Families at the Intersection of the Criminal Justice and Child Protective Services Systems

Abstract

In this article, we first describe the incidence and prevalence of incarceration and CPS involvement in the United States. Second, we outline the reasons that the same individuals and families may be at risk for involvement in both systems and review the limited existing research examining links between incarceration and CPS involvement. Third, we use unique longitudinal data from Wisconsin, spanning from 2004 to 2012, to describe intergenerational and intragenerational overlap in the two systems. Specifically, we calculate (1) the proportion of all CPS-involved children who have an incarcerated parent; (2) the proportion of incarcerated adults who have a CPS-involved child; (3) the proportion of incarcerated young men and women who were involved in the CPS system as adolescents; and (4) the proportion of CPS-involved adolescents who subsequently became incarcerated. We conclude with a discussion of potential directions for future research as well as implications for practice and policy.

The adult incarceration rate in the United States is higher than that in any other industrialized country and it is well known that individuals from socially and economically disadvantaged groups are disproportionately likely to spend time in jail or prison. Less well known is that involvement with child protective services (CPS) is relatively common in the United States as well. CPS involvement also occurs disproportionately for disadvantaged groups. Involvement in one or both of these systems is likely to have a range of independent and interactive influences on parents, children, and families. Furthermore, involvement in one may be associated with subsequent involvement in the other, either intergenerationally (such that a parent is incarcerated and his or her children are the subject of a CPS case) or intragenerationally (such that CPS-involved children subsequently experience incarceration as adults). Yet, there is a dearth of research examining the extent to which individuals and families are sequentially or simultaneously involved in both systems.

The existence of large incarcerated and CPS-involved populations, coupled with the fact that socially and economically disadvantaged groups are disproportionately likely to be involved in each system, highlights a broad range of policy issues. As such, the interaction of the child welfare system and the criminal justice system in the lives of low-income populations has increasingly received attention from policymakers and scholars alike. To date, however, relatively few studies have investigated the extent to which there is overlap between parental incarceration and child CPS involvement¹ despite evidence that increased parental—particularly maternal—incarceration between 1985 and 2000 has been linked to considerable growth in the foster care caseload during that period (Swann and Sylvester 2006). Furthermore, whereas a handful of studies examine the relation between juvenile delinquency and later perpetration of child maltreatment (Colman, Mitchell-Herzfeld, Kin, and Shady 2010; Phillips, Leathers, and Erkanli 2009), little is known about the extent to which incarcerated adults themselves experienced CPS involvement as children.

A better understanding of the intersection between CPS involvement and incarceration can improve services directed to families interacting with either or both the child welfare and criminal justice systems. Yet, a number of measurement and methodological issues have constrained research in this area. One major obstacle has been a lack of matched CPS and incarceration data through which to examine overlap between parental incarceration and child involvement in CPS. Another has been a lack of longitudinal data that combines information on CPS experiences and subsequent incarceration for the same individuals. Our work extends prior research by exploiting unique, matched longitudinal data from Wisconsin to examine overlap between parental incarceration and child CPS involvement, as well as between adolescent CPS involvement and subsequent incarceration in young adulthood. To the best of our knowledge, this is the first U.S.-focused study to use matched longitudinal administrative data on the universe of CPS-involved and incarcerated individuals (in our case, in the State of Wisconsin) to examine both intergenerational and intragenerational overlap in incarceration and CPS involvement.

In the sections that follow, we first briefly describe the incidence and prevalence of both incarceration and CPS involvement in the United States. Second, we outline the potential reasons that the same individuals and families may be at risk for involvement in both systems and review the limited existing research examining links between incarceration and CPS involvement. Third, we use unique longitudinal data from Wisconsin, spanning from 2004 to 2012, to describe intergenerational and intragenerational overlap in the two systems. Specifically, we calculate (1) the proportion of all CPS-involved children who have an incarcerated parent; (2) the proportion of incarcerated individuals who have a CPS-involved child; (3) the proportion of incarcerated young men and women who were involved in the CPS system as adolescents; and (4) the proportion of CPS-involved adolescents who subsequently became incarcerated. We conclude with a discussion of potential directions for future research as well as implications for practice and policy.

Incidence and Prevalence of Incarceration and CPS Involvement

Incarceration

Despite a downward trend in the U.S. prison population in recent years, the absolute number of individuals who are incarcerated remains large. Nearly 1.6 million people were in a prison facility at the end of 2012; the majority of these individuals were black males age 39 or younger (Carson and Golinelli 2013). The total incarceration rate (considering sentenced state and federal prisoners) among U.S. residents age 18 and older in 2012 was 0.48 percent. However, there is substantial variation in imprisonment by race and gender. For example, 0.91 percent of men, but only 0.06 percent of women, were incarcerated in 2012. Furthermore, incarceration rates for black and white adult males were 2.84 percent and 0.46 percent, respectively, whereas incarceration rates for black and white adult females were 0.12 percent and 0.05 percent (Carson and Golinelli 2013). In other words, the racial gap in incarceration rates is greater for men than women, while the gender gap is greater for blacks than whites. By comparison, Wisconsin incarceration rates (in 2005) were 0.42 percent for white adults and 4.42 percent for black adults (Maure and King 2007)—a larger than average racial gap.

Recent national estimates of the prevalence of parental imprisonment suggest that about 25 percent of black children and 4 percent of white children born in 1990 experienced parental incarceration by age 14. Incarceration rates are substantially higher for fathers than mothers (25.1 percent and 3.3 percent for black children’s fathers and mothers; 3.6 percent and 0.6 percent for white children’s fathers and mothers), and especially for less educated parents within each group (Wildeman 2009).

