inequities in poverty that our findings suggest may link systemic racism with racial inequities in CPS contact and also to address directly the persistent, inequitable contact with CPS systems Black families face.

Supplementary Material

Supplementary Material

Footnotes

1

We use the terminology “racialized group” to indicate the active social processes which result in categorizing people into groups characterized by phenotypical, social, and cultural factors that are summarized as ‘race.’

2

Those cities are: Austin, TX; Baltimore, MD; Boston, MA; Chicago, IL; Corpus Christi, TX; Detroit, MI; Indianapolis, IN; Jacksonville, FL; Milwaukee, WI; Nashville, TN; New York, NY; Newark, NJ; Norfolk, VA; Oakland, CA; Philadelphia, PA; Pittsburgh, PA; Richmond, VA; San Antonio, TX; San Jose, CA; and Toledo, OH.

3

The FFCWS cities align closely with county boundaries, and some of our city measures relied on county data.

4

In sensitivity tests, we also examined family characteristics measured at the earliest wave available.

5

Composite International Diagnostic Interview Short Form (CIDI-SF) (Princeton University, 2019)

6

77% of 5-year interviews were conducted between July 2004 and July 2005.

7

We used 2018 CPS and population data for Pittsburgh and Philadelphia as data were not available prior.

8

Measured using pooled data from 2000 through 2010.

9

Additionally, in a series of robustness checks, we examined several other state factors, including city unemployment rates (available as a restricted appendage file to the FFCWS; Princeton University, 2020); state EITC generosity (University of Kentucky Center for Poverty Research, 2020); city poverty rates among families with young children (Ruggles et al., 2020); and prevalence of historic redlining at the city-level (Nelson et al., 2020).

