Projections of Juvenile Confinement Populations
Sound projections require high-quality data. Without data, policymakers have only the opinions and beliefs of practitioners and administrators with which to project future needs for bedspace.
The superintendent of a detention center may offer his or her personal observations about crowding in detention. The administrator of a corrections facility may observe that young offenders are being placed on waiting lists because of insufficient space. A county sheriff may complain that officers are required to transport youth to a neighboring jurisdiction to find an opening in a secure facility. Although personal observations may be helpful in making projections, relying on anecdotal information alone may result in costly errors. Each individual involved in the juvenile justice process can explain the process only from his or her unique perspective. Few are aware of every aspect of the process and of the complex interactions between decisions made at various points in the process.
Once policymakers decide to look beyond personal opinions, they need data about the use of detention and corrections space. Unfortunately, the easiest information to assemble is rarely ideal. In some jurisdictions, the only readily available data may be about past uses of detention and corrections space. An agency might only know that admissions to juvenile corrections grew 50 percent during the past 10 years. Some policymakers might interpret this as a legitimate reason to fund an additional 50-percent increase in corrections space over the next 10 years, but this could be a poor decision. Obviously, a jurisdiction that increased its bedspace significantly in 1999 should not rely on the increase in admissions from 1998 to 2000 to argue for yet more bedspace in 2001. Similarly, it would be unfair to use the lack of an increase to argue that an agency does not require additional space. Perhaps a jurisdiction has not funded any new corrections space during the past 20 years. Flat funding would explain the jurisdictionís flat admission numbers, but this would not necessarily mean that additional space is unwarranted.
Policymakers are better served when agencies can generate additional information. For example, researchers could analyze trends in the use of waiting lists and early releases from confinement. An increase in these practices may indicate a growing demand for space. Even this information, however, does not eliminate the risk of misinterpretation. The fact that a juvenile detention center is constantly full with no waiting lists or early releases could have more than one explanation. It could mean that available space is precisely equal to demand, or it could mean that local decisionmakers have learned to refer just enough youth to detention so that a facility remains full without being oversubscribed.
What would policymakers conclude if the same correctional facility suddenly began to report crowding, early releases, and waiting lists for admission? Such a development might indicate an increase in juvenile crime and the need for more space, or it might mean that local authorities had decided to begin referring all potential detention cases for placement and not concern themselves with availability. Projecting future space needs requires more extensive analysis. The question is what type of analysis?
Limits of Simple Models
Juvenile justice agencies often begin their efforts to project detention and corrections populations with relatively simple models. Simple models may provide projections quickly and at relatively little cost, but they can also produce misleading information. One of the most common simple models assesses the need for secure confinement resources according to expected changes in the juvenile population (e.g., youth ages 10 through 17). If a jurisdiction has 100 detention beds and its juvenile population is expected to increase 20 percent over the next 10 years, policymakers might recommend expanding detention capacity to 120 beds over the same period. This approach may be an improvement in a jurisdiction that has previously used only anecdotal methods to anticipate future space needs, but it has great potential for error. Consider the fact that the national population of juveniles was relatively unchanged between 1970 and 1998, a period when juvenile court caseloads more than doubled. An analyst working with population data alone in the 1970ís or 1980ís could have produced very misleading projections (figure 1).
Most juvenile justice administrators know that projection efforts must include at least some data about the juvenile justice process because the number of offenders referred for placement can differ considerably from trends in the juvenile population. One approach commonly used by State and local agencies is to monitor trends in juvenile arrests and then estimate future demand for detention and corrections space based on expected changes in the number of arrests. For example, some jurisdictions base their projections on trends in juvenile arrests for the most serious offenses, such as the Federal Bureau of Investigation (FBI) Violent Crime Index offenses (i.e., murder and nonnegligent manslaughter, forcible rape, aggravated assault, and robbery). The logic behind this approach is that youth charged with violent and other serious offenses generate most of the space needs in any jurisdiction.
