Chapter 9
CRIMINAL JUSTICE AND CRIME PREVENTION1
By Doris Layton MacKenzie
1. INTRODUCTION
Past behavior is the best predictor of future behavior. From this perspective, it is reasonable to attempt to prevent crime by preventing known offenders from continuing their criminal behavior. This chapter focuses on the options for dealing with actual perpetrators once they are identified so that crime in the community can be reduced. While traditional crime prevention efforts are directed toward those who are not yet involved in crime, our broader definition includes any setting that reduces crime in the community. By definition, therefore, we include as crime prevention, programs in the courts and corrections that focus on reducing the criminal activities of offenders.
For policy purposes, recent interventions for reducing crime through the courts and corrections can be classified into six categories:
(1) Incapacitation or depriving the offender of the capacity to commit crimes usually through detention in prison or capital punishment.
(2) Deterrence or punishment that is so repugnant that neither the punished offender (specific deterrence) nor others (general deterrence) will commit the crime in the future.
(3) Rehabilitation or treatment directed toward changing the offender and thereby preventing future criminal behavior of the treated individual.
(4) Community Restraints or the surveillance and supervision of offenders in the community in order to reduce their capacity and/or opportunity for criminal activities.
(5) Structure, Discipline and Challenge programs that use physically and/or mentally stressful experiences to change the offenders in a positive way or deter them from later crime (specific deterrence).
(6) Combining Rehabilitation and Restraint in order to insure that offenders make changes that are associated with a reduction in future criminal behavior.
As shown in Table 9-1, these are not mutually exclusive categories. The categorization is a heuristic device to classify a wide range of programs currently existing in criminal justice systems throughout the United States. They represent different strategies for controlling crime in the community. Most have some theoretical rationale for expecting a reduction in crime; they differ enormously in the mechanism anticipated to produce the reduction in crime.
Support for these different strategies of crime prevention in the courts and corrections have changed enormously in the past thirty years. In the 1970s, the strong emphasis on rehabilitation that had existed since the turn of the century gave way first to a focus on equality and fairness in sentencing, and then to an increased focus on incapacitation, deterrence and restraint strategies of crime prevention. Today, incapacitation is the primary justification for imprisonment in the U.S. criminal justice system (Zimring and Hawkins, 1995).
A dramatic increase in offender populations accompanied this change in philosophy. The increase was unprecedented. It followed a period of relative stability in incarceration rates that had existed throughout most of the twentieth century. For example, from 1945 to 1974 the average incarceration rate was 106 inmates per every 100,000 individuals in the population. Incarceration rates fluctuated only slightly, from a low of 93 inmates per 100,000 to a maximum rate 119 (Blumstein and Cohen, 1973). Since that time, however, incarceration rates have grown enormously. By 1985, the number of inmates per 100,000 U.S. residents was to 313; this grew to 600 by 1995 (Bureau of Justice Statistics, 1996).
This increase impacted the total correctional populations and not just prisons. Since 1980, the total estimated correctional population rose 179 percent from 1.8 million in 1980 to 5.1 million in 1994 (BJS, 1995). For parole populations the increase was 213 percent, for probation populations the increase was 1,565 percent (BJS, 1995).
While this analysis of crime prevention focuses on how effective these different strategies are in reducing crime, it is important to remember that each strategy has impacts other than crime reduction. For example, analysis of the costs and benefits is critically important in any examination of policy relevant issues. This has been the focus of much of the incapacitation discussion because of the large impact associated with policies that increase the need for building, operating and maintaining the prisons necessary for incapacitation. On the other hand, with the exception of some drug treatment analyses, there are fewer discussions and less research
examining the costs and benefits of rehabilitation. Yet, such analysis is important. A high quality, intensive treatment program for offenders can be relatively costly. The advantages of the program must be weighed against the costs. Such issues, among others, are important in policy decisions.
Table 9-1. Different strategies for preventing crime by the courts and corrections showing the anticipated mechanisms for impact.
CRIME PREVENTION IN THE COURTS AND
CORRECTIONS
Community Structure, Combining
Incapacitation Deterrence Rehabilitation Restraints Discipline and Restraints and
Challenge Rehabilitation
Imprisonment General: Change aspects Increased Experience Offenders can
removes Punitive of offenders surveillance will change be coerced
offenders' punishment will that are and control in offenders in a into
MECHANISM FOR capacity to keep those in changeable and the community positive way rehabilitation
IMPACT commit crimes the community associated with will decrease so they will (forced to
in community from committing criminal offenders not continue take steps to
(General) crimes behavior capacity to to commit positively
commit crimes crimes change)
Small number Specific: Intensive,
of high rate Punitive adequately Increased General and Offenders can
offenders can punishment will implemented surveillance specific be coerced to
be identified keep punished programs (with and control in deterrence remain in
and imprisoned individuals treatment the community treatment
during their from committing integrity) of will decrease longer
active more crimes sufficient offenders'
criminal duration opportunity to Coercion will
career (dosage) commit crimes not diminish
(Specific) the
Target higher Specific effectiveness
General and risk cases deterrence of treatment
Specific
deterrence Cognitive,
skill oriented
and behavioral
treatment
methods
However, the focus of this chapter is on strategies that reduce crime in the community. From this perspective, issues such as costs, prison crowding, reducing risk factors, public opinion, and the politicalization of decision making are considered important only if they have a direct impact on criminal activities and crime prevention. In the following sections of this chapter, these topics are discussed only when they are such an important part of the interpretation of the impact of some programs that they cannot be easily dismissed. In general, the focus of this review is on activities in the courts and corrections that have a direct bearing on preventing crime by reducing the criminal activities of known offenders.
Given the scope of programs and evaluations, examining crime prevention techniques in the criminal justice system is a very large assignment and decisions had to be made about what was important to emphasize in this review. Given the limitations on time, some important topics had to be omitted from this report. Four obvious examples are: Capital punishment; deterrence research not directly related to court or correctional programs; transferring juveniles to adult courts; and community treatment for drug-involved offenders. Another topic that was not described is the relatively new programs on restorative justice and mediation. These are important and current topics, and the interested reader should refer to the recent summaries of the work.
1.1 Examining the scientific evidence
In evaluating the research and assessing the effectiveness of the six identified strategies of crime reduction, this chapter uses three different methods: (1) reviews of the research literature; (2) reviews of meta-analyses used to examine groups of studies; and (3) the scientific methods score combined with significance tests. There were several reasons for this decision to use the different methodologies to review the research and draw conclusions about effectiveness in crime prevention. First, some strategies of crime prevent do not lend themselves to program evaluation that can be easily categorized using the scientific methods score we designed. For example, incapacitation research uses complex statistical models to estimate the crimes prevented by various policy decisions. Such studies do not easily lend themselves to the scoring methodology used for evaluating specific programs. We decided that reviewing the literature on incapacitation, and not scoring each study was more valuable for drawing conclusions about the effectiveness of incapacitation strategies.
Reviews of the literature were also used to judge the effectiveness of several other programs. Most often this was because current reviews of the literature were available, and there was little research that had been completed in the past five or six years that would change the conclusions of the previous reviews. For example, the discussion of shock probation and the Scared Straight programs were based on past literature reviews.
Broad assessments of treatment effects have greatly benefited from the rise of a new statistical technique, meta-analysis that enables researchers to aggregate the continuously growing research literature in order to examine and compare the effect sizes for treatment versus control groups comparisons. In some areas, such as the rehabilitation literature, there is a body of research using meta-analyses to examine the effectiveness of programs. Wherever possible, these meta-analyses are used to draw conclusions about programs. The rationale for this decision is that meta-analysis techniques are respected statistical techniques that are more rigorous than the scientific methods score we are using in this project. The meta-analysis technique permits the aggregation of a large body of research literature in order to examine and compare the effect sizes for treatment versus control group comparisons. The meta-analyses reported herein summarize a large number of studies and control for important methodological issues.
