Assessing the Context of Permanency and Reunification in the Foster Care System

Chapter 6.
The Role of Race in Parental Reunification

[ Main Page of Report | Contents of Report ]

Contents

  1. Introduction
  2. Prior Research
  3. Research Methods
    1. Study Sample
  4. Findings
    1. Best-Fitting Regression Model
    2. Probabilities of Reunification
  5. Discussion
  6. Research Implication

Appendix - Coding of Study Variables

References

1. Introduction

Reuniting children in foster care with their families has been a primary goal of public and private child welfare services for many years. In fact, a major objective of the landmark Adoption Assistance and Child Welfare Act of 1980 was to institute timelines to expedite children's discharge from foster care and, to the extent possible, to facilitate a timely return to their families. Moreover, the Adoption and Safe Families Act of 1997 allows states to use federal funds to provide family reunification services for the first 15 months after a child enters foster care.

However, this legislation also reduces the number of months that a child may remain in foster care without a permanency hearing from 18 to 12 months, and requires states to file for termination of the rights of parents of children who have spent 15 of the most recent 22 months in foster care. These shortened time frames present special challenges for efforts to reunite children with parents who may have serious problems, such as drug addiction and alcoholism (Hohman and Butt 2201).

Such reduced timelines may have special implications for children of color, since many child welfare studies have identified racial differentials among children in out-of-home care. They have documented the fact that black children are more likely than white children: to be placed in foster care, to be placed in kinship care, and to remain in care for longer time periods (McMurty and Lie 1992; U.S. Children's Bureau 1997; Courtney et. al. 1996).

Several studies have also found racial patterns in parental reunification (McMurty and Lie 1992; Courtney 1994; Wells and Guo 1999). For example, white children are reunified more quickly and at higher rates than black children. What are the reasons for the lower reunification rates of African American children? To what extent is it due to the greater economic deprivation and associated social problems among black families relative to white families? Or, is it a result of racial insensitivity, i.e., are black families with the same characteristics and problems as white families treated differently by the foster care system? This study addresses the role of race in reunification patterns.

[Go To Contents]

2.Prior Research

This review focuses on five reunification studies (Goerge 1990; McMurty and Lie 1992; Courtney 1994; Barth at. al. 1987; Wells and Guo 1999). These studies were selected because: (a) reunification was a major dependent variable; (b) race was an important independent variable; (c) they had adequate size samples of African Americans; and (d) they employed rigorous multivariate statistical techniques. Four studies used event history or hazard rates analyses that focused on reunification patterns for first-time cohorts of children based on administrative records (Goerge 1990; Courtney 1994; Wells and Guo 1999; McMurty and Lie 1992). The study by Goerge (1990) tracked a statewide sample of children in Illinois who entered foster care for the first time from 1977 to 1984. He employed proportional hazards analysis to describe the declining probability of reunification the longer children remained in foster care. Interestingly, his analysis did not find lower reunification rates for blacks than whites, nor that race alone had any independent or main effects on reunification rates. But he did find interactions between race and region. For example, black children in Cook County had a lower probability of reunification than non-black children in Cook County. He also found that children with non-relatives had higher reunification rates than children in kinship care.

Another reunification study was conducted by Courtney (1994) in California. He also used the proportional-hazard regression model to track the experiences of a cohort of children who entered foster care for the first time between January 1988 and May 1991. His analysis revealed that less than half of the children entering foster care for the first time would be reunified with their families within three years. Among the half reunified, about half went home within six months and about 70 percent returned within one year. Thus, he concluded the probability of reunification is greatest immediately following placement. His analysis also revealed important differences in reunification patterns between children placed with relatives and those placed with non-relatives: about half of the children placed with non-relatives during the first six months of the study were reunified, compared to only 36 percent of those placed with relatives. But the differences declined between the two groups for children who had been in care for more than six months. As a result of these deviations from the proportional-hazards framework, he decided to stratify his analysis between kin and non-kin.