CPS Involvement

An estimated 6.3 million children, representing about 4.6 percent of the child population, were reported to CPS agencies in the United States in 2012. Just under 20 percent of these children were substantiated or indicated for child abuse or neglect, with the vast majority having experienced child neglect (U.S. Department of Health and Human Services [USDHHS] 2012). In all, more than 250,000 children entered and more than 460,000 were residing in some form of out-of-home placement (OHP) as a result of CPS involvement in 2012 (USDHHS 2013). By comparison, 3 percent of Wisconsin children were reported to CPS in 2012, 13 percent of whom were substantiated (WI Department of Children and Families 2013a). The overall victimization (substantiation) rates for black and white children in Wisconsin in 2012 were 0.82 percent and 0.25 percent, respectively (USDHHS 2012). Approximately 4,453 Wisconsin children entered and 6,255 spent time in OHP in 2012 (WI Department of Children and Families 2013b).

The prevalence of CPS involvement and OHP over the course of childhood is much higher. Current estimates suggest that 13 percent of all U.S. children and 21 percent of black U.S. children are found to be (substantiated as) maltreatment victims by CPS at some point between their birth and age 18 (Wildeman et al. 2014). Furthermore, 6 percent of all U.S. children and 12 percent of black U.S. children experience a foster care placement by age 18 (Wildeman and Emanuel 2014). Thus, CPS involvement and OHP are somewhat common experiences during childhood in the United States, particularly for black and, more generally, disadvantaged children.

Links between Incarceration and CPS Involvement

As noted above, incarceration and CPS involvement may intersect both intergenerationally and intragenerationally. As such, we focus on overlap between parental incarceration and child CPS involvement as well as between adolescent CPS involvement and subsequent incarceration (during young adulthood).

Intergenerational Overlap between Incarceration and CPS Involvement

There are multiple reasons why parental incarceration and child CPS involvement may be linked. First, parents may become incarcerated for engaging in child abuse or neglect that constitutes criminal behavior. In such cases, CPS involvement may directly and mechanically cause parental incarceration. However, existing empirical evidence suggests that perpetrator incarceration for child maltreatment is relatively rare. For example, results from Cross and colleagues’ (2003) meta-analysis of child maltreatment prosecution studies suggest that roughly 56 percent of substantiated child maltreatment cases are referred for prosecution, of which 26 percent result in incarceration. Given that about 20 percent of all CPS referrals are substantiated (USDHHS 2012), this suggests that incarceration of perpetrators (not all of whom are parents) occurs in only about 3 percent of all CPS referrals. Thus, parental incarceration as a result of child maltreatment is not likely to be a primary cause of intergenerational overlap in CPS involvement and incarceration.

Second, children whose parents become incarcerated may be taken into custody by CPS if another adequate caregiver has not been arranged prior to the parent’s incarceration. This may be particularly true in instances of maternal (rather than paternal) incarceration because most children live with their mothers (who are substantially more likely than fathers to be primary or sole caregivers). As such, children are considerably more likely to be in out-of-home care (living with neither parent) when their mother is incarcerated than when their father is incarcerated (Johnson and Waldfogel 2002; Glaze and Maruschak 2010; Mumola 2000). Current estimates from surveys of inmates in state prison indicate that 2.2 percent of fathers and 10.9 percent of mothers report having children in OHP (foster care, agency, or institution) (Glaze and Maruschak 2010).

Maternal incarceration may be directly associated with an increased probability of CPS involvement for children. It may also be indirectly linked to CPS involvement if the quality of care and investments children receive from substitute caregivers during their mother’s imprisonment places them at increased risk of CPS involvement. That is, CPS may become involved with children whose mothers are incarcerated because they are potentially living in an unsafe environment or are being supported by inadequate resources (Phillips, Dettlaff, and Baldwin 2010). Whereas paternal incarceration rates are several orders of magnitude larger than maternal incarceration rates, growth in maternal incarceration has substantially outpaced growth in paternal incarceration in recent years. Indeed, the number of children with incarcerated mothers more than doubled over the last 25 years (Glaze and Maruschak 2010), and Swann and Sylvester (2006) estimate that about 30 percent of the growth in the foster care population between 1985 and 2000 resulted from increased female incarceration (they found no effect of increased male incarceration). On the whole, the limited evidence to date suggests that maternal incarceration is positively and relatively strongly associated with child CPS involvement in the United States at both the aggregate (Swann and Sylvester 2006) and individual levels (Dallaire 2007; Johnson and Waldfogel 2002, 2004; Glaze and Maruschak 2010), although there is no micro-level evidence that this association is likely to be causal.

Relative to maternal incarceration, paternal incarceration should less often directly lead to CPS involvement as a result of children losing their sole or primary caregiver. However, paternal incarceration may indirectly increase risk of CPS involvement in four ways: (1) by reducing family economic stability vis-à-vis fathers’ income or child support payment (Chung 2012); (2) by reducing maternal well-being and social support (Foster 2012; Foster and Hagan 2013); (3) by leading to family instability (new maternal romantic relationships) (Cancian, Chung, and Meyer 2015; Foster and Hagan 2007; see also the articles by Apel, Andersen, and Wildeman, Turney, and Yi in this volume); and (4) by leading to poorer post-incarceration parenting (Andersen and Wildeman 2014; Foster 2012; Foster and Hagan 2013; Turney 2014; Turney, Schnittker, and Wildeman 2012). Each of these factors—economic instability, limited maternal well-being and social support, family instability, and low-quality parenting—is associated with increased child maltreatment and CPS involvement (see, e.g., Berger and Slack 2013; Berger and Waldfogel 2011; Stith et al. 2009). We are aware of no U.S. study to explicitly examine associations between paternal incarceration and CPS involvement at the micro level using data on both fathers and children. However, Andersen and Wildeman’s (2014) analyses leveraging Danish data and a natural experiment indicate that paternal incarceration is associated with a large increase (roughly a doubling) in the risk of foster care placement. This evidence is suggestive of a causal relation.

