Assessing the Context of Permanency and Reunification in the Foster Care System. 4.2 Multivariate Models for Reentry After Reunification

12/01/2001

Tables IIE.1 and IIE.2 present the results of logistic regression models predicting the likelihood of reentry within 12 months of a reunification from foster care. As with the logistic regression models used in the analysis of reunification, these results are computed with multivariate controls. Even so, they tend to reproduce the findings observed from the bivariate proportions described in the previous section.

Table IIE.1 
Logistic Regression Model of Reentry to Foster Care within One Year of a Prior Exit by Reunification.

Predictor Variable

Category Standardized parameter Odds Ratio

Age at Entry

< 3 months 0.021 1.15
3 -11 months 0.015 1.11
1 - 2 years 0.000 1.00
3 - 5 years -0.024 0.89
6 - 8 years -0.043 0.79
9 -11 years * - - 1.00
12-14 years 0.035 1.18
15-17 years -0.114 0.53

Gender

Male 0.003 1.01
Female * - - 1.00

Race/Ethnic

African American 0.039 1.17
Hispanic -0.032 0.86
White, Other * - - 1.00

Region

Primary Urban County -0.053 0.82
Remainder of State * - - 1.00

Sequence

First Spell in Care * - - 1.00
Reentry Spell 0.087 1.68

Care Type

Congregate Care 0.045 1.24
Foster Care, Kin Care * - - 1.00

Stability

One Placement in Spell * - - 1.00
2+ placements 0.050 1.21

Duration

Log(months in spell) -0.111 0.84

State

Alabama -0.014 0.81
California 0.031 1.12
Illinois 0.074 1.61
Maryland 0.014 1.17
Michigan * - - 1.00
Missouri 0.020 1.17
New Mexico 0.022 1.34
New York 0.011 1.05
Wisconsin 0.080 1.69

Year of Entry

1990 -0.022 0.91
1991 -0.017 0.93
1992 * - - 1.00
1993 0.002 1.01
  Intercept -1.516  
exp(intercept) 0.220  
Concordance 0.61  
N 138,503  
N reentries 23,131  
Proportion Reentry 0.17  

Note: Predictors noted with "*" are not contained in model, but are the "excluded" category for their variable. Predetermined odds ratios of 1.00 are assigned to these categories. Odds ratios for other categories of the same variable express effects in relation to that of the excluded category.


The pooled nine-state results in the model in Tables IIE.1 suggests the following:

  • Reentry is most likely for children who entered care at ages under one year or between the ages of 12 and 14 years.
  • African American children are more likely to reenter after reunification, Hispanics are less likely to reenter, with the likelihood for whites falling in between.
  • Children from primary urban places have lower reentry rates than children from the other regions of their states.
  • Reentry spells are substantially more likely than first spells to result in subsequent reentry. Some children seem to cycle in and out of care.
  • Children in congregate care are more likely to reenter than are children in foster care.
  • Children who experienced more than one placement during the previous spell are more likely to reenter than children who had a single placement.
  • Reentry is negatively related to the duration of the prior spell. Reunification after a short stay in care is more likely to lead to reentry than reunification after a longer spell in care.
  • The rates of reentry increased moderately between the 1990 and 1993 cohorts.
  • Controlling for the other variables in the model, the likelihood of reentry is highest in Wisconsin and Illinois, and relatively high in New Mexico. The likelihood of reentry is lowest in Alabama and Michigan.

The individual state models in Table IIE.2 again tend to resemble the pooled model. Some of the differences observed are:

Table IIE.2 
Logistic Regression Model of Reentry to Substitute Care within 12 months of Reunification Exit. All Child Entrants Between 1990-93 That Exited by Reunification Within 3 years. By State.
  AL CA IL MD MI MO NM NY WI 9-State

Age at Entry

< 3 months 0.79 1.30 1.08 0.91 1.19 1.08 0.61 0.98 0.76 1.12
3 -11 months 0.74 1.25 1.02 1.07 1.13 1.09 1.34 0.93 1.02 1.10
1 - 2 years 1.07 1.10 1.00 0.84 1.04 1.03 1.27 0.84 0.83 1.00
3 - 5 years 1.00 1.00 0.94 0.77 0.86 0.78 1.36 0.76 0.72 0.89
6 - 8 years 0.62 0.87 0.85 0.70 0.80 0.72 0.98 0.70 0.69 0.80
9 -11 years * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
12-14 years 0.94 1.10 1.52 1.21 0.97 1.28 0.98 1.10 1.34 1.18
15-17 years 0.36 0.56 0.90 0.65 0.47 0.60 0.54 0.34 0.60 0.56