Courtney found racial differentials in reunification. Whether the children were placed with kin or non-kin, African American children were reunified at about half the rate as white children--even after controlling for child disabilities, family structure and economic deprivation. Further analyses revealed that these differences may have been due in part to the lower reunification rates of black children who were infants or older youth than white children who were infants or older youth. In fact, race and age interactions revealed that--among children placed with non-kin--black children between the ages of 1-12 years old have similar reunification rates as comparable-age white children. Yet, no similar race and age interactions between blacks and whites were found among children placed with relatives.

Wells and Guo (1999) conducted a longitudinal study of children who entered the foster care system in Cuyahoga County, Ohio for the first time between January 1, 1992 and December 31, 1993. Their experiences were tracked--based on administrative records--over four years and three months from January 1, 1992 through March 31, 1996. Rates of reunification were examined 24 months after each child's first entry into foster care. They employed proportional hazards regression models to conduct event history analysis of rates of reunification and reentry. However, this study will focus on their findings related to reunification. These researchers also identified racial differentials in reunification: black children had lower rates of reunification than white children. They also found an interaction between race and age of entry. African American infants were reunified more slowly than non-African American infants. However, these differentials declined as the children got older, such that, by age 13, there were no differences in reunification rates between the racial groups. Interestingly, the researchers did not find different reunification rates between children placed with kin and non-kin.

McMurty and Lie (1992) conducted an events history analysis of the exit patterns of white and minority children who had been placed in foster care in Maricopa County, Arizona from 1979 through 1984, and whose cases were closed from 1979 through 1986. The authors found: (1) African American children were half as likely to be reunified as white children; and (2) children with disabilities were less likely to be reunified than children without disabilities. Barth and colleagues (1987) conducted a stepwise discriminant analysis of the discharge status of children who had been placed in foster care for physical abuse in San Mateo County, California. Among the study sample of 101 children who had been discharged between 1980 and 1984, 80 were reunified and 21 had other permanent out-of-home placements (including adoption, placement with kin, guardianship, and emancipation). The study sought to determine which background characteristics or type of services more strongly predicted reunification or other permanency placements. The authors concluded that black children were less likely to be reunified than white children. They also found the intensity of direct services (i.e, the total hours of phone or face-to-face contact between social workers and clients in the home, the worker's office, or in the courtroom) to be a stronger predictor of reunification than the intensity of in-home services (i.e., the total hours of all services received in the home and rendered by all providers.)

What conclusions can be drawn from these five studies? Four studies found lower rates of reunification for black than white children (Barth et. al. 1987; McMurty and Lie 1992; Courtney 1994; Wells and Guo 1999). Two studies found interactions between race and age of entry: wide racial differentials in reunification rates at younger ages, which declined as the children grew older (Courtney 1994; Wells and Guo 1999). Two studies found lower reunification rates among children who were placed with kin than among those placed with non-relatives (Courtney 1994; Goerge 1990).

[Go To Contents]

3. Research Methods

This study seeks to answer the question, "Is race a strong predictor of reunification when combined with other important child, family or case history characteristics?" Related questions are: (a) "What other child, family or case history characteristics are also strong predictors of reunification?" and (b) "Are the main effects of race reduced when controlling for other important predictors?"

The data for this study are from the 1994 National Study of Protective, Preventive and Reunification Services Delivered to Children and Their Families (NSPPRS), which will be referred to as the "National Study." An objective of the National Study, which was conducted by Westat, was to document the number and characteristics of children and families who received various in-home or out-of-home child welfare services between March 1, 1993 and March 1, 1994 based on a national sample of 2109 children. This analysis will focus on the subsample of 1034 children who had foster care experiences (i.e., were placed in foster care) during the study period, had left foster care or who were still in foster care by February 28, 1994.

The National Study has a longitudinal component, since it obtains outcome data on the segment of children who were discharged during the six-month period after February 28, 1994. This analysis, however, will rely on the cross-sectional aspects of the study. Thus, the NSPPRS is essentially a point-in-time study that suffers from the limitations of many cross-sectional studies: it over-represents cases that remain in care for longer periods of time. On the other hand, a unique advantage of the National Study is that, since it is based on interviews with caseworkers who provided information for specific foster children based on their case records, the survey was able to include variables that are not often found in most administrative records. Because of the small sample size of Hispanics in the survey, this analysis is restricted to comparisons between blacks and whites.