Third, there may be social selection such that the same types of parents who are likely to become incarcerated are also likely to be involved with CPS for suspected or confirmed child maltreatment, independent of being incarcerated. Current evidence suggests that children whose caregivers have been arrested (whether or not they are subsequently incarcerated) are exposed to greater levels of economic insecurity, parental mental health problems, substance abuse, and domestic violence than both other children involved in CPS and children in the general population (Phillips and Dettlaff 2009). These same factors are independently predictive of child maltreatment and CPS involvement (Berger and Slack 2013; Berger and Waldfogel 2011; Stith et al. 2009). Furthermore, parents’ criminal behaviors may directly place children at risk of CPS involvement for various types of maltreatment. For example, exposing children to domestic violence or illicit activities such as prostitution and the manufacture, sale, or purchase of illegal drugs constitutes child maltreatment in many states.

It is also possible that socially and economically disadvantaged individuals and families may be systematically more likely to experience both CPS involvement and incarceration than their more advantaged counterparts exhibiting identical behaviors. This may reflect differential levels of surveillance, monitoring, or exposure to law enforcement and/or child maltreatment reporters based on race or socioeconomic status. It may also reflect differential decision making in the criminal justice and/or child welfare systems based on these factors. Indeed, large literatures have documented that there are racial and socioeconomic disparities in both CPS and criminal justice involvement. Furthermore, evidence suggests that parental criminal history is used by CPS as a marker for child maltreatment risk and, thereby, influences CPS decisions to substantiate maltreatment or remove children (Phillips, Dettlaff, and Baldwin 2010). Taken together, these factors have important implications for intergenerational transmission of social and economic disadvantage, as well as social exclusion (Foster and Hagan 2007, 2009), which may at least in part reflect disproportionate representation of poor and minority families in both systems.

Finally, it is important to note that parental incarceration may lead to poorer quality parenting and caregiving environments and, thereby, child maltreatment and CPS involvement. If so, children may continue to be at risk of CPS involvement even after parents are released from prison. A large body of research has established that incarceration is associated with a range of adverse post-release outcomes spanning employment and earnings, relationship instability and dissolution, paternal financial contributions to children, material hardship, social support, housing instability and homelessness, and health and mental health (for reviews see Andersen and Wildeman 2014; Foster and Hagan 2007, 2009; Wildeman 2009, 2012; Wildeman and Muller 2012). Deficits in these areas are, again, positively associated with child maltreatment (Stith et al. 2009). Existing empirical evidence also suggests that, even after parents are released from jail or prison, children in OHP may remain there, in part because their parents face limited economic resources, social supports, and unstable housing (Katz 1998).

Intragenerational Overlap between CPS Involvement and Incarceration

Experiencing child maltreatment (and subsequent CPS involvement) is associated with a range of social and emotional problems that may place adolescents at increased risk of engaging in delinquent or criminal behavior. Existing empirical evidence suggests associations of child maltreatment with subsequent behavior problems; post-traumatic stress disorder; substance abuse; poor cognitive development, achievement, and school-related outcomes; poor economic outcomes (earnings/income, work trajectory, occupation, wealth); intergenerational transmission of maltreatment; delinquency and violence; and criminal behavior (see, e.g., Berger and Slack 2013; Berger and Waldfogel 2010; Font and Berger 2015; Gilbert et al. 2009; Krug et al. 2002; World Health Organization 2006). Indeed, the trauma and stress associated with such childhood experiences are associated with poorer social psychological development and functioning during emerging adulthood, particularly for young women (Foster and Hagan 2007; Foster, Hagan, and Brooks-Gunn 2008). Although research has not established whether these associations are causal, these factors are also associated with incarceration. In addition, as noted above, there may be social selection such that the same types of adolescents who are likely to have been involved in CPS are also likely to engage in delinquent or criminal behavior, independent of CPS involvement. There may also be systematic explanations for why particular groups of adolescents are disproportionately likely to both become engaged with CPS and to subsequently be incarcerated, all else equal (e.g., conditional on equivalent experiences and behaviors relative to other groups), including differences in surveillance, monitoring, or exposure as well as decision making in the criminal justice and/or child welfare systems.

Few empirical studies have explicitly addressed links between overall CPS involvement as a child and subsequent incarceration. However, a large body of research documents that individuals who spent time in OHP as children are disproportionately likely to be incarcerated as adults. In the most rigorous study to date, Doyle (2008) uses an instrumental variables strategy, identified by random variation in assignment to CPS caseworkers with varying propensities for child removal, to estimate the causal effects of OHP on criminal justice system involvement and incarceration and finds them to be large. However, this evidence relates only to OHP.

Inter- and Intragenerational Overlap in Incarceration and CPS Involvement in Wisconsin

We use longitudinal data from Wisconsin, spanning from 2004 to 2012, to describe inter- and intragenerational overlap in incarceration and CPS involvement. We first calculate the proportion of all CPS-involved children who have an incarcerated parent and the proportion of incarcerated adults who have a CPS-involved child. We then calculate the proportion of incarcerated young men and women who were involved in the CPS system as adolescents and the proportion of CPS-involved adolescents who subsequently became incarcerated (as a young adult). We focus on two measures of CPS involvement: screened-in CPS reports and OHPs. Furthermore, because we have statewide data on incarceration in state prison, but data on incarceration in county jail only for Milwaukee County, we present results separately for statewide samples (for analyses involving incarceration in state prison) and Milwaukee County samples (for analyses involving incarceration in state prison, in county jail, and in either type of institution).