Gender

Male 0.96 1.00 1.07 1.05 1.05 1.00 1.06 0.99 1.02 1.02
Female * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Race/Ethnic

                   
African American 1.37 1.30 1.23 0.97 1.29 1.30 0.85 1.04 1.03 1.17
Hispanic - - 0.88 0.71 0.62 0.99 0.97 0.95 0.95 0.96 0.85
White, Other * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Region

Primary Urban County 0.86 0.88 0.95 0.97 0.74 0.98 1.45 0.76 0.59 0.82
Remainder of State 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Sequence *

First Spell in Care 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Reentry Spell 1.19 1.73 1.58 1.56 1.63 1.59 1.69 1.40 1.76 1.67

Care Type

                   
Kinship Care - - 0.71 0.86 0.64 - - 0.56 0.66 0.59 - - 0.77
Congregate Care 1.29 1.39 1.44 1.23 1.22 1.22 1.05 1.07 0.85 1.16
Foster Care * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00

Stability

One Placement in Spell * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
2+ placements 0.76 1.07 1.13 1.27 1.38 1.24 1.38 1.23 1.29 1.20

Duration

Log(months in spell) 0.74 0.99 0.85 0.82 0.90 0.94 0.92 0.71 0.72 0.85

Year of Entry

                   
1990 0.86 0.83 0.96 0.87 1.08 1.12 0.69 1.00 0.93 0.91
1991 0.88 0.87 1.00 0.97 1.11 1.07 0.78 1.02 0.81 0.93
1992 * 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00 1.00
1993 1.14 0.99 0.95 1.02 1.29 1.09 0.93 1.01 0.95 1.01

Intercept

-1.32 -1.60 -1.20 -1.24 -1.84 -1.68 -1.48 -0.90 -0.58 -1.31
  0.27 0.20 0.30 0.29 0.16 0.19 0.23 0.41 0.56 0.27

Concordance

0.66 0.60 0.60 0.60 0.59 0.60 0.63 0.65 0.66 0.60

N

1,962 56,518 12,129 3,745 10,838 8,080 2,666 31,047 11,518 138,503

N reentries

281 8,733 2,706 675 1,590 1,495 520 4,628 2,503 23,131

Prop. reenter in 12 mos.

0.14 0.15 0.22 0.18 0.15 0.19 0.20 0.15 0.22 0.17

Note: Predictors noted with "*" are not contained in model, but are the "excluded" category for their variable. Predetermined odds ratios of 1.00 are assigned to these categories. Odds ratios for other categories of the same variable express effects in relation to that of the excluded category.


Age: In Alabama, New York, and Wisconsin, the youngest children (those who entered their prior spell under one year of age) do not have higher reentry rates. In New Mexico, the highest reentry rates are for children who entered their first spell between 3-5 years of age-while in most states this was one of the lower reentry age groups. Also, Alabama, Michigan, and New Mexico do not show an increase in reentry for children ages 12-14.

Race and Ethnicity: Reentry levels are lower for Hispanics in all states. Reentry levels are much higher for African Americans than whites in Alabama, California, Illinois, Michigan, and Missouri. In Maryland, New York, and Wisconsin, the reentry levels for African Americans and whites are very similar. In New Mexico, African American reentry rates are lower than those for Hispanics or whites.

Region: Reentry levels are lower in the primary urban county for all states except New Mexico, where they are substantially higher in Bernalillo County. Reentry rates are especially low in the primary urban places in Michigan (Wayne County.), New York (New York City) and Wisconsin (Milwaukee County.).

Duration: The inverse relation between spell duration and the likelihood of reentry is strongest in Alabama, New York, and Wisconsin. Note that these include one low-reunification state (Alabama), one medium-reunification state (New York), and one high-reunification state (Wisconsin).

Time Trends: As was observed in the reunification models, while the pooled nine-state data shows a clear trend across entry cohorts (for reentry it is increasing), the individual state trends are less clear or mixed. Alabama, California, and Maryland are the only states where the odds ratios present a clear picture of increasing likelihood of reentry.