Assessing the Context of Permanency and Reunification in the Foster Care System. 6. Summary of Administrative Data Analysis


The comparison of the foster care reunification patterns from nine states extends our perspective about the processes involved in child discharges from the custody of state child welfare agencies.

This work is based on analysis of the population of children who have been admitted to foster care, and examines information from the administrative records of their foster care history. As such, it does not examine the underlying social contexts of the communities and families from which the children have been removed, nor does it consider the substance of reunification work -- such as the nature and quantity of services provided to the child and/or the family of origin. Instead, it provides comparable information about the outcomes of these forces, that is, the child's history in the foster care system.

There are similarities in reunification patterns across all states. This is seen most clearly in the timing of discharges from care -- exits by family reunification tend to occur earlier in foster care episodes than do other types of exits. The likelihood of reunification also tends to be similarly related to the same child and case attributes across most states.

But there is also extensive variability in reunification patterns across states. This is seen most clearly in the overall prevalence of reunification -- as the proportion of foster care exits that are achieved via family reunification varies widely across the nine states.

These apparently contradictory statements are drawn from empirical observation of reunification patterns during the mid-1990s. It is important to note that foster care reunification is not an isolated process, but rather one that is embedded in many other dynamics of state child welfare systems. Reunification levels are influenced greatly by the composition of the foster care population, and depend in part on the decisions that determine placement in foster care. For example, an emphasis on placement prevention strategies might be expected to reduce reunification levels, because a segment of "likely" cases for reunification would be diverted from placement in foster care.

The prevailing feature of the reunification process is that the likelihood of exit by reunification is highest at the beginning of a child's stay in foster care, and gradually decreases as time in care elapses. For all foster care episodes observed in these data, approximately 8 percent ended in reunification during the first month of care, about 30 percent during the first year in care, and about 40 percent during the first three years in care. This basic pattern exists for each state, as shown in the graphic of the likelihood of reunification over time in Figure C.

As a result, reunification tends to occur earlier (or for episodes of shorter duration) than other types of foster care exit. This pattern is consistent with two widely held tenets of child welfare practice: that foster care is intended to serve as a temporary intervention in a time of family crisis, and that family reunification is the preferred outcome of placement in alternative care arrangements. For a certain segment of the foster care population, relatively rapid reunification is achieved. During the time period observed, the following categories describe basic foster care (based on Table II.B.1, Panels C and D, numbers rounded):

  • 20% reunification within 6 months of placement
  • 20% reunification during months 6-36
  • 30% other exit by month 36
  • 30% still in care at month 36

Although the temporal pattern of reunification exits is similar across states, there is extensive variability in the actual levels. The percent of exits by reunification varied between 29 percent in Alabama and 67 percent in California and Wisconsin. Certain states (NM, MO, WI, CA) had high proportions of reunification levels during the first month. This may reflect entry decisions as well as exit practice. Some states (WI, NM, MI) demonstrated higher levels of continuing reunification after 6 months.

Analysis of Reunification

The likelihood of reunification is related to a number of variables describing both child and case characteristics. These were analyzed using multivariate techniques to control for the fact that the characteristic variables are themselves highly interrelated.

Child Characteristics:

  • Children more likely to become reunified are those who entered care between the ages of 1 and 11 years, white and Hispanic children, and children from outside of the largest urban counties.
  • Children less likely to be reunified include those who entered foster care as infants or teens, African-American children, and those from the largest urban counties.
  • A child's gender had no relation to reunification.

Case Characteristics:

  • Children are more likely to be reunified when they are in their first foster care placement, when their episode involves only one residence (no moves), and for those states where it is measurable, when they are placed with non-relatives.
  • Children are less likely to be reunified if they have reentered foster care, if their episode involves movement between residences, and if they are in kinship placements.


  • Under multivariate controls, the likelihood of reunification was decreasing steadily over time. The children who entered foster care in 1994 were 17 percent less likely to exit by reunification than those who entered during 1990.


  • For the most part, the direction of most of these relationships was consistent across states, although there was some variability in magnitudes. This suggests that the reunification process is similar in the different systems (Alabama is the outlier, with opposite relationships for age and region). Overall, controlling for the child and case characteristics, reunification was most likely in Wisconsin and New Mexico.

Analysis of Reentry After Reunification

Overall, 17 percent of the observed family reunifications resulted in the child re-entering foster care within one year of exit. Reentry to foster care is a significant signal that the reunification was not fully successful. A multivariate analysis was performed on the likelihood of reentry after an exit by reunification.

Higher reentry levels were observed for children who entered care during their teenage years, children who had been placed in congregate care arrangements during their stay in foster care, children whose initial spell in care was shorter in duration, children who had previously experienced a reentry to foster care, children from the non-urban counties, and for African-American children.

Lower reentry levels were observed for children who had entered as pre-teens, children who had been placed in family foster homes, children whose previous spell had been longer in duration, children who had experienced only one stay in foster care, children from large urban counties, and for Hispanic children.

The states with the highest reentry rates were Wisconsin and Illinois. While Wisconsin was also one of the states with higher levels of reunification, Illinois was one of the lower reunification states, so there is no evidence of a clear relationship between reunification rates and reentry rates at the aggregate level. Alabama and Michigan had the lowest reentry levels.

The likelihood of reentry after reunification exits was increasing moderately over time.

Adoption Analysis

Exits from foster care by completed adoption were analyzed using the same multivariate tools that were applied in the analysis of exits by reunification.

The age of the child at entry to foster care is the dominant factor in these models. Children who were placed in foster care as newborns (0-2 months of age) had a much greater likelihood of adoption than other children. (Approximately 3 times higher than children who entered at 3-12 months of age, 5 times higher than 1-2 years, etc.) This is a combination of the "attractiveness" of young children to potential adoptive parents and of the increased chance that the birth parents of newborns placed in foster care might be at higher risk of losing their parental rights than parents of older children.

Other factors related to increased likelihood of adoption include: children from non-urban counties, white children, children in first foster placements, children in family foster arrangements, and children with longer duration in alternative care.

Factors related to a lower likelihood of adoption include: children from large urban counties, African-American children, children who have reentered foster care at least once, children in congregate care placements, and children with briefer duration of time in foster care.

A substantial increase in adoptions was noted over time, with children who entered foster care in 1993 being over 20 percent more likely to exit via adoption than the children who entered foster care in 1990.

Michigan has a much higher likelihood of foster care adoption than any of the other eight states. Alabama, California, New York and Illinois have the lowest likelihoods of adoption.