Data

Our data are drawn from the 2013 Multi-Sample Person File (MSPF) housed at the Institute for Research on Poverty (IRP) at the University of Wisconsin–Madison. The MSPF includes linked administrative data on the full universe of participants in most State of Wisconsin social welfare programs (see Brown et al. 2011 for a full description of the MSPF database), such as Wisconsin’s Temporary Assistance for Needy Families (TANF) program; Supplemental Nutrition Assistance Program (SNAP); Medicaid; child support enforcement; child protective services; and Unemployment Insurance (wage and benefit data). It also includes data from the Department of Corrections (on state prison incarceration) and Milwaukee County Sheriff (on incarceration in Milwaukee County jails).

Inclusion in the MSPF necessitates that an individual or his or her family or household members participated in one or more of these programs at some point. Any individual who is observed as being part of a social welfare benefit case (i.e., is a member of a family or household unit considered for eligibility or benefit determination) is included, regardless of being the actual benefit recipient. Once an individual enters the MSPF, they continue to be included, backward and forward in time, regardless of prior or ongoing benefit receipt. For example, once an individual enters the MSPF, their earnings can be obtained (through UI wage records) for all subsequent and prior periods regardless of program participation in those periods. Administrative data on incarceration in state prison facilities and Milwaukee County Jail is available from 1993 through 2012. Unfortunately, however, data on child welfare involvement is available only from 2004 through 2012, based on the development of the current data system known as the Wisconsin Statewide Automated Child Welfare System, or WiSACWIS. Given this, we focus our analyses only on the 2004 to 2012 period.

Each of our analyses uses different samples, which are summarized in Table 1. We first calculate the proportion of all CPS-involved children and all children in OHP in Wisconsin between 2004 and 2012 (both statewide and in Milwaukee County) who had a parent who was incarcerated within 12 months of the date of a screened-in CPS report or the date they began an OHP spell. Specifically, we identify whether a child was observed to have an incarcerated parent in same month as a screened-in CPS report (or the first month of an OHP spell) or in any of the subsequent 11 months, regardless of whether the incarceration period began prior to the CPS event. To do so, we use the full universe of CPS-involved children in each of those years, excluding only those children who had missing information on birthdate or gender (< 1 percent of all reports) or on CPS referral or investigation date (~5 percent of all reports). Our statewide sample for this analysis includes 222,664 CPS-involved children, of whom 99,752 had experienced OHP. Our Milwaukee County sample includes 80,146 children who were involved in CPS in one or more of those years, of whom 45,199 had experienced OHP. We calculate the statewide proportion of CPS-involved children who had a parent in state prison, as well as the proportions of CPS-involved children in Milwaukee who had a parent in state prison, Milwaukee County Jail, and either institution. We present these figures for the full samples and by parent gender and child race. Note that, in all of our analyses, incarceration in Milwaukee County Jail and in state prison are not mutually exclusive; indeed; a considerable proportion of adults experience incarceration in both institutions over the course of the observation period.

Table 1.

Description of analysis samples

Population	Observation Period	Observations: Statewide	Observations: Milwaukee County
Intergenerational analyses
Analysis #1: Proportion of CPS-involved children with incarcerated parents
CPS-involved children	2004–2012	222,664	80,146
Children in OHP	2004–2012	99,752	45,199
Analysis #2: Proportion of incarcerated adults with CPS-involved children
Incarcerated adults	2004–2012	Total: 76,844	In State Prison
		Men: 69,850	Total: 30,377
		Women: 6,994	Men: 27,215
			Women: 3,162
			In Milwaukee County Jail
			Total: 88,242
			Men: 65,919
			Women: 22,323
			In Prison or Jail
			Total: 95,189
			Men: 72,008
			Women: 23,181
Intragenerational analyses
Analysis #3: Proportion of CPS-involved adolescents (age 15–16) who were incarcerated as young adults (age 18–21)
CPS-involved adolescents (age 15–16)	2004–2007 (age 15–16)	Total: 13,131	Total: 3,340
	2007–2012 (age 18–21)	Men: 5,203	Men: 1,266
		Women: 7,928	Women: 2,074
Adolescents in OHP (age 15–16)	2004–2007 (age 15–16)	Total: 6,133	Total: 1,681
	2007–2012 (age 18–21)	Men: 3,328	Men: 816
		Women: 2,805	Women: 865
Analysis #4: Proportion of incarcerated young adults (age 18–21) who were CPS involved as adolescents (age 15–16)
Incarcerated young adults (age 18–21)	2004–2007 (age 15–16)	Total: 9,605	In State Prison
	2007–2012 (age 18–21)	Men: 7,658	Total: 597
		Women: 1,947	Men: 570
			Women: 27
			In Milwaukee County Jail
			Total: 5,529
			Men: 4,051
			Women: 1,478
			In Prison or Jail
			Total: 6,126
			Men: 4,621
			Women: 1,505

Note: All data are drawn from the 2013 MSPF. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail.

Second, we calculate the proportion of incarcerated adults for whom one or more of their children was the subject of a CPS case within 12 months of the start of an incarceration spell. We focus on all incarcerated adults rather than incarcerated parents because incarcerated adults in the MSPF can only be identified as parents if one or more of their children was included in social welfare benefit program. Identifying parents in this way would undercount the proportion of incarcerated adults who are actually parents and, in particular, would exclude the most socioeconomically advantaged incarcerated parents (who are potentially least likely to have CPS-involved children). For these analyses, we use the universe of adults who were incarcerated in Wisconsin state prison (76,844; 69,850 men and 6,994 women), and the universe of Milwaukee County adults who were incarcerated in Wisconsin state prison (30,377 adults; 27,215 men; 3,162 women), in Milwaukee County Jail (88,242 adults; 65,919 men; 22,323 women), and in either state prison or Milwaukee County Jail (95,189 adults; 72,008 men; 23,181 women) at any point between 2004 and 2012. We present the proportion of all Wisconsin adults in state prison, as well the proportions of Milwaukee County parents in state prison, county jail, and either state prison or county jail, who had CPS-involved children. We also present these figures by gender and race.