There is an enormous body of literature on crime prevention efforts in criminal justice. Much of this literature is not research examining the impact of crime prevention strategies. Few of the research studies are of sufficient quality to permit conclusions regarding the effectiveness of the program studied. In order to evaluate the quality of the research, as described in Chapter 1, this report uses a scale of 1 to 5 to summarize the scientific rigor of the studies examined. A score of 5 indicates the strongest evidence for inferring cause and effect, a score of 1 indicates the weakest. Studies scoring a 1 were considered so low in scientific rigor that they were excluded from conclusions about a topic. Topics were chosen and the previously described categories were identified by examining current research publications. The focus was on research that had been completed since 1987. In most other cases, the discussion of a topic is based on published reviews of the literature or meta-analyses.
Most of the studies reviewed examine the recidivism of offenders who receive some treatment, service or regime. Studies with scientific method scores of 5 are random assignment studies with successful assignment of cases to treatment conditions, sufficient numbers to reasonably expect that the experiment had sufficient statistical power to detect a difference in recidivism if indeed one was there, and limited attrition during the study.
Two problems that continually arose in the research were the small number of subjects and attrition. Even if a strong experimental design is used, a study will not have sufficient power to detect a difference that may exist if the number of subjects are small. This was particularly a problem with studies using juvenile subjects. Attrition is another problem. Some studies used as comparison groups those who had dropped out of the program being studied. At times this is referred to a comparison between the motivated individuals and others. The problem is these groups can be assumed to differ prior to receiving any treatment and, therefore, no conclusion can be drawn about the effect of the program being studied. This is a serious problem with some of the drug treatment studies.
Frequently studies did not fulfil the requirements for a score of 5 because the assignment process was not successful, there were too few subjects, or subjects were lost during the study. If the effect of these problems could be assumed to be minimal from a research perspective, such studies were given a score of four. Other studies that scored a 4 were quasi-experimental designs with careful statistical controls for differences among subjects. These studies were also required to have limited attrition during the study and sufficient numbers of subjects.
A score of 3 indicated more serious problems with the research design such as limited information on the subjects and comparison groups so that it was impossible to determine how similar the groups were before the study began. Attrition, uncontrolled group differences, and few subjects would also contribute to a score of 3.
A score of 2 revealed serious flaws with the research design, therefore only tentative conclusions could be drawn from such a study because the scientific rigor was so limited.
The next sections of this report review the proposed strategies for preventing crimes through the use of the criminal justice system. Studies for juveniles and for drug-involved offenders are evaluated in separate sections because of the particular focus of these studies. Conclusions about what works, what doesn't and what is promising are based on a careful examination of the literature reviews, meta-analyses, statistical significance, and the scientific methods scores.
2. INCAPACITATION
The concept of incapacitation is simple -- for as long as offenders are incarcerated they clearly cannot commit crimes outside of prison. Crime is reduced because the incarcerated offenders are prevented from committing crimes in the community. At least while they are in prison, they cannot continue to commit crimes. A secondary benefit of incarceration is thought to be the indirect effect of deterring (or inhibiting) others from committing crimes because of the threat of incarceration (general deterrence effect). Furthermore, individuals who spend time in prison may be deterred from continuing their criminal activities when they are released (a specific deterrence effect).
Most people accept the notion that crime prevention through incapacitation is a primary justification of imprisonment (Zimring and Hawkins 1995). Generally accepted also, is the fact that some individuals should be incapacitated for long periods of time because of the seriousness of their offenses and the threat they pose if released. Questions arise over how broadly the incapacitation strategy should be used and whether it is a cost efficient and effective crime prevention strategy. Some ask that prison space be reserved for only a small select group of dangerous repeat offenders. Others advocate a general incapacitation strategy that would incarcerate a substantial number of felons. The success of incapacitation in reducing crime in the community remains a controversial subject.
During the mid-1970s, interest in incapacitation as a crime prevention strategy grew, in part due to concerns about the efficacy of rehabilitation, rising crime rates and public fear of crime. Originally, incapacitation strategies were supported because of what seemed to be the logical utility of keeping offenders in prison so they could not commit crimes. In some jurisdictions increases in incarceration rates were found to be accompanied by decreases in crime rates. This correlation was used to justify incapacitation policies. However, careful scientific examination requires more than an association between two variables because both could easily be caused by some third factor. Furthermore, correlational studies examining the association between incarceration rates and arrest rates within jurisdictions have not found any consistent relationship between the two (Zimring and Hawkins 1995).
More rigorous research examining the effectiveness of incapacitation in reducing crime has focused on developing models to estimate the impact of incarceration on individual level offending (Zimring and Hawkins 1995; Spelman 1995). Estimating the crime prevention benefits that can be obtained through incarceration is a complicated process. The researcher must estimate how frequently offenders commit crimes and the duration of active criminal careers. The majority of studies examining incapacitation effects demonstrate a small but positive effect in reducing crime. Frequently, however, this crime prevention effect is associated with significant increases in prison populations. Issues of concern relate to whether this reduction is worth the additional costs for building and maintaining prisons and jails, and whether there are other more cost-effective methods of crime reduction.
Early research on incapacitation used official-records to estimate individual-level offense rates (Clarke 1974; Greenberg 1975; Van Dine, Conrad, and Dinitz 1977; Petersilia and Greenwood 1978; Blumstein and Cohen 1979). In a review of these studies, the National Academy of Sciences Panel on Research on Deterrent and Incapacitative Effects (1978) reported that the incapacitation studies offered widely divergent estimates of the incapacitative effect of imprisonment. While some studies indicated that incapacitative effects were negligible, others claimed major potential impacts on crimes through increased use of imprisonment. The panel concluded that the principal disagreement was over the value of the individual crime rates that were used to estimate the effectiveness of incapacitative policies. That is, models of the crime reduction effectiveness of imprisonment required estimates of how frequently individuals commit crimes when they are in the community. There were no generally accepted estimates of these rates nor did researchers know how long criminals continued to commit crimes (e.g., the length of the "career" of crime).
Surveys of prisoners conducted during the late 1970s and early 1980s were designed to answer the questions about the individual crime rates and criminal careers of offenders (Peterson, Greenwood, and Lavin 1977; Peterson and Braiker 1980; Peterson and Braiker with Polich 1981; Marquis and Ebener 1981; Peterson, Chaiken, Ebener, and Honig 1982; Chaiken and Chaiken 1982; Greenwood with Abrahamse 1982). If researchers could discover how frequently individuals committed crimes when they were in the community (e.g., individual crime rates) and for how long they continued committing crimes (the career), the information could be used to refine the models predicting how much crime would be reduced by locking them up in prison. The surveys asked prisoners to report on their criminal activities before they had been incarcerated.
Using these estimates, researchers examined the number of crimes prevented by actual and hypothetical criminal justice practices and sanctioning policies. In general, reviews of these "collective" incapacitation strategies demonstrated a modest reduction in crime combined with substantial increases in prison populations. For instance, in a 1987 review of the research on general incapacitation, Visher (1987) concludes that the sentencing practices and policies, that doubled prison populations during the 1970s and early 1980s, resulted in an estimated crime reduction of 10 to 30 percent.
Increases in prison populations and the research findings of large differences in crime rates of individual offenders moved attention towards a more selective strategy of incapacitating a small group of offenders. Encouragement for this selective incapacitation as a crime control strategy also came from research that revealed a small number of very active offenders (six percent of the cohort) accounted for a disproportionately large number of the arrests (52 percent) in a Philadelphia birth cohort (Wolfgang, Figlio, and Sellin 1972). That is, a relatively small number of offenders were responsible for a large amount of crime. Incapacitation advocates argued that crime could be reduced if these "career criminals" were identified and incapacitated. This "selective incapacitation" strategy would identify the offenders who were predicted to commit serious crimes at high rates so that they could be incarcerated for long periods of time. Further support for the benefits of incapacitation as a correctional strategy came from the proposal that, although there were enormous costs to incarcerating large numbers of felons, there were also substantial costs if they were released and continued to commit crimes (in terms of such factors as criminal processing, loss to victims, etc.) (Zedlewski 1987). Some of the practices that can be attributed to these incapacitation strategies are habitual offender laws, mandatory sentences and the more recent three-strikes laws.