Logistic regression models are the primary method of analysis in this study. Most of the study variables were also used in other reunification studies. The independent variables fall into three groups: children's characteristics, family characteristics and case history characteristics. Children's characteristics include: race, gender, age at entry, and disability status. Family characteristics include: social class index (combining parent's education and employment status), whether the parent lacks job skills, has substance abuse problems, or is a single mother. Case history characteristics include: reasons for placement (abuse allegations or neglect allegations), whether the child was placed with kin or nonkin, and whether the parental caretaker was provided various kinds of services. Caseworkers were asked whether the caretakers received any of the following 23 services: parent training (at home or in classes), household management, homemaker services, day care, respite care, emergency financial aid, family planning, legal services, schooling, employment training, health care, psychological assessments, out-patient or inpatient mental health treatment, case management or counseling, out-patient or inpatient substance abuse treatment, self-help groups, housing, housing payment, temporary shelter and transportation.

The dependent variable in this analysis is parental reunification (i.e., the proportion of children who were reunified vs. those who were discharged for other reasons or are still in care.) The specific codes for the independent and dependent variables are described in detail in the appendix.

3.1 Study Sample

Selected characteristics of the sample of children in the reunification study are presented in Table 1. These characteristics were reported by caseworkers to the Westat interviewers based on a careful examination of the children's case records. Of the 1,034 foster children in the study, 21 percent were reunified, seven percent had other discharges and 72 percent were still in care. Half of them (52%) were white, 38 percent were black and one-tenth were Hispanic. Six out of ten had a disability. Thirty-seven percent entered foster care under the age of five, 17 percent entered between the ages of 5-8 years old, and 46 percent entered at nine years or older. How long were they in foster care? Twenty-eight percent were in care less than one year, 38 percent were in foster care between 1 and 2 years, while 34 percent were in foster care three or more years. One-fifth of the children were placed with relatives.

What are the characteristics of the children's caretakers? Only three out of ten were single mothers. Over half (54%) had less than a high school education, 64 percent were not employed and half received AFDC. Thirty-one percent of the caretakers lacked job skills, 41 percent had substance abuse problems, and 44 percent did not receive any of the 23 services listed. Three out of ten (30%) caretakers had allegations of abuse, while one-fourth (26%) had allegations of neglect. Most data on foster children reveal that the number placed for neglect usually far exceeds those placed for abuse. Yet, the data in Table 1 provide comparable numbers of abuse and neglect allegations. There are two reasons for these results. First, these study variables refer to the subset of ONLY abuse and ONLY neglect allegations--excluding cases in which both abuse and neglect allegations may have been present. Second, neglect allegations were restricted to the three types that almost all states include in their definitions: physical, emotional and medical. Allegations of abandonment, lack of supervision, or failure to thrive were excluded, since many states do not include them in their classifications of neglect.

Table 1.
Characteristics Of Children In Reunification Study
Discharge Status Caretaker Education

Reunified

21%

Less than HS education

54%

Other discharge

7%

HS or higher education

46%

Still in care

72%

Percent

100%

Percent

100%

Total N

555

Total N

1,034    
   

Employment

Race of Child

Employed

36%

White

52%

Not employed

64%

Black

38%

Percent

100%

Hispanic

10%

Total N

743

Percent

100%    

Total N

989

AFDC

   

Receives AFDC

49%

Child Disability

No AFDC

51%

Disabled

60%

Percent

100%

Not disabled

40%

Total N

717

Percent

100%    

Total N

1,017

Job Skills

   

Has skills

69%

Age Entered Foster Care

No skills

31%

Under 1 year

17%

Percent

100%

1-4 years

20%

Total N

793

5-8 years

17%    

9-12 years

19%

Drug Problem

13 or more years

27%

Has problems

41%

Percent

100%

No problem

59%

Total N

946

Percent

100%
   

Total N

792

Time in Foster Care

   

Under 1 year

28%

Caretaker Services

1-2 years

38%

1 or more services

56%

3-4 years

18%

No services

44%

5-9 years

13%

Percent

100%

10 or more years

3%

Total N

1,034

Percent

100%    

Total N

1,033

Abuse Allegations

   