Third, we calculate the proportion of young men and women who were involved in the CPS system as adolescents and were subsequently incarcerated in Wisconsin state prison or Milwaukee County Jail. Given that our CPS data span only from 2004 to 2012, we are able to observe during both childhood and adulthood only those cohorts of men and women with incarceration spells between ages 18 and 21 (in 2007 through 2012), and only whether they had CPS involvement when they were ages 15 to 16 (in 2004 through 2007). Thus, our sample for these analyses consists of the entire universe of 13,131 CPS-involved adolescents in Wisconsin (5,203 men and 7,928 women) and 3,340 CPS-involved adolescents in Milwaukee County (1,266 men and 2,074 women) between 2004 and 2007, excluding only those adolescents who had missing information on birthdate or gender (< 1 percent of all reports) or on CPS referral or investigation date (~5 percent of all reports). Of these, 6,133 statewide (3,328 men and 2,805 women) and 1,681 in Milwaukee County (816 men and 865 women) had experienced OHP as adolescents. We present the proportion of CPS-involved adolescents who were subsequently in state prison as young adults in Wisconsin, as well as the proportions of CPS-involved adolescents in Milwaukee County who were subsequently in state prison, Milwaukee County Jail, and either institution as young adults. We present these figures for the full sample as well as by gender and race/ethnicity.

Finally, we calculate the proportion of incarcerated 18- to 21-year-olds who had been CPS-involved between ages 15 and 16. The statewide sample for these analyses consists of the entire universe of 9,605 young adults (7,658 men and 1,947 women) in state prison in Wisconsin between 2007 and 2012, again excluding only those who had missing information on birthdate, gender, or on CPS referral or investigation date. The Milwaukee County sample includes the universe of 6,126 young adults (4,621 men and 1,505 women) in either state prison (597 young adults; 570 men and 27 women) or Milwaukee County Jail (5,529 young adults; 4,051 men and 1,478 women) during those years. We present the overall proportions of young adults incarcerated in Wisconsin state prison who were CPS involved as an adolescent, as well as the proportions of Milwaukee County young adults in state prison, Milwaukee County Jail, and either institution and who had been CPS involved as an adolescent. Again, we present results for the full sample as well as by gender and race/ethnicity.

Results

CPS-Involved Children with Incarcerated Parents

Figure 1 presents the proportion of all CPS-involved children (and those in OHP) in Wisconsin between 2004 and 2012 for whom one or both parents was incarcerated within 12-months of their screened-in CPS report (or the start of an OHP spell).² The first bar indicates that 8 percent of all children experiencing a screened-in report had a parent in state prison at some point during 12 months following the report. Parental incarceration in state prison was even more common for children whose CPS cases were substantiated (11 percent; not shown in figure), and particularly for those experiencing OHP, of which 11 percent had a parent in state prison.³ The remaining bars in Figure 1 present the proportions of CPS-involved children (and of children in OHP) in Milwaukee County who had incarcerated parents. The second set of bars shows the proportion of such children who had a parent in state prison; the third shows the proportion who had a parent in Milwaukee County Jail; and the fourth shows the proportion who had a parent in either state prison or Milwaukee County Jail. Parental incarceration in state prison was considerably more common for CPS-involved children (and children in OHP) in Milwaukee than for such children statewide, at 12 percent versus 8 percent (and 14 percent versus 11 percent), respectively. Furthermore, within the Milwaukee County sample, parental incarceration in county jail was much more common than parental incarceration in state prison (23 and 12 percent versus 27 and 14 percent for all CPS-involved children and children in OHP). In all, 28 percent of children with a screened-in CPS report and 34 percent of those in OHP in Milwaukee County had a parent who was incarcerated in either state prison or Milwaukee County Jail at some point during the 12 months following the screen-in date or start of the OHP spell, although the parent’s incarceration spell may have begun before such date.

Figure 1. Proportion of CPS-involved children with incarcerated parents, State of Wisconsin and Milwaukee County, 2004–2012.

Note: The proportion of children observed to have an incarcerated parent in same month as a screened in CPS report (or the first month of an OHP spell) or in any of the subsequent 11 months, regardless of whether the incarceration period began prior to the CPS event, is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of CPS-involved children (or children in OHP) in Wisconsin, 2004 to 2012; and (2) through (4) the universe of CPS-involved children (or children in OHP) in Milwaukee County, 2004–2012.

Figure 2 presents the proportion of all CPS-involved children and children in OHP who had an incarcerated parent, by parent gender and child race. Given that parental incarceration in state prison was more common in Milwaukee than statewide, and also that parental incarceration in the Milwaukee County Jail was much more common than parental incarceration in state prison, we focus our discussion of these (and all subsequent) results on findings from the statewide sample vis-à-vis state prison, and findings from the Milwaukee sample vis-à-vis incarceration in state prison or Milwaukee County Jail. Statewide, children screened-in by CPS were dramatically more likely to have their father (7 percent) than their mother (1 percent) incarcerated in state prison in the month of a CPS report or in the 11 months following a CPS report. Furthermore, about 9 percent of children in OHP had an incarcerated father and 2 percent had an incarcerated mother. For the Milwaukee county sample, 11 percent of CPS-involved children’s mothers and 22 percent of their fathers were incarcerated in either state prison or Milwaukee County Jail in the month of a CPS report or at some point in the 11 months following their CPS report. For children in OHP in Milwaukee County these figures were 17 percent and 24 percent, respectively. Turning to child race, CPS-involved black children (including those experiencing OHP placement) were substantially more likely to have an incarcerated parent than other CPS-involved children. Strikingly, 40 percent of black children who were screened in by CPS and 43 percent of black children who experienced OHP in Milwaukee County also experienced parental incarceration in either prison or jail in the month of a screen-in date or start of an OHP spell or within 11 months of a screen-in date or start of an OHP spell.