The complexity of juvenile justice decision-making, however, virtually guarantees that detention and corrections populations will not follow Violent Crime Index arrest trends so closely. National changes in juvenile arrests during the 1990ís underscore this point. The 1990ís were a virtual case study in how difficult it can be to predict juvenile justice trends. No statistical model could have anticipated the changes in serious juvenile crime that occurred between 1985 and the end of the 1990ís (figure 2).
Consider what would have happened if an analyst working in 1985 had projected changes in the nationwide demand for bedspace using 5-year trends in FBI Violent Crime Index arrests. The projection of bedspace needs in 1990 would have been produced by multiplying 1985 levels of placement resources by the percentage change in Violent Crime Index arrests between 1980 and 1985óa decrease of 9 percent. Arrests for violent offenses, however, were about to jump sharply. A projection from 1985 would have underestimated the volume of arrests in 1990 by 33 percent. An analyst working in 1990 would have been more fortunate using the percentage change in arrests from 1985 to 1990 (up 37 percent) to project space needs in 1995. Yet, a few years later, in 1993, the same technique would have produced estimates for 1998 that were far larger than actual need. No statistician using this method in 1993 would have predicted that juvenile arrests for violent offenses would drop 25 percent between 1994 and 1998.
Extending the period of calculation by using 10-year trends rather than 5-year trends would ameliorate the problem somewhat but not resolve it entirely because the number of arrests is not directly linked to the number of placements. Analysts will produce more useful projections when they include juvenile court processing data in projection models. The juvenile court process is the principal gatekeeper for placements in juvenile bedspace. The juvenile court usually approves detention decisions, or at least it must approve the continuation of detention beyond some statutorily defined limit (e.g., 72 hours). The juvenile court is also the main access point for placement in (or commitment to) long-term facilities. To be admitted to a juvenile correctional facility, young offenders must be referred to court, officially charged with delinquency, adjudicated a delinquent, and then committed by the court. Thus, changes in detention and corrections populations are likely to be more closely related to changing court actions than to changes in juvenile arrests.
This is clear when trends in juvenile arrests are compared over time with trends in juvenile court delinquency cases (figure 3). Between 1980 and 1997, for example, increases in delinquency cases outpaced increases in juvenile arrests. According to the Office of Juvenile Justice and Delinquency Preventionís (OJJDPís) Juvenile Court Statistics program at the National Center for Juvenile Justice, U.S. juvenile courts handled slightly more than 1 million delinquency cases in 1980, just half the number of arrests involving youth younger than age 18 that year. By 1997, the total number of delinquency cases handled by juvenile courts represented 62 percent of the number of arrests.1
Law enforcementís increasing use of court referrals for arrested youth is also apparent when the analysis examines only court cases and arrests that involved FBI Crime Index offenses (i.e., all offenses on the Violent and Property Crime Indexes). In the early 1980ís, the number of court cases involving Crime Index offenses equaled about 70 percent of the number of juvenile arrests involving Crime Index offenses. By the late 1990ís, the number of juvenile court cases involving these offenses represented nearly 90 percent of the number of arrests.
Projection efforts are more useful if they can account for changing patterns in court processing. A changing rate of formal prosecution in juvenile courts, for example, could have a dramatic effect on the number of youthful offenders placed in secure facilities. National data about juvenile court processing reveal, in fact, that the proportion of delinquency cases handled formally (with prosecutor petitions rather than informal agreements for diversion or dismissals) increased from 49 percent to 57 percent between 1983 and 1997 (figure 4).
This shift toward more formal handling could have been expected to increase the number of juveniles eligible for out-of-home placement. An analyst projecting future space needs with this information might still have made significant errors, however, unless the analysis was amended to include an additional factornamely, changes in the use of formal adjudication. Between 1983 and 1997, as the use of formal petitioning increased, the use of adjudication saw a corresponding decrease from 68 percent to 58 percent. When both changes are considered together, it is clear that the total rate of adjudication (adjudication as a percentage of referrals) remained unchanged between 1983 and 1997 (33 percent in both years). This example demonstrates that projection models are likely to perform better when they include more than a single source of information and when they analyze more than a single point in the juvenile justice process.