Only abuse

30%

Kin Care

No abuse

70%

With Kin

21%

Percent

100%

Not with Kin

79%

Total N

763

Percent

100%    

Total N

976

Neglect Allegations

   

Only neglect

26%

Single Mom

No neglect

74%

Single

30%

Percent

100%

Not single

70%

Total N

792

Percent

100%  

Total N

1,034  
(Total N = unweighted sample sizes)

[Go To Contents]

4. Findings

What is the relationship between race and reunification? The data in Table 2 reveal that 34 percent of whites are reunified, compared to only nine percent of blacks. Thus, white children are four times more likely than black children to be reunified.

Table 2.
Reunification By Race
 

White

Black

Total

Reunified

34%

9%

21%

Not Reunified 1

66%

91%

79%

Total Percent

100%

100%

100%

(Unweighted N)

(519) (371) (890)

(Weighted N)

(277,858)

(303,840)

(581,698)

p=.00*
1 Not reunified includes children discharged for reasons other than reunification and those children who are still in care.

In order to identify the independent variables that were strong predictors of reunification, logistic regressions were run separately with single factors. The rankings of 15 variables by size of R2 are presented in Table 3. Eleven variables have significant bivariate relationships with reunification: race, services provided to caretakers, age of entry, parental job skills, socioeconomic status (SES), parental employment status, parental substance abuse problem, child placed because of neglect, child placed because of abuse and parental education. These findings reveal that reunification rates were higher among caretakers: who received services, who had job skills, who had no substance abuse problem, who had completed high school, and who were currently working. However, white and older children had higher reunification rates than black and younger children. Yet, it should be noted that abuse and neglect had opposite relationships with reunification. Children placed for abuse were more likely to be reunified, while children placed for neglect were less likely to be reunified. On the other hand, four variables were not significantly related to reunification: child's disability, child's gender, if the parent was a single mother, and whether the child's parental household received AFDC. Thus, it was decided to drop these four variables from any further analysis.

Table 3.
Single-Factor Predictors of Reunification

Predictors

Estimates P-level (100)R 2 F-values

Race (black=1)

-1.61

.00*

9.2%

22.98

No Caretaker Services

-1.36 .00* 6.4% 19.97

Age at Entry

.11 .00* 5.5% 24.81

Lack of Job Skills

-1.35 .00* 5.2% 23.22

Low SES

-1.27

.00*

4.9% 16.74

Not Employed

-.99

.00*

3.5% 11.60

Substance Abusers

-.91

.00*

3.2% 11.32

Neglect Allegations

-.98

.01*

2.7%

9.95

Abuse Allegations

.90

.00*

2.7%

9.95

Low Education

-.63

.03*

1.6%

5.01

Kin Caretaker

-.68 .03* 1.2% 5.14

Child Disability

-.15 .64 0.1% .23

Child Gender

-.09 .65 0.0% .21

Single Mom

.14

.48

0.0% .52

AFDC Recipient

.09 .73 0.0% .13

*Statistically significant relationships at p=< .05

To examine the possible presence of multicollinearity, a correlation matrix was run among nine variables that were significantly related to reunification. (Parental education and employment status were not included in the matrix, since they comprised the SES index.) The strongest correlations were found between substance abuse problems and SES (-.61), between abuse and neglect allegations(.53) and race and placement with kin (-.53). To minimize collinearity, abuse was excluded from the initial analyses of the best-fitting models for predicting reunification and neglect was retained. Despite the high correlations between kin placement and race, and SES and substance abuse problem, it was decided to retain them to assess their effects on the best-fitting models, since they were found to be important predictors of reunification in prior studies (see Table 4).