In support of the selective incapacitation sentencing policy, Greenwood and Abrahamse (1982) argued that increasing the length of time served by the predicted high-rate offenders while at the same time reducing the time served by those who were predicted to be low-rate offenders could reduce crime rates without a corresponding increase in prison populations. From this perspective, in the face of constraints on available prison resources, the issue for criminal justice policy was how to allocate a limited number of cells among competing offenders. Ideally, the worst (those who commit the most and/or the most serious crimes), active (not yet at the end of the "career" of crime) criminals would be identified and put in the prison cells. Greenwood and Abrahmse (1982) examined whether such policies could reduce the robbery rate in California. They found evidence that through the use of a selective incapacitation strategy the robbery rate could be reduced by about 15 percent, and the number of incarcerated robbers would be reduced by about five percent, but they cautioned that their analysis had several limitations and they suggested that the work should be replicated in other jurisdictions.
Other researchers reviewed Greenwood and Abrahamse's results and concluded that the original analysis greatly overstated the effects of the proposed selective incapacitation (Cohen 1983, 1984; Spelman 1984; von Hirsch and Gottfredson 1984; Visher 1987). And, in 1983 the National Academy of Sciences panel on Criminal Careers commissioned a reanalysis of the original survey data. The estimates resulting from this study indicated substantially smaller incapacitative effects than those found by Greenwood and Abrahamse. Furthermore, substantial increases in the prison population were predicted (Blumstein et al. 1986).
For selective incapacitation to be effective, it must be possible to identify and incarcerate offenders who will commit the most crimes in the future. The estimates of crime commission rates used in the incapacitation studies were derived from the RAND studies using inmate self-reports of criminal activities prior to incarceration. Greenwood and Turner (1987) investigated whether this retrospective data could be used to predict future arrests. They found their data had poor predictive accuracy and, furthermore, the small differences in arrest rates between the groups classified as high- and low-risk did not justify the large differences in sentence lengths that would be required for these offenders if significant selective incapacitation effects were to be achieved. This research, as well as the more recent work of Gottfredson and Gottfredson (1994), suggests that identifying future offenders in order to selectively incarcerate them will prove difficult.
Surprisingly there was little research on the magnitude of incapacitation effects during the decade of the 1980s when the incapacitation philosophy drove the largest increase in incarceration in American history (Zimring and Hawkins 1995). A few more recent studies were completed in the early 1990s (Miranne and Geerken 1991; Horney and Marshall 1991; English and Mande 1992). In one recent study, Cohen and Canela-Cacho (1994) studied the relationship between incarceration and levels of violent crime using both national data and corrections data from six states. The study focused on the crime control effects of incarceration especially whether incarceration was an effective strategy for controlling violent crimes and the merits of pursuing alternative incarceration policies. Their data indicate that changing prison policies such as guidelines, mandatory minimum prison terms, and restrictions on parole release, have played a major part in the rising prison populations over the last decade. In comparison to the past, a higher proportion of those who are arrested in the U.S. are sentenced to prison and those who are committed to prison stay there for longer periods of time. This increased risk of being sent to prison after arrest, and of spending more time if committed to prison was true, in general, for those convicted of various types of crimes and across different jurisdictions.
In their research, Cohen and Canela-Cacho (1994) used sophisticated estimating techniques that took into consideration the fact that high-rate offenders are over-represented among inmates while low-rate offenders are disproportionately found among the offenders who remain free, and the fact that termination of criminal careers reduces the crime prevention effects derived from increased in incarceration. They estimate that the incapacitation effects during periods of low incarceration rates are probably much greater than previously estimated, and that the increasing numbers of offenders being incarcerated today bring only marginal returns for incapacitation effects. This occurs because the expanding prison populations are likely to include an increasing number of offenders who would be low-rate.
The focus on tougher sentencing laws has led to increasingly rigid sentencing statutes and these have particularly impacted repeat offenders. By 1994, 30 states had introduced "three-strikes" legislation and ten had passed tougher sentencing for repeat offenders (Benekos and Merlo 1996). The "three-strikes and you're out" baseball metaphor is used throughout the country in reference to criminal sanctions that become increasingly severe for each conviction an offender receives until they are consider to be "out" or in prison for life. Greenwood, Rydell, Abrahamse, Caulkins, Chiesa, Model and Klein (1994) estimated the crime prevention impact of the California three-strikes law, one of the most sweeping of the laws. Although the first two "strikes" accrue for serious felonies, the third strike that triggers the life sentence can be any felony. According to their estimates the new law would reduce serious felonies committed by adults in California between 22 and 34 percent below what would have occurred had the previous law remained in effect. One third of these prevented felonies would be violent crimes such as murder, rape, and assaults. The remaining two-thirds of the prevented crimes would be less violent felonies such as less injurious assaults, and most robberies, and burglaries of residences. Several alternative models were tested to see if other less costly options would be predicted by the model to be as effective as the three-strikes laws. Although these options were predicted to drop the costs, they would also drop the effectiveness. The researchers caution, that while these results appear encouraging for crime prevention, it will come at great financial cost due to the large estimated increase in prison population. For example, the California three-strikes law, if applied in all eligible cases, would reduce the number of serious felonies in a year by about 28 percent or 329,000 crimes. However, this would cost an additional $5.5 billion a year in additional criminal-justice funding for the additional costs of the construction and operation of prisons. This can be translated as a cost of $16,000 per serious felony prevented.2
In summary, there is now a body of research examining the crime prevention effectiveness of incapacitation policies. In general the results indicate:
o Incapacitation policies prevent crime because offenders who are imprisoned do not have the opportunity to commit crimes;
o There are a small number of offenders who commit a large number of crimes. If they could be incapacitated a large number of crimes would be prevented.
However, there are many unresolved questions that make the effectiveness of this strategy questionable. Most important are the following issues:
o It is not yet possible to predict who will be the high frequency offenders in the future; therefore targeting them for increased prison sentences is impossible;
o Increased use of incapacitation as a crime prevention strategy must address the expected increases in imprisonment rates and the associated financial costs that accompany such strategies;
o Large increases in the use of incapacitation may have limited returns because the additional offenders not now incarcerated are lower frequency offenders who would not be committing many crimes in the community, thus, reducing the return on investment for every new dollar expended;
o Large increases in the use of incapacitation may also have limited returns because offenders who are incarcerated for lengthy periods of time may be at the end of their criminal career and therefore would not be committing crimes in the community;
o True estimates of the crimes prevented are difficult to obtain because both the frequency of criminal participation and the duration of careers must be estimated.
Furthermore, recent studies of the impact of the increases in imprisonment rates that have occurred in the past twenty five years have revealed that the impact has had a major impact on minority populations in urban environments (Tonry, 1995). Other disadvantages of increased use of imprisonment strategies are the unintended consequences of imprisonment on the families and communities of those who are imprisoned (Clear, 1996).
3. DETERRENCE
Deterrence strategies are based on early criminological theory proposing that sufficiently repugnant punishments will inhibit individuals from committing crime. As is obvious from Table 1, deterrence could be an expected impact of incapacitation, community restraints and challenge programs. However, this is secondary to the primary mechanism that is expected to have an impact on crime prevention for these strategies. Here, the programs classified as deterrence are those with a primary purpose of deterring either the individual offender or others through the repugnant nature of the sanction. At the individual level, specific deterrence is explained by the fact that the pain generated by the punishment will serve to discourage future criminality. It assumes a rational choice model of decision making where the offender perceives that the cost and benefits of punishment are not outweighed by the crime. General deterrence refers to the impact of the threatened punishment has on other potential offenders, thus reducing the chance that they will commit crimes.
Deterrence is the rationale given for programs such as Scared Straight, chain gangs and shock probation. These are distinguished from other strategies because the major emphasis is on the punitive nature of the punishment and not on reducing crime through restraint, discipline or challenge. Another deterrence strategy is that of fines, particularly day fines. These fines are designed to be fair given the difference in the economic circumstances of the individual offenders thereby making this sanction more punitive than the tradition fines.
Research examining two types of deterrence programs is reviewed in this section. It is important to note, as shown in Table 9-1 that other programs such as incapacitation policies that threaten offenders with longer prison sentences as well as the programs requiring offenders to participate in emotionally and physically strenuous programs (e.g., structure, discipline and challenge) are also expect to deter offenders. However, they are also viewed as potentially having other impacts and, therefore, they have been examined in separate sections.
