Research Update: Understanding complexity in the foster care system
In child welfare research, it is common to represent foster care outcomes using what statisticians call a multivariate linear model: where each child’s unique characteristics and circumstances determine their likelihood of entering and leaving care. In a basic linear model, these individual entries and exits take place independently of each other. However, research being carried out at the Center for State Child Welfare Data at Chapin Hall suggests that such approaches may fall short of capturing the complex dynamics that shape what happens to children, when, and why.
DECISION MAKING SHAPED BY RESOURCE CONSTRAINTS
In the foster care system, caseworkers (and others) make decisions that affect how children enter, move through, and exit care. Although these decisions – to admit, move, or discharge a child – are made on a daily basis, over time (weeks, months, and years) they determine whether the foster care caseload grows, shrinks, or remains constant. From a policy and practice perspective, it is important to consider whether these decisions are shaped purely by the child’s individual needs and circumstances or whether they are also influenced by the system itself. For example, whether a child is admitted to care is likely to be affected by whether a bed is available at the time the decision is made. The decision might also be affected by how much time the caseworker has to make the decision. This means that even if a child in one foster care system has exactly the same characteristics as a child in another foster care system, their trajectories through care may very well differ, due to resource constraints in their respective foster care systems.
TWO-WAY CAUSALITY BETWEEN ENTRIES AND EXITS
Although most observers of foster care and foster care systems understand that resource constraints influence what happens to children, there has been relatively little empirical research that explores these themes at the system level. In part, this is a conceptual problem. Are resource constraints observable in the micro decisions made by caseworkers? Or, is it possible to observe the effects of resource constraints in the dynamic interplay of children entering and leaving foster care? To put the question in simple terms, are entries and exits tied together through time in ways that are consistent with an underlying causal dynamic: do entries cause exits and do exits cause entries?
Using entry/exit time series data from the Foster Care Data Archive, our study examined the number of children entering and leaving the placement system each week for a period of 15 years. In this study, we hypothesized that beds were the relevant resource constraint. We found evidence of a causal relationship at the aggregate level. When children are placed into care, the resource constraints (i.e., the supply of beds) create the conditions that favor children leaving care. Similarly, we found that when children leave care, entries into care followed. Interestingly, the strength of the causal signal was more significant in the congregate care system then in either the foster or kinship care subsystems. We believe this is because of the fixed cost nature of congregate care – vacancies in fixed cost systems have different implications when compared to systems with a different fixed cost structure.
A NON-LINEAR COMPLEX SYSTEM
Our findings bring studies of foster care placement into the growing body of research on complex, non-linear, adaptive systems. Within this literature, if the system is governed by resource constraints, then the size of the population tends towards a value determined in part by capacity and resource constraints as opposed to population-based need. If we can grasp the existence and nature of these constraints, then we may be able to optimize resource allocation so as to better manage the availability and accessibility of high-quality care.
SYSTEM SCIENCE IN ACTION IN CHICAGO
If you would like to share ideas about how to apply similar multidisciplinary system science research in other policy domains, you can register for our workshop System Science in Action on October 3, 2017, at Chapin Hall at the University of Chicago.