Table 4.
Correlations of Estimates
  No Services Entry Age Lack Skills Sub Abuser Allegations With Kin Low  SES
Neglect Abuse  

Race

-.12 .38 .18 -.36 .19 .39 -.53 .01

No Services

  -.15 .21 .05 -.16 -.16 .05 .21

Age of Entry

    .11 .15 -.11 -.04 -.25 .08

Lack Job Skills

      .30 -.06 .23 -.14 -.28

Substance Abusers

        .04 .15 -.08 -.61

Neglect Allegations

          .53 .01 -.37

Abuse Allegations

            -.34 -.27

Kin Caretaker

              -.06

4.1 Best-Fitting Regression Model

To examine interaction effects among the stronger predictors, many regression models (such as race by entry age, race by services, race by job skills, race by substance abuse problem, race by kinship placement, substance abuse by services, etc.) were run that included various combinations of those variables. However, none of the interaction terms maintained significance in combination with the remaining variables. In some cases, the addition of interaction terms reduced the size of the R2 . Thus, it was decided to focus on regression models that excluded interactions.

In order to determine the best-fitting logistic regression model, the top-ranking eight variables were entered in a stepwise sequence from stronger to weaker predictors. The resulting eight models are presented in Table 5. The top-ranking five variables (race, age at entry, job skills, caretaker services, substance abuse problem) are entered sequentially in Models 1-5. As each variable is entered, the others continue to be significant predictors of reunification. In Model 5, their combined effects accounted for a 20.5 percent variance (R2 ) in reunification with an F-value of 15.45.

However, when neglect was added in Model 6, it was not significantly related to reunification, while four of the five other predictors continued to be significant predictors. Similarly, when kin caretaker is added in Model 7, it no longer was significantly related to reunification, while all five of the other variables remained significant predictors. It is interesting that the strong correlation between race and parental reunification was not reduced when kinship placement was added to the model. Finally, when SES was added in Model 8, it also was no longer significantly related to reunification, while four of the five other variables continue to be significant predictors. Nor does the addition of SES reduce the influence of race. on reunification. Since time in foster care, abuse, and caretaker's education were found to be important predictors in past studies, additional regression models were run separately for each of these three variables. Moreover, a broader definition of neglect (including physical, emotional, medical, abandonment, lack of supervision and failure to thrive) was also included in a separate regression model. However, none of these additional variables continued to be significantly related to reunification when combined with the other five predictors. Consequently, Model 5 was considered to be the best-fitting model for predicting parental reunification, since it was the only model in which all of the predictors were able to maintain a significant relationship with reunification in combination with one another.

Table 5.
Steps To Obtain Best-Fitting Model For Predicting Reunification

Predictors

Model 1

Model 2

Model 3

Model 4

Intercept

-69 (.00)*

-.32 (.07)

-1.05 (.00)

-.55 (.06)

Race (black=1)

-1.61 (.00)*

-1.54 (.00)*

-1.50 (.00)*

-1.57 (.00)*

No Services

-----

-1.06 (.00)*

-1.07 (.00)*

-.97 (.01)*

Age at Entry

-----

-----

.09 (.00)*

.08 (.00)*

No Job Skills

-----

-----

-----

-1.40 (.00)*

(100)R 2 =

9.2%

12.7%

16.8%

20.1%

F-Value =

22.8 (.00)*

28.12 (.00)*

21.28 (.00)*

15.78 (.00)*

*Statistically significant relationships at p=< .05

Table 5.
Steps To Obtain Best-Fitting Model For Predicting Reunification
(Continued)

Predictors

Model 5

Model 6

Model 7

Model 8

Intercept

-.33 (.33) -.18 (.69) -.22 (.64) -.33 (.53)

Race (black = 1)

-1.46 (.00)* -1.47 (.00)* -1.49 (.00)* -.93 (.05)*

No Services

-1.04 (.00)* -1.04 (.00) -1.08 (.01) -.99 (.01)

Age at Entry

.08 (.01)* .08 (.02)* .09 (.02)* .11 (.01)*

No Job Skills

-1.36 (.00)* -1.36 (.00)* -1.36 (.00)* -1.07 (.01)

Substance Abuse Problem

-.48 (.04)* -.46 (.10) -.46 (.00) -.04 (.92)

Neglect Allegations

----- -.33 (.39) -.27 (.47) .04 (.93)

Kin Caretaker

----- ----- .10 (.79) .18 (.66)

Low SES

----- ----- ----- -.78 (.09)