3.1 Monetary Penalties
Fines are frequently used as criminal penalties for a wide variety of cases in American courts (Hillsman, 1990; Hillsman, Sichel, and Mahoney 1984; Casale and Hillsman 1986; Cole et al. 1987). However, many of the fine sentences are composites of fines and other noncustodial sanctions and not stand alone sanctions. Judges have wide discretion in setting fines. They are not uniformly imposed, and jail sentences are sometimes used as alternatives to fines particularly for the poor. Rarely are fines in the U.S. used as the sole sanction for more serious cases or for repeat offenders. In contrast, in Western Europe fines are the most often imposed sentence for most crimes and are a major alternative to imprisonment (Hillsman 1990). One of the differences between the use of fines in the United States and other countries is the fact that American judges are not able to set fines that are proportionate to the severity of the offense but are also equitable and fair given the difference in the economic circumstances of the individual offenders. "Day" or "unit" fines as they are called in Western Europe are linked to both the offender's daily income and to the gravity of the crime.
In terms of crime prevention, fines may act as a deterrent to criminal activities. Most studies of fines, however, have focused on setting just and proportionate levels for the amount of the fine, or on compliance, cost savings, or prison population impact issues. Until recently little was known about the use of fines as criminal penalties in the United States. NIJ has filled this gap by supporting studies to examine fining practices (see Hillsman et al. 1984; Casale and Hillsman 1986; Cole et al. 1987; Glaser and Gordon 1988; Hillsman and Green 1987, 1988). These studies are discussed in this chapter only if they examine the impact of fines on crime prevention. However, it is important to note that the results from these studies demonstrate that the use of fines is widespread throughout the U.S. and they are used for a wide range of offenses. Collection rates vary greatly by court, however, these difference may be due to differences in collection techniques and enforcement strategies. The assumption is that these practices could be improved to achieve greater compliance (Casale and Hillsman 1986).
Gordon and Glaser (1991) did examine the impact of traditional fines on recidivism in a quasi-experimental study comparing financial penalties versus similar sentences (probation or probation plus jail). While there were no significant difference between groups, as shown in Table 9-2, offenders who received a fine with probation have lower recidivism rates than offenders who received only probation. Similarly, those who received a fine with probation and jail had lower recidivism than offenders who received only probation and jail without the fines.
As yet, there are few jurisdictions in the United States that use the day fine concept. However several courts are currently adapting day fines to the American context and they are experimenting with their use (Hillsman 1990). Two studies have examined the impact of day fines on the recidivism of offenders. One study assessed the recidivism of offenders sentenced in Milwaukee's Municipal Court Day-Fine Pilot Project and compared the recidivism rates to a comparison group who received traditional fines (Worzella 1992). A larger proportion of the day-fine offenders paid their fine in full. There was no difference in the percent of the groups who committed further violations of municipal ordinances but the day-fined group had fewer arrest warrants (neither measure was significantly different).
In 1991, the Bureau of Justice Assistance funded a multi-site demonstration project, "Structured Fines Demonstration Project," designed to enhance the application and enforcement of structured fines (day fines) as sanctions for drug offenders and other misdemeanants and felons. NIJ awarded funding to Turner and Petersilia (1996 ) to complete an evaluation the project. While most of the research focused on the implementation and development of the day fine programs, there was some outcome data from one of the jurisdictions. As shown in Table 2 , day fines were associated with reductions in both technical violations and rearrests.
Overall, there is a limited amount of research examining the effectiveness of fines in reducing the recidivism of offenders. The Gordon and Glaser (1991) study suggests that fines as additions to other sanctions may be effective in reducing recidivism. Since fines could potentially reduce the cost of courts and corrections, and day fines address the problems of inequality, this strategy appears to be a promising avenue for future research.
3.2 Shock Probation, Shock Parole, and Split Sentences
The programs is this section have been grouped together because the major emphasis of the programs has been on specific deterrence of the offender participants. Shock probation or parole programs are a form of split sentence in which offenders are incarcerated for unspecified short periods of time in prisons or jails followed by a period of community supervision. The idea is that a short period of time incarcerated would "shock" offenders into abandoning criminal activity and into more conventional and law-abiding behavior. During their incarceration there are no special programs for them and they are mixed with other offenders in the jail or prison. Reviews of the research examining shock programs has provided little evidence of a deterrent effect. Studies examining the recidivism of shock probationers with similar probation groups have found no differences and in some cases the shock probationers have done demonstrably worse (Vito, 1984; Vito and Allen, 1981; Boudouris and Turnbull, 1985; Finckenauer, 1982).
"Scared Straight" is another program designed to deter young offenders or at-risk juveniles from continuing criminal activities. They are taken to maximum security institutions where inmates tell them the horrors and difficulties of life in prison. Studies of these programs have not indicated any differences between those who participated in the programs and comparison groups and in some cases the re-arrest rates were higher for those who participated in the program (Buckner and Chesney-Lund 1983; Lewis 1983).
Overall there is no evidence that deterrence programs such as these effectively reduce the future criminal activities of the offender participants (see also section on rehabilitation in this chapter).
Table 9-2. Studies of fines and recidivism showing scientific methods score and findings.
Study Scientific Findings
Methods Score
Gordon and Glaser (1991) 3 Probation + fine had fewer arrests
(25%) than probation only (36%), S.
Probation + jail + fine had fewer
arrests (37%) than probation + jail
only (50%), S.
Worzella (1993) 3 No differences between day fine
group in new ordinance violations
(33%) and conventional fine group
(34%), NS.
Turner and Petersilia 5 Day fine group had fewer technical
(1996) violations (9%) than conventional
sentenced group (22%), S.
Day fine group had fewer rearrests
(11%) than conventional sentenced
group (17%), NS.
Note: NS=nonsignificant, S=significant
4. REHABILITATION AND TREATMENT
In contrast to incapacitation, rehabilitation strategies focus on changing individual offenders so they will not continue their criminal activities. The research goal is to identify and understand the individual differences that explain criminal behavior and how interventions can be used to change individuals so they will not continue to commit crimes. The work is based on psychological theories of learning, cognition and the general principles of human development applied to the analysis of illegal behavior (Andrews and Bonta 1994). Research has focused on examining the components of programs that are effective in reducing recidivism.
Since the mid-1970s there have been major changes in how the courts and corrections manage offenders in this country. One of the most visible influences on this change was the report by Martinson (1974) that was widely interpreted as showing that "nothing works" in rehabilitation. Distilled from a larger coauthored research report by Lipton, Martinson, and Wilks (1975), Martinson's essay described the results of the research teams' assessment of 231 evaluations of treatment programs conducted between 1945 and 1967. From this research, he concluded "With few and isolated exceptions the rehabilitative efforts that have been reported so far have had no appreciable effect on recidivism (1974:25; see also Sechrest, White, and Brown 1979 and Martin, Sechrest, and Redner 1981).
Critics (Gendreau 19811; Gendreau and Ross 1979, 1981, 1987; Gottfredson 1979; Cullen and Gilbert 1982, Greenwood and Zimring 1985; Halleck and Witte 1977; Palmer 1975; Palmer 1983; Van Voorhis 1987) argued against this conclusion, saying it was not that treatment programs could not potentially reduce recidivism, but instead that it was impossible to draw any conclusions from the research because:
1. The research methodology was so inadequate that few studies warranted any unequivocal interpretations about what works; and,
2. The programs studied were so poorly implemented and delivered in such a weakened form that they would not reasonably be expected to have an impact.
The predominantly negative reviews of rehabilitation that dominated the 1970s were challenged by researchers such as Palmer (1975; 1983) who argued that the broad generalizations of the conclusions overlooked many positive instances of success and the researchers gave little attention to such important issues as the fit between the type of offender and the type of treatment provided. Reviews of evaluations published after Martinson's essay indicated that substantial research exists showing the effectiveness of correctional treatment (Gendreau 1981; Gendreau and Ross 1979, 1981, 1987; Gottfredson 1979; Cullen and Gilbert 1982;, Greenwood and Zimring 1985; Halleck and Witte 1977; Van Voorhis 1987). However, despite the critiques of the work and its questionable validity, the phase "nothing works" became an instant cliche and exerted an enormous influence on both popular and professional thinking (Walker, 1985; Cullen and Gendreau 1989; Tonry 1996; Stojkovic 1994). The perception of the conclusion became widespread throughout the U.S. and it gave rise to a strong movement to change both the philosophy and control of imprisonment policy and this impact was felt throughout the 1980s.