References

  1. Agan AY, & Starr SB (2016). Ban the box, criminal records, and statistical discrimination: A field experiment. SSRN Electronic Journal. [Google Scholar]
  2. Alexander M (2010). The new Jim Crow: Mass incarceration in the age of colorblindness. New Press. [Google Scholar]
  3. Andrews R, Casey M, Hardy BL, & Logan TD (2017). Location matters: Historical racial segregation and intergenerational mobility. Economics Letters, 158, 67–72. [Google Scholar]
  4. Annie E Casey Foundation. (2021). Kids Count Data Center. https://datacenter.kidscount.org
  5. Barth R, Miller J, Green R, & Joy N (2001). Children of color in the child welfare system. University of North Carolina. [Google Scholar]
  6. Bartholet E, Wulczyn F, Barth RP, & Lederman C (2011). Race and child welfare. Chapin Hall. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1889239 [Google Scholar]
  7. Berger LM, Cancian M, Cuesta L, & Noyes JL (2016). Families at the intersection of the criminal justice and child protective services systems. The ANNALS of the American Academy of Political and Social Science, 665(1), 171–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Centers for Disease Control and Prevention. (2020). Bridged-race population estimates 1990–2018. https://wonder.cdc.gov/Bridged-Race-v2018.HTML
  9. Collyer S, Curran MA, Hartley RP, Parolin Z, & Wimer C (2021). The Potential Poverty Reduction Effect of the American Families Plan. Center on Poverty and Social Policy, Columbia University. [Google Scholar]
  10. Dale MK (2014). Addressing the underlying issue of poverty in child neglect cases. American Bar Association. [Google Scholar]
  11. Derenoncourt E, & Montialoux C (2020). Minimum wages and racial inequality. The Quarterly Journal of Economics. [Google Scholar]
  12. Dettlaff AJ, Rivaux SL, Baumann DJ, Fluke JD, Rycraft JR, & James J (2011). Disentangling substantiation: The influence of race, income, and risk on the substantiation decision in child welfare. Children and Youth Services Review, 33(9). [Google Scholar]
  13. Dettlaff AJ, Weber K, Pendleton M, Boyd R, Bettencourt B, & Burton L (2020). It is not a broken system, it is a system that needs to be broken: The upEND movement to abolish the child welfare system. Journal of Public Child Welfare, 14(5), 500–517. [Google Scholar]
  14. DeWitt L The decision to exclude agricultural and domestic workers from the 1935 Social Security Act. Social Security Bulletin 2010; 70: 49–68. [PubMed] [Google Scholar]
  15. Drake B, Jolley JM, Lanier P, Fluke J, Barth RP, & Jonson-Reid M (2011). Racial bias in child protection? A comparison of competing explanations using national data. Pediatrics, 127(3), 471–478. 10.1542/peds.2010-1710 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Drake B, & Rank MR (2009). The racial divide among American children in poverty: Reassessing the importance of neighborhood. Children and Youth Services Review, 31(12), 1264–1271. 10.1016/j.childyouth.2009.05.012 [DOI] [Google Scholar]
  17. Fong K (2019). Concealment and constraint: Child protective services fears and poor mothers’ institutional engagement. Social Forces, 97(4), 1785–1810. [Google Scholar]
  18. Font SA, Berger LM, & Slack KS (2012). Examining racial disproportionality in child protective services case decisions. Children and Youth Services Review, 34(11). [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Foster CH (2008). The welfare queen: Race, gender, class, and public opinion. Race, Gender, & Class, 15(3/4), 162–179. [Google Scholar]
  20. Harris MS, & Hackett W (2008). Decision points in child welfare: An action research model to address disproportionality. Children and Youth Services Review, 30(2), 199–215. [Google Scholar]
  21. Heflin CM, & Iceland J (2009). Poverty, Material Hardship, and Depression. Social Science Quarterly, 90(5), 1051–1071. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hoynes HW, & Patel AJ (2018). Effective policy for reducing poverty and inequality?: The Earned Income Tax Credit and the distribution of income. Journal of Human Resources, 53(4), 859–890. 10.3368/jhr.53.4.1115.7494R1 [DOI] [Google Scholar]
  23. Jones CP (2000). Levels of racism: A theoretic framework and a gardener’s tale. American Journal of Public Health, 90(8), 1212–1215. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Jonson-Reid M, Drake B, & Kohl PL (2009). Is the overrepresentation of the poor in child welfare caseloads due to bias or need? Children and Youth Services Review, 31(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Karpman M, Zuckerman S, & Gonzalez D (2018). Material hardship among nonelderly adults and their families in 2017. Urban Institute. [Google Scholar]
  26. Kim H, Wildeman C, Jonson-Reid M, & Drake B (2017). Lifetime prevalence of investigating child maltreatment among US children. American Journal of Public Health, 107(2), 274–280. 10.2105/AJPH.2016.303545 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Klein S & Merritt DH (2014). Neighborhood racial & ethnic diversity as a predictor of child welfare system involvement. Children and Youth Services Review, 41, 95–105. [Google Scholar]
  28. Kovski NL, Hill HD, Mooney SJ, Rivara FP, Morgan ER, & Rowhani-Rahbar A (2021). Association of state-level earned income tax credits with rates of reported child maltreatment, 2004–2017. Child Maltreatment. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Krieger N (2014). Discrimination and health inequities. International Journal of Health Services, 44(4), 643–710. 10.2190/HS.44.4.b [DOI] [PubMed] [Google Scholar]
  30. Lanier P, Maguire-Jack K, Walsh T, Drake B, & Hubel G (2014). Race and ethnic differences in early childhood maltreatment in the United States. Journal of Developmental & Behavioral Pediatrics, 35(7), 419–426. [DOI] [PubMed] [Google Scholar]