(100)R2 =

20.5% 22.4% 22.1% 19.8%

F-Value =

15.45 (.00)* 18.77 (.00)* 14.77 (.00)* 7.08 (.00)*
*Statistically significant relationships at p=< .05

4.2 Probabilities of Reunification

Based on the parameters of the best-fitting model, the probabilities of reunification are derived for various subgroups of children who entered foster care at seven years old, which was the average age. Using the methodology employed in the reanalysis of NIS-2 data (Sedlak 1993), probabilities of reunification are derived for eight scenarios that involve various combinations of caretaker services, job skills and substance abuse problems (see Table 6). The highest probabilities of reunification are represented in scenario eight: children whose parental caretakers have job skills, received services and do not have any substance abuse problems. For this scenario, the probability of reunification is 56 percent among white children, compared to only 23 percent among black children.

Table 6.
Probabilities Of Reunification For Children Entering Foster Care At Age 7 By Scenario And Race
 

Job Skills

Caretaker Services Substance Abuse Problem Probability of Reunification
(100* Probability)
Scenario   Black White
1 No No Yes 2% 7%
2 No No No 3% 10%
3 No Yes Yes 4% 17%
4 Yes No Yes 6% 22%
5 No Yes No 7% 24%
6 Yes No No 9% 31%
7 Yes Yes Yes 15% 44%
8 Yes Yes No 23% 56%

Scenario six represents intermediate probabilities: children whose parental caretakers have job skills, have no substance abuse problem, but received no services. For this scenario, whites have a probability of reunification that is three times higher than blacks (31% vs. 9%). Other intermediate probabilities appear in scenario four: children whose parental caretakers have job skills, but have substance abuse problems, and received no services. In this scenario, whites have a probability of reunification that is about four times higher than blacks (22% vs. 6%). Yet, scenario one represents the lowest probabilities of reunification: children whose parental caretakers have no job skills, received no services, and have substance abuse problems. In this scenario, the probability of reunification among white children is 7 percent, compared to only 2 percent among black children.

To depict visually the various probabilities of reunification by race, a bar graph (Fig. 1) was created that compares four scenarios: one, four, six and eight. It reveals that the probability of reunification increases as the parental caretakers' characteristics are more advantaged. Yet, it also reveals that even when parental caretakers of black children have the same desirable (or undesirable) characteristics as white caretakers, white children are still three or more times likely to be reunified than black children. Moreover, the probability of reunification for blacks with the more desirable characteristics in scenario eight (23%) is about the same as it is for whites in scenario four (22%) in which caretakers have substance abuse problems, received no services, but have job skills. Clearly, race continues to be a strong predictor of reunification in each of the scenarios.

Figure 1
Probability of Parental Reunification for Children Who Entered Foster Care At Age 7 by Race

Figure 1. Probability of Parental Reunification for Children Who Entered Foster Care At Age 7 by Race.

[Go To Contents]

5. Discussion

Although this study is based on cross-sectional data, it is interesting to note that many of its findings are consistent with those from several longitudinal studies. This is especially evident when some of the bivariate relationships of important variables with reunification are compared with those of prior research. First, this analysis found race to be significantly related to reunification: white children were more likely to be reunified than black children. This finding agrees with those in four reunification studies: Barth (1987), Courtney (1994), Wells and Guo (1999) and McMurty and Lie (1992). While Goerge (1990) found no direct racial differences, he obtained interactions between race and region. For example, blacks in Cook County had lower reunification probabilities that whites in Cook County.

Second, this study found kinship placement to be inversely related to reunification: children who were placed with kin were less likely to be reunified than children placed with non-relatives. These findings agree with those of two studies: Goerge (1990), and Courtney (1994). However, Wells and Guo (1999) did not find kinship placement to be related to reunification. Moreover, this study found that kinship placement did not continue to be significantly related to reunification, when combined with race and the other predictors in the regression models. Thus, it was concluded that the higher kinship placements of black children do not explain their lower reunification rates relative to white children.