Figure 2. Proportion of CPS-involved children with incarcerated parents, State of Wisconsin and Milwaukee County, 2004–2012, by parent race and child gender.

Note: The proportion of children observed to have an incarcerated parent in same month as a screened in CPS report (or the first month of an OHP spell) or in any of the subsequent 11 months, regardless of whether the incarceration period began prior to the CPS event, is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of CPS-involved children (or children in OHP) in Wisconsin, 2004 to 2012; and (2) through (4) the universe of CPS-involved children (or children in OHP) in Milwaukee County, 2004–2012.

Incarcerated Adults with CPS-Involved Children

Figure 3 presents the proportion of incarcerated adults in Wisconsin between 2004 and 2012 who had CPS-involved children (and children in OHP) at the time of or within 11 months of the start of an incarceration spell. Again, we show figures for incarceration in state prison, for the statewide sample, as well as for incarceration in state prison, Milwaukee County Jail, and either institution for the Milwaukee sample; however, we limit our discussion to the first and last of these categories. Statewide, just over 15 percent of adults in prison had one or more CPS-involved children and just almost 6 percent had children in OHP. By comparison, nearly 18 percent of incarcerated (in either prison or jail) Milwaukee adults had CPS-involved children and nearly 6 percent had children in OHP either in the month of their incarceration or at some point during the 11 months following the start of an incarceration spell.

Figure 3. Proportion of incarcerated adults with CPS-involved children, State of Wisconsin and Milwaukee County, 2004–2012.

Note: The proportion of incarcerated parents observed to have a child with a CPS report (or to be in OHP) in the month that the incarceration spell began or in any of the subsequent 11 months is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of adults incarcerated in Wisconsin state prisons, 2004 to 2012; (2) the universe of Milwaukee County adults incarcerated in Wisconsin state prisons, 2004–2012; (3) the universe of adults incarcerated in Milwaukee County Jail, 2004–2012; and (4) the universe of Milwaukee county adults incarcerated in either Wisconsin state prison or Milwaukee County Jail, 2004–2012.

Figure 4 presents these figures by gender and race. Incarcerated women were substantially more likely than incarcerated men to have CPS-involved children (21 percent versus 15 percent for men and women statewide and 21 percent versus 16 percent for Milwaukee adults in either prison or jail) and children in OHP (12 percent versus 5 percent and 8 percent versus 5 percent). Black adults were more likely than white adults (and slightly more likely than adults of another race/ethnicity) to have CPS-involved children and children in OHP.

Figure 4. Proportion of incarcerated adults with CPS-involved children, State of Wisconsin and Milwaukee County, 2004–2012, by parent race and child gender.

Note: The proportion of incarcerated adults observed to have a child with a CPS report (or to be in OHP) in the month that the incarceration spell began or in any of the subsequent 11 months is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of adults incarcerated in Wisconsin state prisons, 2004 to 2012; (2) the universe of Milwaukee County adults incarcerated in Wisconsin state prisons, 2004–2012; (3) the universe of adults incarcerated in Milwaukee County Jail, 2004–2012; and (4) the universe of Milwaukee county adults incarcerated in either Wisconsin state prison or Milwaukee County Jail, 2004–2012

CPS-Involved Adolescents Who Were Incarcerated as Young Adults

Figure 5 shows rates of incarceration between ages 18 and 21 for young adults who were CPS-involved (and experienced OHP) between ages 15 and 16. Again, we focus our discussion on results from the statewide sample when considering incarceration in state prison, and the Milwaukee sample when considering overall incarceration (in either state prison or county jail).⁴ Almost 5 percent of CPS-involved adolescents in Wisconsin spent time in state prison between age 18 and age 21, whereas just under 29 percent of all CPS-involved Milwaukee adolescents spent time in either Milwaukee County Jail or Wisconsin state prison between age 18 and age 21.⁵ For those who experienced OHP as an adolescent, these figures were 10 percent and 34 percent, respectively. Young men who experienced CPS involvement (and OHP) as adolescents were substantially more likely to become incarcerated in young adulthood than young women who had such experiences. In addition, the Milwaukee sample data reveal that young black adults who were CPS involved as an adolescent were substantially more likely to be incarcerated than their counterparts of other races and ethnicities. However, this largely reflects a much greater rate of incarceration in the Milwaukee County Jail rather than a greater rate of incarceration in state prison.⁶ Indeed, within the Milwaukee County sample, black CPS-involved adolescents were less likely than those of other racial/ethnic groups to experience incarceration in state prison, but vastly more likely to experience incarceration in county jail.

Figure 5. Proportion of incarcerated young adults (age 18–21 in 2007–2012) who were CPS involved as adolescents (age 15–16 in 2004–2007), State of Wisconsin and Milwaukee County, 2004–2012.

Note: The proportion of young adults who were incarcerated between ages 18 and 21 and were also CPS-involved between ages 15 and 16 is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of CPS-involved 15 and 16 year-olds (or those in OHP) in Wisconsin in 2004 to 2007; and (2) through (4) the universe of CPS-involved 15 and 16 year-olds (or those in OHP) in Milwaukee County in 2004 to 2007.