  31. Lee T (2016). Processes of racialization in New York City’s child welfare system. City & Society, 28(3), 276–297. [Google Scholar]
  32. Lu YE, Landsverk J, Ellis-Macleod E, Newton R, Granger W, & Johnson I (2004). Race, ethnicity, and case outcomes in child protective services. Children and Youth Services Review, 26(5), 276–297. [Google Scholar]
  33. Maguire-Jack K, Lanier P, Johnson-Motoyama M, Welch H, & Dineen M (2015). Geographic variation in racial disparities in child maltreatment: The influence of county poverty and population density. Child Abuse & Neglect, 47, 1–13. [DOI] [PubMed] [Google Scholar]
  34. Mallon GP (2020). Black and brown children’s and families’ lives matter. Child Welfare, 98(3), v–ix. [Google Scholar]
  35. Nelson RK, Winling L, Marciano R, Connolly N, & et al. (2020). Mapping Inequality. American Panorama. https://dsl.richmond.edu/panorama/redlining [Google Scholar]
  36. Omeish Y, & Kiernan S (2020). Targeting bias to improve maternal care and outcomes for Black women in the USA. EClinicalMedicine, 27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Pac J, Garfinkel I, Kaushal N, Nam J, Nolan L, Waldfogel J, & Wimer C (2020). Reducing poverty among children: Evidence from state policy simulations. Children and Youth Services Review, 115. 10.1016/j.childyouth.2020.105030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Pargas DA (2009). Disposing of human property: American slave families and forced separation in comparative perspective. Journal of Family History, 34(3), 251–274. [Google Scholar]
  39. Parolin Z (2019). Temporary Assistance for Needy Families and the Black–White child poverty gap in the United States. Socio-Economic Review, 0(0), 1–31. [Google Scholar]
  40. Parke M (2004). Marriage-related provisions in welfare reauthorization proposals: A summary. Center for Law and Social Policy (CLASP). [Google Scholar]
  41. Pew Research Center (2018). Demographic and economic trends in urban, suburban and rural communities. https://www.pewsocialtrends.org/2018/05/22 [Google Scholar]
  42. Princeton University. (2019). User’s guide for the Fragile Families and Child Wellbeing Study https://fragilefamilies.princeton.edu/sites/fragilefamilies/files/year_5_guide.pdf
  43. Princeton University. (2020). FFCWS restricted use contract data > Labor market and macroeconomic. https://fragilefamilies.princeton.edu/restricted/macroeconomic
  44. Putnam-Hornstein E, Needell B, King B, & Johnson-Motoyama M (2013). Racial and ethnic disparities: A population-based examination of risk factors for involvement with child protective services. Child Abuse & Neglect, 37, 33–46. [DOI] [PubMed] [Google Scholar]
  45. Ratcliffe C (2015). Child poverty and adult success. Urban Institute. https://www.urban.org/research/publication/child-poverty-and-adult-success [Google Scholar]
  46. Reardon SF (2016). School segregation and racial academic achievement gaps. The Russell Sage Foundation Journal of the Social Sciences, 2(5), 34–57.30775444 [Google Scholar]
  47. Reichman NE, Teitler JO, Garfinkel I, & McLanahan SS (2001). Fragile Families: Sample and design. Children and Youth Services Review, 23(4–5), 303–326. [Google Scholar]
  48. Roberts DE (2002). Shattered bonds: The color of child welfare. New York: Basic Books. [Google Scholar]
  49. Rosenthal L, & Lobel M (2016). Stereotypes of Black American women related to sexuality and motherhood. Psychology of Women Quarterly, 40(3), 414–427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Rothstein R (2017). The color of law: A forgotten history of how our government segregated America. Liveright Publishing Corporation. [Google Scholar]
  51. Ruggles S, Flood S, Goeken R, Grover J, Meyer E, Pacas J, & Sobek M (2020). IPUMS USA, v. 10.1. IPUMS. https://usa.ipums.org/usa/index.shtml [Google Scholar]
  52. Slack KS, Berger LM, DuMont K, Yang M-Y, Kim B, Ehrhard-Dietzel S, & Holl JL (2011). Risk and protective factors for child neglect during early childhood: A cross-study comparison. Children and Youth Services Review, 33(8), 1354–1363. [Google Scholar]
  53. Tausanovitch C, & Warshaw C (2016). County-level public preference estimates. https://americanideologyproject.com/
  54. The Tax Policy Center. (2020). What are marriage penalties and bonuses? https://www.taxpolicycenter.org/briefing-book/what-are-marriage-penalties-and-bonuses
  55. University of Kentucky Center for Poverty Research. (2020). National Welfare Data. http://ukcpr.org/resources/national-welfare-data
  56. U.S. Department of Health & Human Services. (2020). Child maltreatment 2018. https://www.acf.hhs.gov/cb/data-research/child-maltreatment
  57. Vera Institute of Justice. (2020). Incarceration Trends, v. 2.1. https://github.com/vera-institute/incarceration-trends
  58. Waldfogel J (1998). The future of child protection: How to break the cycle of abuse and neglect. Cambridge, MA: Harvard University Press. [Google Scholar]
  59. Webb E, Maddocks A, & Bongilli J (2002). Effectively protecting black and minority ethnic children from harm: Overcoming barriers to the child protection process. Child Abuse Review, 11(6), 394–410. 10.1002/car.760 [DOI] [Google Scholar]
  60. Western B, & Wildeman C (2009). The Black family and mass incarceration. The Annals of the American Academy of Political and Social Science, 621(1), 221–242. [Google Scholar]
  61. Zuberi T (2001). Thicker than blood: How racial statistics lie. University of Minnesota Press [Google Scholar]
  62. Zhang L, Nepomnyaschy L, and Simmel C (2021). So close yet so different: neighborhood inequality and child maltreatment. Child Abuse & Neglect. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material
NIHMS1796178-supplement-Supplementary_Material.docx (78KB, docx)

RESOURCES

OSZAR »