Third, this study found age of entry to be significantly related to reunification: rates of reunification rose directly with increases in the age that the children entered foster care. Several studies also found lower reunification rates among infants and young children than older children (Goerge 1990; Courtney 1994). Moreover, Courtney (1994) found large racial differences in reunification between blacks and whites at young ages, but declining racial differentials as the age of the children increased. While Wells and Guo (1999) found no significant relationship between age of entry and reunification, they found a positive relationship between the interaction term "African American and age of entry" and reunification.

Fourth, although this study found the reasons for placement to be significantly related to reunification, there were wide differences regarding the specific reasons in prior research. For example, Goerge (1990) found neglected children to have higher reunification rates than abused children. But Wells and Guo (1999) found that children placed for abuse had higher rates of reunification than those placed for neglect or dependency. While Courtney (1994) found--among children placed with non-kin--that children placed for sexual abuse had higher reunification rates than those placed for neglect, he found no significant relationships between any of the reasons for placement and reunification among children placed with kin. And, while this study found that abuse and neglect were both significantly correlated with reunification, they were related in the opposite directions. Children placed for abuse were more likely to be reunified, while children placed for neglect were less likely. However, neither reason for placement continued to be significantly related to reunification -- when combined with other key predictors in the regression models.

On the other hand, unlike the studies by McMurty and Lie (1992), Courtney (1994) and Wells and Guo (1999), this analysis did not find child disability or health problems to be significantly related to reunification. Those studies found children with health problems to be reunified more slowly than children without any health problems. Nor did this study find, as did Courtney (1994), that AFDC status was significantly related to reunification. Yet, this analysis found two other measures of socioeconomic status -- parental education and employment status -- to be positively related to reunification. Children whose parental caretakers had completed high school and were currently employed were more likely to be reunified than children whose parental caretakers were high school dropouts or were not employed.

The questions raised at the beginning of this study will now be addressed. First, "Is race a strong predictor of reunification when combined with other important child, family or case history characteristics?" This analysis revealed that race is a strong predictor of reunification when combined with other important child, family or case history characteristics. A second question was: "What other child, family or case history characteristics are also strong predictors of reunification?" This study reveals four important predictors of reunification in addition to race: age of entry, caretaker job skills, caretaker substance abuse problems and caretaker services.

A third question was: "Are the main effects of race reduced when controlling for other important predictors?" This study found that controlling for the other key predictors does not reduce the independent effects of race. In sum, it was concluded that race continues to play a major role in the reunification of children in addition to other child, family and case history characteristics.

[Go To Contents]

6. Research Implications

This analysis suggests a number of future directions in research on race and reunification. First, as Courtney and colleagues (1996) conclude in their comprehensive review of the literature, it is important to acknowledge the role of race and ethnicity in future child welfare research:

We encountered many studies in which these factors were not mentioned as variables, although the sample size and location of the study would have lent themselves to such analysis. The failure or unwillingness to at least acknowledge the relationships among race, child welfare services, and child welfare outcomes may only serve to invite uninformed speculation about the reasons for these relationships. Whenever methodologically possible, child welfare researchers should include race as an explanatory factor in research designs and consider their theoretical justification for doing so (i.e., why does the researcher think that race might play a role?) (Courtney 1996).

Second, there is a need for reunification studies that not only include adequate sample sizes of African Americans, but of Hispanics, Native Americans, and Asian and Pacific Islanders as well. While many studies have often obtained similar results for Caucasians and Hispanics, little is known about reunification rates for Native Americans or Asians because they are most often excluded due to their small sample sizes. Third, there is a need for more longitudinal studies that are based on both administrative records and surveys that track cohorts of children over time to better understand the dynamics of reunification patterns among racial and ethnic groups.

Fourth, additional research is needed on the type of services that are most appropriate for enhancing reunification among parental caretakers of different racial and ethnic groups (Maluccio, Fein, and Davis 1994). This analysis found lower reunification rates among families that did not receive any services. Reviews of child welfare research revealed that families of color are less likely to receive services than Caucasian families (Courtney 1996). Further research is needed to determine, "Which mix of specific services are most effective for reunification with which types of parental caretakers among various racial and ethnic groups?" Fifth, more research is needed that continues to include common measures from prior studies in order to more carefully examine their separate and combined effects on reunification. Such an approach might resolve contradictory findings related to the importance of race, ethnicity, kinship placement, SES, child disability, age of entry, and reasons for placement as predictors of reunification.