Today, while there is still some debate about the effectiveness of rehabilitation (e.g., Lab and Whitehead 1988; Whitehead and Lab 1989) recent literature reviews and metaanalyses demonstrate that rehabilitation programs can effectively change offenders (Andrews and Bonta 1994; Andrews, Bonta, and Hoge 1990; Andrews, Zinger, Hoge, Bonta, Gendreau, and Cullen 1990; Palmer 1975; Gendreau and Ross 1979, 1987). In general, according to Andrews et al. (1990), reviews of the literature show positive evidence of treatment effectiveness. For example, in a series of literature reviews, the proportion of studies reporting positive evidence of treatment effectiveness varied from near 50 percent to 86 percent: 75 percent (Kirby 1954), 59 percent (Bailey 1966), 50 percent (Logan 1972), 48 percent (Palmer's 1975 retabulation of studies reviewed by Martinson in 1974), 86 percent (Gendreau and Ross 1979) and 47 percent (Lab and Whitehead 1988).3 In reviewing these studies, Andrews et al. (1990) conclude that "This pattern of results strongly supports exploration of the idea that some service programs are working with at least some offenders under some circumstances." The important issue is not whether something works but what works for whom.
What is clear is that some approaches to treatment are better than others. Psychological researchers emphasize that effective treatment programs must follow some basic principles (Gendreau and Ross 1979, 1987; Gendreau and Cullen 1989). First, treatment must directly address characteristics that can be changed (dynamic) and that are directly associated with an individual's criminal behavior (criminogenic factors). There are numerous risk factors associated with criminal activity. Age, gender and early criminal involvement are some examples. In comparison to others, young males who began criminal activities at a young age are higher risks for future criminal activities. But these "static" characteristics such as age, gender and past history, while predictive of recidivism, cannot be changed in treatment. Instead, the "dynamic" or changeable factors should be the target of treatment programs.
Equally as important is the distinction between factors that are criminogenic and those that are not. Criminogenic factors are those that are directly associated with criminal behavior. Research has revealed some dynamic factors that are also criminogenic: attitudes, cognitions, behavior regarding employment, education, peers, authority, substance abuse and interpersonal relationships that are directly related to an individual's criminal behavior. Less promising targets for reducing future criminal behavior include increasing selfesteem without touching antisocial propensity, or increasing the cohesiveness of antisocial peer groups. While factors such as self esteem many be correlated with criminal behavior, changing them will not necessarily reduce future criminal activities. That is, criminals may have relatively strong self concepts but they may continue to commit crimes. Treatment programs that target such noncriminogenic factors will not be particularly successful in reducing recidivism. In order to be successful, treatment must address factors that can be changed (e.g., dynamic factors) and that are directly related to an individual's criminal behavior (criminogenic).
In a recent metaanalysis examining predictors of adult recidivism, Gendreau, Little and Goggen (1996) found antisocial cognitions, values, and behaviors (dynamic, criminogenic factors) along with static factors (history, age, gender, race) were the strongest predictors of recidivism. This provides support for the proposal that these changeable factors should be targeted in treatment. In contrast, self esteem, depression and anxiety were relatively weak predictors of recidivism. These characteristics are commonly targets of treatment despite the fact that they appear to have little association with recidivism.
Also important in determining whether a treatment program will be effective is the therapeutic integrity of the program or the need for effective programs to be delivered as planned and designed. Poorly implemented programs, delivered by untrained personnel, where offenders spend only a minimal amount of time in the program, can hardly be expected to successfully reduce recidivism.
Furthermore, programs must target offenders who are at sufficient risk for recidivism so that this reduction is measurable. Many offenders are low risk for future recidivism. Treatment programs that provide intensive services for such offenders will show little reduction in future criminal activities because few of these offenders will recidivate anyway.
The final principle of effective treatment is the need to deliver treatment in a style and mode that addresses the learning styles and abilities of offenders. For example, more effective programs follow a cognitive behavioral and social leaning approach rather than nondirective relationship-oriented counseling or psycho-dynamic, insight-oriented counseling.
Using these principles as the basis to classify studies of treatment as appropriate or inappropriate, Andrews et al (1990) undertook a meta-analysis of 154 treatment comparisons.4 Most often studies were classified as appropriate because the treatment was behavioral in nature. Few studies could be classified on the basis of risk or treatment integrity. Inappropriate treatments were those that employed deterrence (e.g., "Scared Straight"), nondirective approaches, non-behavioral milieu approaches, and group interactions. They found an effect size of .63 for appropriate treatment and this was significantly larger than the mean values for inappropriate services and criminal justice sanctions (warnings, probation, intensive probation, custody). Over all they found an effect size of .21 for the effectiveness of treatment programs. The researchers do note that, considering the enormous number of offenders who have passed through the criminal justice system, there are a comparatively small number of evaluations of appropriate correctional programming.
Lipton and Pearson (1996) found some, but limited, evidence corroborating the finding that treatment programs could be classified by the appropriateness of the treatment provided. These researchers are currently working on a comprehensive, detailed review of the evaluation research on rehabilitation programs for offenders, the Correctional Drug Abuse Treatment Effectiveness Project (CDATE). The project, funded by The National Institute on Drug Abuse (NIDA) at the National Institutes of Health, has been collecting evaluation studies conducted from 1968 until 1994 in order to assess the effects of correctional interventions on various outcome measures (e.g., drug use, recidivism). Preliminary findings from a meta-analysis of the first 500 coded evaluation studies (they anticipate over 1,500 studies) replicated the findings of the Andrews et al. (1990) study on the significance of the appropriateness of treatment. However, while Andrews et al. found a correlation of .69 between appropriateness of correctional service (as defined by the Andrews group) and recidivism (using effect size), Lipton and Pearson found a correlation of only about half that size (.34). They speculate that part of this difference may be because the conceptualization of appropriateness of correctional service still has some ambiguity that results in differences in the categorization of studies used in meta-analyses. From the perspective of crime prevention, the implication is that there are some difficulties in identifying what is appropriate treatment. However, the preliminary analysis of the CDATE data does support the conclusion that specific types of correctional treatment are associated with lower rates of recidivism. The difficulty appears to be in identifying exactly what characteristics are associated with effective treatment.
In summary, there is evidence that:
o Rehabilitation is effective in reducing the criminal behavior of at least some offenders.
The evidence from the meta-analyses suggests that effective correctional treatment programs appear to follow some basic principles. In order to effectively reduce recidivism, treatment programs appear to need to:
o Be carefully designed to target the specific characteristics and problems of offenders that can be changed in treatment (dynamic characteristics) and those that are predictive of the individual's future criminal activities (criminogenic) such as antisocial attitudes and behavior, drug use, anger responses;
o Be implemented in a way that is appropriate for the participating offenders and utilizes therapeutic techniques that are known to work (e.g., designed by knowledgeable individuals, programming provided by appropriately educated and experienced staff, use of adequately evaluated programs) and require offenders to spend a reasonable length of time in the program considering the changes desired (deliver sufficient dosage);
o Give the most intensive programs to offenders who are at the highest risk of recidivism;
o Use cognitive and behavioral treatment methods based on theoretical models such as behaviorism, social learning or cognitive behavioral theories of change that emphasize positive reinforcement contingencies for prosocial behavioral and is individualized as much as possible.
More information is needed regarding: (1) how to ensure that treatment programs have adequate integrity; (2) what should be targeted in the treatment (antisocial attitudes, values, employment behavior, education, etc.); and, (3) what method should be used to deliver the treatment (required staff training, outpatient, in-prison programs).