Incarcerated Young Adults Who Were CPS-Involved as Adolescents

Figure 6 presents rates of CPS-involvement (and OHP) during adolescence for individuals who were incarcerated as young adults. On the whole, about 18 percent of all 18- to 21-year-olds in state prison in Wisconsin had been involved with CPS between ages 15 and 16, whereas 8 percent had experienced OHP. By comparison, 19 percent (and 7 percent) of Milwaukee young adults who were in either state prison or county jail had experienced CPS involvement (and OHP) as adolescents.⁷ Most strikingly, more than a quarter of all women between ages 18 and 21 who were in state prison had been involved with CPS as an adolescent, and about 9 percent had experienced OHP. For the Milwaukee sample, these figures were 32 percent and 10 percent with respect to incarceration in either state prison or county jail.⁸ The data reveal relatively little variation in prior (adolescent) CPS involvement or OHP experiences by race and ethnicity among incarcerated young adults, with the exception that, in the Milwaukee sample, black young adults in state prison were considerably less likely than nonblack young adults in state prison to have been involved in CPS or to have experienced OHP as an adolescent.

Figure 6. Proportion of incarcerated young adults (age 18–21 in 2007–2012) who were CPS involved as adolescents (age 15–16 in 2004–2007), State of Wisconsin and Milwaukee County, 2004–2012.

Note: The proportion of adolescents who were CPS-involved between ages 15 and 16 and were also incarcerated between ages 18 and 21 is presented. CPS-involvement and OHP are not mutually exclusive, nor are incarceration in state prison and Milwaukee County Jail. Denominators are: (1) the universe of 18 to 21 year-olds incarcerated in Wisconsin state prisons, 2007 to 2012; (2) the universe of Milwaukee County 18 to 21 year-olds incarcerated in Wisconsin state prisons, 2007–2012; (3) the universe of 18 to 21 year-olds incarcerated in Milwaukee County Jail, 2007–2012; and (4) the universe of Milwaukee county 18 to 21 year-olds incarcerated in either Wisconsin state prison or Milwaukee County Jail, 2007–2012.

Implications for Research and Policy

Incarceration of adults and CPS involvement for children and families are relatively common in the United States, particularly among disadvantaged populations. Involvement in each system is associated with a similar set of environmental factors (not addressed here), as well as individual and family characteristics, including limited economic resources and psychosocial problems. Parental incarceration may also directly lead to child CPS involvement and child CPS involvement may be a marker for potential risk of future incarceration. These factors point to the importance of understanding both inter- and intragenerational overlap in incarceration and CPS involvement, which has increasingly drawn the attention of policy makers and scholars. Yet, few existing studies explicitly address these issues, largely due to limitations of existing data.

Our descriptive analyses using linked data on the universes of adults incarcerated in state prison and Milwaukee County Jail, and children involved in CPS from 2004 to 2012 in Wisconsin yield new insights into the degree of inter- and intragenerational overlap between incarceration and CPS involvement. First, we find that, across the state of Wisconsin, more than 8 percent of CPS involved children and nearly 11 percent of children in OHP have a parent who is incarcerated in state prison in the month of or 11 months following a screened-in CPS report or the start of an OHP. In Milwaukee County, fully 28 percent of all CPS-involved children and 34 percent of children in OHP have a parent in jail or prison. That the Milwaukee figures are so much higher than the statewide figures likely reflects three factors. To begin with, the Milwaukee figures include incarceration in both prison and jail, whereas the statewide figures include only incarceration in state prison. In addition, the Milwaukee population is both much more urban and considerably more disadvantaged than the balance of the state. Finally, Milwaukee has the largest concentration of blacks in the state and, as discussed above, blacks are disproportionately represented in both the CPS and incarceration populations. Nonetheless, both sets of figures represent a large number of children given that more than 6 million children are reported to CPS each year (USDHHS 2012) and that nearly one-half million children are in OHP in a given year (USDHHS 2013). Consistent with what we know about incarceration rates in general, these children’s fathers are considerably more likely to be incarcerated than their mothers, and black children are substantially more likely than white children to have an incarcerated parent. Black children are also disproportionately likely to be CPS involved. These analyses indicate that a considerable proportion of CPS-involved children have incarcerated parents.

Second, statewide more than 15 percent of incarcerated adults have children who become CPS-involved in the month of or 11 months following their incarceration; almost 6 percent have children in OHP. In Milwaukee County, these figures are 18 percent and 6 percent when incarceration in both state prison and county jail are considered. Again, this suggests that a large number of children and families are affected given that about 1.6 million adults are incarcerated in the United States (Carson and Golinelli 2013). In addition, female prisoners are substantially more likely than male prisoners to have a CPS-involved child following their incarceration and black prisoners are disproportionately likely to have CPS-involved children. Taken together, our findings regarding intergenerational overlap between parental incarceration and child CPS involvement lend support to prior theoretical and empirical work suggesting that parental incarceration may contribute to intergenerational transmission of disadvantage and social exclusion (Foster and Hagan 2007, 2009, 2013), potentially as a result of individual and family stress (Foster 2012; Foster and Hagan 2013).

Third, roughly 5 percent of all CPS-involved 15- to 16-year-olds in Wisconsin and nearly 10 percent of 15- to 16-year-olds who experienced OHP were subsequently incarcerated in state prison between age 18 and age 21. In Milwaukee, about 29 percent of all CPS-involved 15- to 16-year-olds and nearly 34 percent of 15- to 16-year-olds who experienced OHP were in jail or prison at some point between age 18 and age 21; rates of incarceration for CPS-involved adolescents are particularly high for men and black individuals. These are extremely high proportions. Furthermore, they may even underestimate the full extent of overlap between CPS involvement and childhood and adult incarceration given that we are able only to observe CPS involvement at ages 15 and 16 (and do not observe CPS involvement at younger ages) and incarceration from age 18 to age 21 (and do not observe incarceration beyond age 21). In addition, individuals who are incarcerated by age 21, and particularly those who are in state prison by age 21, may be an especially select group of the most troubled young adults. If so, it is possible that those who are incarcerated at older ages will be less likely to have experienced CPS involvement during their youth than the young adults we observe. Future research in this area can shed better light on rates of overlap based on age at incarceration as well as reasons for intragenerational overlap between CPS involvement and incarceration.