[Go To Contents]

Appendix: Coding of Study Variables

Independent Variables

Child Characteristics

Race: black=1; white=0.
Gender: male=1; female=0
Age of Entry: Continuous variable (average age was seven years old)
Child Disability Status: with disability=1; no disability=0

Family Characteristics

Parent Job Skills: lack job skills=1; has job skills=0
Parent Substance Abuse Problem: has problem=1; has no problem=0
Parent Single Mother: is a single mom=1; not a single mom=0
Parent Social Class: finished HS and employed=0; all other combinations=1

The social class index comprises only parental education and employment status.
The variable "AFDC recipient" was excluded from the index, because it was not correlated with the other two measures of socioeconomic status.

Case History Characteristics

Reasons for Placement: only abuse allegation=1; no abuse allegation=0.
                                     only neglect allegation =1; no neglect allegation=0.

The variable "only abuse allegations" applies only to allegations that include: sexual abuse, physical abuse and emotional abuse; it excludes any neglect allegations. The variable "only neglect allegations " applies only to allegations that include: physical neglect, emotional neglect; and medical neglect, it excludes any abuse allegations.

Type of Placement: with kin=1; not with kin=0

Parental Caretaker Services: no services received=1; some services received=0.

The 23 services included: parent training at home, parent training in classes, household management, homemaker services, day care, respite care, emergency financial aid, family planning, legal services, schooling, employment training, health care, psychological assessments, outpatient mental treatment: inpatient mental health treatment, case management or counseling, outpatient substance abuse treatment, in-patient substance abuse treatment, self-help groups, housing, housing payment, temporary shelter and transportation. The number of services provided to caretakers could range from 0-23.

Dependent Variables

Parental Reunification: reunified=1; other discharges and still in care=0

[Go To Contents]

References

Barth, Richard, et al. (1987) Contributors to Reunification or Permanent Out-of-Home Care for Physically Abused Children. Journal of Social Service Research, 9(2/3): 31-45

Courtney, Mark E. (1994) Factors Associated with the Reunification of Foster Children with Their Families. Social Service Review, 68 (1): 81-108.

Courtney, Mark E. (1995) Reentry to Foster Care of Children Returned to Their Families. Social Service Review, 69 (2): 226-241.

Courtney, Mark E., et al. (1996) Race and Child Welfare Services: Past Research and Future Directions. Child Welfare, 75 (2): 99-137.

Frame, Laura, Jill Duerr Berrick and Melissa Lim Brodowski (2000) Understanding Reentry of Out-of Home for Reunified Infants. Child Welfare, 79 (4): 339-369.

Goerge, Robert M. (1990) The Reunification Process in Substitute Care. Social Service Review, 64 (3): 422-457

Hohman, Melinda M. and Rick L. Butt (2001) How Soon is Too Soon? Addiction Recovery and Family Reunification. Child Welfare, 80 (1): 53-67.

Maluccio, Anthony N., Edith Fein, and Inger Davis (1994) Family Reunification: Research Findings, Issues, and Directions. Child Welfare, 73 (5): 489-504.

Maas, Henry and Richard Engler, Jr. (1959) Children in Need of Parents. New York: Columbia University Press.

McMurty, S. and G. Lie (1992) Differential Exit Rate of Minority Children in Foster Care. Social Work Research and Abstracts, 28 (1): 42-48.

Sedlak, Andrea (1993) Study of High Risk Child Abuse and Neglect Groups, NIS-2 Reanalysis: Report to Congress. Appendix C.

U. S. Children's Bureau (1997) National Study of Protective, Preventive and Reunification Services Delivered to Children and Their Families. Washington, D. C: U. S. Government Printing Office.

Wells, Kathleen and Shenyang Guo (1999) Reunification and Reentry of Foster Children. Children and Youth Services Review, 21 (4): 273-294


Where to?

Top of Page | Contents

Main Page of Report | Contents of Report

Home Pages:
Human Services Policy (HSP)
Assistant Secretary for Planning and Evaluation (ASPE)
U.S. Department of Health and Human Services (HHS)