5. COMMUNITY RESTRAINTS
Many of the sanctions and correctional options categorized as community restraints are frequently referred to as intermediate sanctions or alternative punishments. However, here the term community restraint refers to the fact that a group of these alternative punishments increase the amount of surveillance and control over offenders while they are in the community. In a sense, they might be referred to as "semi-incapacitation" because they are expected to reduce offenders' ability to commit crimes. Examples of restraint programs are intensive supervision, house arrest, electronic monitoring, and halfway houses. Theoretically, increasing the surveillance and control over offenders in the community will prevent criminal activities by reducing both their capacity and their opportunity to commit crimes. Additionally, it is expected that the punitive nature of the sanctions will act as a specific deterrence to reduce the offenders future criminal activity.
In response to the record numbers of convicted offenders and widespread prison crowding, correctional officials in recent years have expanded the range of intermediate sanctions that fall between traditional probation and complete incarceration (Cullen, Wright, and Applegate 1996; Tonry and Lynch 1996; Byrne, Lurigio and Petersilia; Harland 1996; Smykla and Selke 1995). House arrest, intensive supervision, curfew, day reporting and other intermediate sanctions fulfill many purposes. They provide graduated punishments that may be more appropriate than either probation or prison for some offenses, and they maintain a higher level of offender restraint and accountability than does standard probation or parole supervision. In addition, intermediate sanctions may provide enhanced levels of treatment or services for problems that are common among criminal offenders, such as drug abuse, low education levels and unemployment. Finally, when used in lieu of confinement, intermediate sanctions may reduce prison or jail populations and associated costs.
This section examines sanctions that increase the restraints on offenders in the community and studies assessing the effectiveness of these restraints in reducing criminal activity. The term restraints is used to refer to activities such as contacts with agents, urine testing (see section on drug-involved offenders, this chapter) and employment verification that represent control over offenders and increased accountability. Since these restraints and not rehabilitation are the primary focus of the research, this section examines whether the sanctions are effective in preventing the future criminal activities of these offenders. That is, the research is designed to investigate such factors as the number of contacts with probationers, curfews, or confinement to home, and not the amount of rehabilitation provided within the programs. However, when available, follow-up studies or components of studies that examine the effectiveness of rehabilitation within these restraint-focused sanctions are reviewed as well. All of these sanctions increase the restraints above what would usually occur in traditional probation or parole.
Most jurisdictions in the U.S. have some type of intermediate sanctions programs. These programs have variously been called correctional alternatives, intermediate sanctions, community corrections or, more recently, correctional options. As a result of disillusionment with the effectiveness of rehabilitation and the focus on justice and incapacitation, intermediate sanctions were proposed as an ideal way to provide a range of sanctions between probation and parole (Morris and Tonry 1990; Tonry 1996). Theoretically, these sanctions could be scaled in severity so as to be proportionate to the seriousness of the crimes committed. Furthermore, the additional control and threat of sanctions were expected to either deter offenders from future criminal acts or restrict them (in a sense incapacitate them) so they would not have the opportunity to reoffend.
Throughout the 1980s and 1990s, NIJ funded a wide range of evaluations of different correctional alternatives. There is now a body of research that permits us to draw some conclusions about the crime prevention effectiveness of these programs. This section reviews the literature and research on some of the major alternative sanctions. The focus of most of these studies has been the recidivism rates of offenders who are given sanctions that increase the degree of control and surveillance over their activities. In the majority of cases no significant differences are found between offenders placed in alternative sanctions and the comparison groups. Except in a few instances, there is no evidence that these alternatives are effective in reducing crime as measured by official record data. The problem is that most of these alternatives increase the probability of detection. It is unknown whether the actual offense rates change. That is, the increased probability of detection may mean that the intensively supervised offenders are at higher risk of being caught when a criminal act is committed, compared to the comparison offenders, who may commit crimes much more frequently.5
The most hopeful sign from this work is in the exploratory research that has followed most of these evaluations. This research focused on alternative sanctions that increase the treatment and therapeutic aspects of the programs and compares the effectiveness of such programs with similar alternatives that do not include treatment or therapy. The findings suggest that if sanctions include appropriate treatment, the recidivism of the offenders receiving the treatment may be reduced. From this perspective, it is not the restraints that are effective in reducing the criminal activities of the offenders, but rather, their criminal activities are reduced through the treatment they receive.
5.1 Intensive Community Supervision
Compared to regular probation and parole services, Intensive Community Supervision, usually called Intensive Supervised Probation (or Parole) or ISP, was designed to provide increased restraints on offenders in the community (Lurigio and Petersilia 1992; Petersilia and Turner 1993; Cullen, Wright and Applegate 1996; Tonry and Lynch 1996) Studies of ISP do indeed reveal that there are increased direct contacts between the offenders and the supervising probation or parole agent. Many programs combine other options such as electronic monitoring and/or home confinement with the increased agent-offender contacts. Furthermore, indirect methods of observation are also frequently combined with the ISP programs. Many times offenders are required to report for more frequent urine testing or agents may conduct regular employment verification. In all, these direct and indirect observations provide substantially increased levels of control within probation and parole programs. However, the type and level of demands placed on offenders differs enormously by jurisdiction. Offenders are often required to pay fines, keep a mandatory curfew, or provide community service and these additional requirements also differ by jurisdiction.
ISP programs grew dramatically in the 1980s, and by 1990 virtually every state in the nation had developed some type of ISP program. In part, this was the result of the initial research examining the programs in New Jersey and Georgia where the findings suggested that ISP led to a significant decrease in reincarceration (in Georgia, see Erwin 1986) and rearrests (in New Jersey, see Pearson 1987). However, critical reviews of the research demonstrated that the data did not support the initial unqualified conclusions about the ability of the ISP programs to reduce crime.
Recognizing the limitations of the prior research, the National Institute of Justice funded RAND to evaluate fourteen ISP programs in nine states using an experimental design (Petersilia and Turner 1993). The research was greatly facilitated by the addition of funds from the Bureau of Justice Assistance to assist sites in the additional costs for the increased probation and parole staff required to provide the additional supervision. This experimental design with random assignment of offenders to ISP and control groups eliminated many of the past methodological problems of the earlier studies. Recidivism was measured using arrests and technical violations. The results were disappointing for the ISP advocates. When ISP participants were compared to the control group, there were no significant differences in arrests. At the end of the one year study period, about 37 percent of the ISP participants and 33 percent of the control offenders had been arrested. In comparison, the researchers found a significant difference when the technical violation rates were examined. The average ISP violation rate was found to be 65 percent for ISP participants compared with 38 percent for the controls. In summary, while there was no evidence that the increased surveillance in the community deterred offenders from committing crimes, it did appear that this additional control increased the probability that criminal or technical violations would be detected.
In another study funded by NIJ, Deschenes, Turner and Petersilia (1995) found similar conclusions -- no difference between comparison groups and two groups of offenders on ISP programs. One ISP group was diverted from prison and the other was given ISP after release from prison. Supporting the results of the previous 14 site study, findings revealed no significant difference between groups. As shown in Table 9-3, there is a fairly substantial body of research now available on ISP. Few of the studies found statistically significant differences and the direction of the differences between the ISP groups and the comparison groups varied, sometimes favoring ISP, sometimes favoring the alternative.
Although research has not revealed a significant relationship between levels of surveillance and recidivism, there is some evidence that increased treatment of offenders in ISP programs may be related to significant reductions in rearrests. Follow-up analyses by the RAND researchers (Petersilia and Turner 1993) and also researchers evaluating ISP programs in Massachusetts (Byrne and Kelly 1989), Oregon (Jolin and Stipack 1991) and Ohio (Latessa 1993) have found evidence that rearrests are reduced when offenders receive treatment services in addition to the increased surveillance and control of the ISP programs. For example, Petersilia and Turner (1993) reported a 10 to 20 percent reduction in recidivism for those who were most active in programs while they were in the community. However, the research designs used in these evaluations do not reach the experimental rigor of the random assignment study by RAND that examined the effect of increasing the surveillance and control of ISP participants.
In summary the results indicate:
o Increasing the surveillance and other restraints on offenders in ISP programs is not associated with decreases in recidivism;
o The increased surveillance of ISP is often associated with increases in technical violations of the conditions of the ISP programs;
o Incorporating treatment into the requirements of ISP programs may lead to a reduction in recidivism but this research has not be as rigorously examined.