Finally, and perhaps most strikingly, incarcerated young adults ages 18 to 21 are extremely likely to have experienced CPS involvement at ages 15 to 16; this is especially true for incarcerated women. Statewide, approximately 18 percent of all incarcerated 18- to 21-year-olds were CPS-involved as adolescents and 8 percent spent time in OHP. Whereas this was true for 15 percent and 8 percent of men, it was true for 29 percent and 9 percent of women. Figures for Milwaukee were roughly comparable, though modestly higher. These findings support prior theory and empirical research suggesting that the trauma and stress associated with experiences such as child maltreatment (and, in our case, associated CPS involvement and OHP) may adversely influence social, emotional, and psychological development and functioning vis-à-vis the transition to adulthood, particularly for women (Foster and Hagan 2007; Foster, Hagan, and Brooks-Gunn 2008). We note, however, that when interpreting these results it is important to recognize that we did not consider reasons for CPS involvement and that some proportion of the individuals we observe likely became CPS involved due to their own delinquency, truancy, or criminality, rather than as the result of their parent being subject to an investigation of abuse or neglect. Future research should examine this possibility.

In short, large proportions of children who become involved with CPS experience parent incarceration either before or after being screened into CPS or beginning an OHP. Furthermore, incarcerated adults are quite likely to have CPS-involved children and children in OHP. Most notably, more than a fifth of incarcerated women have a CPS-involved child and 8 to 14 percent have a child in OHP within a year of becoming incarcerated. Prior research suggests that this is not likely driven by incarceration of parents for reasons of child maltreatment perpetration (Cross et al. 2003). Furthermore, incarcerated young adults—especially women—are highly likely to have experienced CPS involvement as adolescents. CPS-involved adolescents are at relatively high risk of incarceration during young adulthood; this association is particularly strong for black individuals and men.

There are two major limitations that should be considered when interpreting our results. First, Wisconsin may not be representative of the nation as a whole. Indeed, its incarceration rate of white adults is slightly lower than the national average, whereas its incarceration rate of black adults (and its black-to-white incarceration rate ratio) is among the highest in the nation (Mauer and King 2007). Furthermore, Wisconsin has a relatively low rate of CPS involvement compared to the nation as a whole (USDHHS 2012; WI Department of Children and Families 2013a). However, here too the black-white disparity in Wisconsin is among the worst in the nation. Likewise, Milwaukee may not be representative of other large cities. It is characterized by high levels of social and economic disadvantage as well as racial disparities in both CPS and criminal justice involvement. Second, we focus only on overlap between incarceration and CPS involvement, but other aspects of the criminal justice systems, including arrest, criminal history, probation, and parole may also overlap with CPS involvement in a range of ways (Phillips and Dettlaff 2009; Phillips, Dettlaff, and Baldwin 2010).

With these caveats in mind, our analyses have several implications for research, policy, and practice. For research, first and foremost, our analyses—like most in this area—are purely descriptive. Additional research on the extent to which inter- and intragenerational links between incarceration and CPS involvement (and vice versa) may be causally linked is clearly needed. Second, research examining the reasons for which adolescents became CPS involved and whether there is variation in their likelihoods of subsequent incarceration is necessary to gain a better understanding of intragenerational links between CPS involvement and subsequent incarceration. This may be especially true with respect to women given that incarcerated young women are particularly likely to have experienced CPS involvement as adolescents. Likewise, future research is necessary to better understand the mechanisms linking parental incarceration and child CPS involvement, particularly with respect to paternal incarceration, as well as whether there may be important differences in these mechanisms by race. In addition, future research should examine reasons for differential overlap in systems involvement across racial groups in order to design effective prevention policies that address particular risk factors.

High levels of inter- and intragenerational overlap in incarceration and CPS involvement also have implications for policy and practice. To begin, child welfare policies regarding child safety, removal, and permanency should carefully consider the special needs of children with incarcerated parents, including their family situations before, during, and after the period of incarceration. They should also recognize that this is likely to be a heterogeneous group. Furthermore, parental gender and whether the parent is the child’s primary or sole caretaker should be given particular weight in designing interventions. Children whose mothers are incarcerated are particularly likely to become involved with CPS. These children may have more intensive needs than those whose fathers become incarcerated. Finally, the high degree of overlap between adolescent CPS involvement and young adult incarceration—particularly for women—suggests that such adolescents are at particularly high risk for future incarceration and may require specialized and intensive preventive efforts.

Biographies

Lawrence M. Berger is Director of the Institute for Research on Poverty and Professor of Social Work at the University of Wisconsin–Madison. He studies how socioeconomic factors (family structure and composition, economic resources) and public policies influence parental behaviors and child and family functioning and well-being.

Maria Cancian is Professor of Public Affairs and Social Work and an Affiliate and former Director of the Institute for Research on Poverty at the University of Wisconsin–Madison. Her research considers the relationship between public policies and changes in marriage, fertility, employment and family well-being.

Laura Cuesta is assistant professor of social work at Rutgers University. Her current projects examine the effects of child support policy; international approaches to child support; and the intersection of the criminal justice and child protective services systems.

Jennifer L. Noyes is a Researcher and Associate Director of Programs and Management with the Institute for Research on Poverty at the University of Wisconsin–Madison as well as an Honorary Fellow of the Robert M. La Follette School of Public Affairs. Her research focuses on the integration of services and programs designed to enhance the employment prospects and economic well-being of low-income families and children.