5.2 Home Confinement
Home confinement is designed to regulate and restrict the freedom of the offender within the community (Renzema, 1992; Baumer and Mendelsohn 1992). The terms "house arrests," "home confinement" and "electronic monitoring" are often used interchangeably. However, it is important to note that house arrest, home confinement and more recently "community control" are terms describing the programs, while electronic monitoring is a tool used to monitor the compliance with the requirements of the sentence. During the 1980s, technological advances made it possible to monitor offenders electronically to insure that the offender was complying with the requirements of the program. Unlike ISP, house arrest is usually a sentence given by the court that are much more restrictive than ISP.
In general home confinement programs had targeted low-risk offenders such as those convicted of Driving While Intoxicated (DWI). However, more recently home confinement has been used for parolees (Beck and Klein-Saffran 1989) or other more serious offenders (Baumer, Maxfield and Mendelsohn 1993; Baumer and Mendelsohn 1991; Austin and Hardyman 1991). Early research examining the effectiveness of the home confinement programs suffered from poor research designs, lack of program integrity, and the low risk offenders placed in the programs.
Recidivism rates of the low-risk offenders placed in home confinement programs are usually very low. Therefore, many studies do not have the power to detect small differences that might be expected between the participants and control groups. Two studies using experimental designs shown in Table 9-4 found no significant difference in recidivism when the behavior of offenders who are electronically monitored on home confinement is compared with those being manually supervised (Baumer and Mendelsohn 1991; Austin and Hardyman 1991).
In summary:
o Most likely because of the low-risk offenders in the programs, most home confinement programs have low recidivism and technical violations rates;
o The available evidence indicates no significant differences in recidivism outcome when the participants are compared to control groups.
Table 9-3. Studies of intensive supervised probation/parole (ISP) showing scientific methods score and findings (from Cullen et al. 1993 and Byrne and Pattavina 1992).
Study Scientific Methods Findings
Score
Fallen et al. (1981) 3 Recidivism lower for ISP
Erwin (1986) 3 Recidivism rates lower for ISP
than probationers and lower for
prison releasees.
Mitchell et al. (1986) 3 Recidivism higher for ISP than
parolees and CCC.
Pearson (1987) 3 ISP recidivism rates lower
Byrne and Kelly (1989) 3 ISP lower recidivism
Molof (1991) 2 ISP lower recidivism than
probationers
Jolin and Stipack (1991) 4 Recidivism rates higher for ISP
than EM and work release groups
Latessa (1991a) 3 Recidivism rates higher for ISP
offenders than three comparison
samples
Austin and Hardyman 3 ISP-EM offenders arrested more
(1991) than parolees
NCCD (1991) 3 Recidivism rates for ISP-jail
probationers and ISP-parolees
higher than comparisons but
rates lower for ISP parolees
than comparison
Latessa (1992) 3 Recidivism rates higher for ISP
Latessa (1993) 2 Offenders in drug and mental
health units had higher
recidivism rates than those in
sex-offenders and alcohol
offender units.
Moon and Latessa (1993) 3 ISP drug program participants
had lower recidivism rates
Latessa (1993b) 3 ISP groups had higher recidivism
than probationers.
Petersilia and Turner 5 ISP sample in 10 states had
(1993) higher recidivism than
comparison. ISP samples in 4
states had lower recidivism than
comparison.
Table 9-4. Studies of home confinement and electronic monitoring showing scientific methods score and findings.
Study Scientific Findings
Methods Score
Baumer and Mendelsohn 5 EM had more revocations (21%) than
(1991) Manual supervision (18%), NS.
Austin and Hardyman 5 EM arrested (14%) more than
(1991) controls (11%), NS.
Note: NS=nonsignificant, S=significant5.3 Community Facilities: Residential and Day Reporting
Halfway houses, also called community residential centers, pre-release centers, or restitution centers, are nonconfining residential facilities for adjudicated adults or juveniles, or those subject to criminal or juvenile proceedings (pre-trial period). They are intended as an alternative to confinement for persons not suited for probation or who need a period of readjustment to the community after imprisonment. More halfway houses provide services for juveniles than for adults, and some houses specialize by client or treatment modality such as women only, pre-release, substance abusers, or developmentally disabled (Latessa and Travis 1991). The facilities are included as community restraints because most of the research reviews have focused on their use as additional restraint and not on the details of the services provided.
Research examining the effectiveness of halfway houses in reducing recidivism has indicated mixed results. In an early evaluation of correctional halfway houses, Allen et al. (1976) reviewed 35 studies. The majority of the studies used quasi-experimental designs or nonexperimental designs; only two used true experimental designs. The evidence was about equally divided between lower recidivism for the halfway house residents and no differences in recidivism in comparison to control groups in the quasi-experimental and experimental designs. In a later study focusing on parolees in halfway houses, Latessa and Allen (1982) examined 44 studies with sufficiently rigorous methodology to enable the researchers to draw reasonable assessments of post-release outcomes. As Allen et al. (1976) had found earlier, the results were mixed -- at times showing halfway house residents having lower recidivism rates and at time showing no differences or that halfway house residents did worse on recidivism rates.
Day reporting centers are a more recent correctional option that require offenders who are on pretrial release, probation, or parole to appear at specific location on a frequent and regular basis. Unlike the halfway houses, the day reporting centers are non-residential, offenders are required to report to the centers but they return to their homes to sleep at night. While at the centers they are required to participate in services (treatment, employment search, etc) or activities (urine test, meetings with agents) provided by the agency or other community agencies. These programs are currently being widely implemented in the United States. In 1990, a study by the NIJ found only 13 centers in the United States, by 1995 there were at least 114 centers in 22 states (Parent, Byrne, Tsarfaty, Valdae and Esselman 1995). The centers emphasize both strict surveillance and high levels of treatment and other services to offenders. As with the other intermediate sanctions, there is a tension between providing increased surveillance and increased treatment in the day reporting centers, and centers vary greatly in the emphasis placed on one or the other. While there have been some descriptive studies of day reporting programs, to date there have been no impact evaluations examining the effectiveness of the programs in preventing crime.
5.4 Summary of Community Restraints
A large body of research, much of it funded by NIJ, including random assignment studies, consistently shows the failure of these programs to lower recidivism. Restraining offenders in the community by increasing surveillance and control over their activities does not reduce their criminal activities. In general, they are arrested as often as their counterparts who receive less surveillance. Most research has focused on the restraining aspects of these community programs and not the treatment services delivered to the offenders. That is, the research fails to clearly identify and rigorously examine (from a research perspective) the impact of the therapeutic aspects of the community programs. When the researchers have mentioned the therapeutic integrity of the programs, it is often to note that the anticipated services or staffing did not occur (see for instance, Sontheimer and Goodstein's study or Greenwood et al.'s study of the Skillman aftercare program discussed in the juvenile programs section of this chapter). Questions remain about the impact of additional treatment within a program that increases restraints.
6. STRUCTURE, DISCIPLINE AND CHALLENGE
Correctional boot camps for adults and for juveniles and wilderness programs have been grouped together because they all focus on structure, discipline and physical and/or mental challenge. The experiences of the offenders in the programs is anticipated to change them in a positive way so that their future criminal activities will be reduced. The mechanism for this change is attributed to various factors such as self esteem, or increased bonds with staff and peers. Some also expect that these punitive programs will discourage others from committing crimes or that the individuals who spend time in the programs will be deterred from future criminal activities. At times programs combine therapeutic programming with the structure, discipline and challenge aspects. The studies of the programs focus on the recidivism rates of those who are released from the programs and compare these rates to comparison groups who served different sentences. Thus, the studies examine the specific deterrence or positive change impacts of the programs.
6.1 Boot Camps for Adults
Boot camp prisons, alternatively called shock incarceration, regimented discipline or intensive incarceration, are correctional programs designed to be similar to military basic training. These relatively new programs began in 1983 in Georgia and Oklahoma but rapidly spread throughout the nation (MacKenzie 1990; MacKenzie and Parent 1991). The early programs emphasized the military aspects of discipline, comportment and drill and ceremony.
More recently the programs have changed to include more programming and treatment and many have de-emphasized the military focus of the programs. As has occurred with other correctional options, the boot camp programs vary tremendously when cross-program comparisons are made in type of population served, treatment components, aftercare or follow-up supervision, and emphasis on military drill and ceremony. The majority of the state department of corrections have opened boot camp programs and increasing numbers of programs are being opened for juveniles and for jail inmates.
To date, there have been no random assignment studies examining the effectiveness of boot camp prisons for adult offenders. As shown in Table 9-5, most of this research had limited scientific rigor (scores of 3). Some of the research has made use of statistical controls to adjust for original differences between the boot camp releasees and comparison groups to examine their performance in the community (see for example, MacKenzie et. al 1995). In general, the results show no significant differences in recidivism between offenders who are sent to boot camps when compared to others including those who either served a longer period of time in prison or those who served their sentence on probation (MacKenzie et. al. 1993; MacKenzie and Shaw 1993; Flowers, Carr and Ruback 1991; Florida Department of Corrections 1990). However, in programs where a substantial number of offenders were dismissed from the boot camp prior to completion, the recidivism rates for those who completed the program were significantly lower than the rates for those who were dismissed (MacKenzie et al. 1995; NYDCS and NYDP 1993). Thus, while there is no evidence that the boot camps actually change offenders, there is some indication that the programs can be used to "signal" which offenders will have difficulty completing probation or parole. That is, offenders who remain in the program and complete it are at less risk for recidivism than those who are dismissed (either voluntarily dropping out or for misbehavior).
In a further exploratory analysis examining program differences and recidivism rates, MacKenzie et al. (1995) found some commonality among programs where the boot camp releasees had lower recidivism rates than comparison groups on some but not all, measures of recidivism. In particular, these programs: (1) devoted more than three hours per day to therapeutic activities such as therapy, counseling, drug treatment and education; (2) there was some type of follow-up for the offenders in the community after they left the boot camp; and (3) participants had to volunteer for the program.
The following conclusions can be drawn from the available research examining boot camps for adult offenders:
o The military atmosphere, structure and discipline of correctional boot camps does not significantly reduce the recidivism of releasees in comparison to offenders serving time on parole or probation;
o In programs where a substantial number of offenders are dismissed from the boot camp, the recidivism rates for those who complete the boot camp are significantly lower than the rates for those who were dismissed;
o Exploratory analyses suggest that programs incorporating components such as therapeutic activities during the boot camp and follow-up in the community may be successful in reducing recidivism but this conclusion is tentative until more research is completed.
6.2 Boot Camps for Juveniles
Recently, four random assignment studies have been completed examining the recidivism of juveniles released from boot camps. With cooperative funding from The NIJ and the Office of Juvenile Justice and Delinquency Prevention (OJJDP) a carefully designed experimental study examining boot camps in three sites was completed. Funding was provided for sites to develop innovative demonstration programs, if they were willing to permit research to randomly assign juveniles to the boot camps or some alternative. Three sites were selected to participate. The research results are considered preliminary because data collection continued after these analyses were completed.
The fourth study of juvenile boot camps is being conducted by The California Youth Authority (CYA). This study used random assignment to evaluate the effectiveness of the CYA juvenile boot camp. The results are considered preliminary at this point because the juveniles have been in the community for only a short period of time.
As shown in Table 9-6, all of these studies are judged to be rigorous (methods score of 5). The results from three of the studies reveal no significant differences in recidivism between the boot camp youth and the control groups. In the fourth site, the CYA, more of the boot camp youth were reincarcerated than the control youth. Obviously, this presents little support for these boot camps as crime prevention techniques.
Table 9-5. Studies of correctional boot camps showing scientific methods score and findings.
Study Scientific Findings
Methods Score
N.Y. Department of Male BC graduates reincarcerated
Correctional Services and 2 less for
N.Y. Division of Parole new crimes and parole violations
(1992, 1993) compared
to: (1) parolees sentenced before
the program began; (2) those who
refused to enter; (3) dismissals
from the program.
N.Y. Department of Female BC graduates reincarcerated
Correctional Services and 2 less for
N.Y. Division of Parole new crimes and parole violations
(1993) compared
to: (1) parolees sentenced before
the
program began; (2) those who refused
to enter; (3) Dismissals from the
program.
Compared with those sentenced to
Flowers and Ruback (1991) 3 various alternatives, male BC
graduates were reincarcerated less
often. But when graduates where
compared with groups most similar to
BC graduates the reincarceration
rates were higher for BC graduates.
BC graduates had fewer arrests and
reconvictions for new crimes when
MacKenzie (1991) 3 compared to samples of parolees and
probationers but more arrests for
technical violations. Those
dismissed from the program had fewer
arrests than graduates but were the
same in reconviction rates.
Texas Department of 2 BC releasees were reincarcerated
Criminal Justice (1991) more than
parolees, ISP and restitution
center releasees.
Compared to probationers BC
releasees had fewer rearrests in one
MacKenzie et al (1995) 4 state and more in two states; fewer
revocations in three states, NS.
Compared to parolees BC releases had
fewer rearrests in four states;
fewer revocations in five states and
more revocations in one state NS.
Florida Department of There were no differences between BC
Corrections (1990) 2 graduates and prison releasees in
new felonies. BC graduates had more
new misdemeanors but fewer technical
violations.
Note: NS=nonsignificant, S=significant
Table 9-6. Studies of juvenile boot camps and recidivism showing scientific methods score and findings.
Study Scientific Findings
Methods Score
Peters et al. (1996a) 5 More boot camp youths (38.8%)
recidivated than control group
(35.5%), NS.
Peters et al. (1996b) 5 Fewer boot camp youths (28.1%)
recidivated than control group
(31%), NS.
Peters et al. (1996c) 5 More boot camp youths (71.8%)
recidivated than control group
(50%), S.
Bottcher et al. (1996) 5 More boot camp youths (77.7%)
rearrested than control group
(77.1%), NS.
Note: NS=nonsignificant, S=significant
6.3 Summary
The boot camps reviewed in this section do not, as a whole, appear to be good candidates for crime prevention. In general, findings indicate no difference between the offenders who participated and those who did not. There was some suggestion in the research examining adult boot camps that enhanced therapeutic programming within the boot camps may have had an impact on reducing recidivism, but the research is exploratory and did not use a strong methodology. The juvenile programs appeared less hopeful. Several questions remain. First, more information is needed about the therapeutic integrity of the programs and how the programs compare to the alternatives where the control groups spent time. Possibly, the failure to find differences in recidivism may be because the control groups were receiving enhanced treatment while the juveniles in the boot camps were spending time on physical activities. Such physical activities may have health benefits but they may not address the criminogenic needs of these offenders. Questions remain about how rehabilitation can be combined with these programs and whether this would enhance or conversely reduce the effectiveness of the rehabilitation.
7. JUVENILE OFFENDERS
7.1 Treatment Programs for Juveniles
Rehabilitation has particular appeal for use with juveniles. Juvenile crime is often serious and it may represent a large proportion of the total criminal activity in a community. However, it is usually assumed that adolescents deserve and require special handling because at this stage of life they are in a formative period and criminal behavior at this stage will not necessarily be continued into adulthood. Theoretically, rehabilitation has been the focus of correctional programs for juveniles. However, in practice, as occurs with adult programs, juvenile programs are generally poorly implemented. Juveniles have a potentially long adulthood in front of them, therefore, strategies that reduce the future criminal activities of juveniles are also particularly important. An effective preventive intervention at an early age, that results in reduced criminality over a lifetime, can have a substantial payoff.
The most extensive meta-analysis examining the effectiveness of delinquency outcome studies was conducted by Lipsey (1992 ). In a meta-analysis of juvenile delinquency programs, Lipsey examined the effectiveness of 443 different research studies.6 This meta-analysis improved on previous reviews of delinquency treatment research by: (1) broadening the coverage of the literature through an exhaustive search for relevant studies; and, (2) coding sufficient detail from each eligible study.
Among other criteria, the studies included in Lipsey's analysis were those that provided some intervention or treatment that had as its aim the reduction, prevention, treatment, or remediation of delinquency or antisocial behavior problems similar to delinquency. Delinquency was defined as behavior chargeable under applicable laws. Studies were included in the analysis only if the majority of the subjects were between the ages of twelve and twenty one.
Findings from the Lipsey (1992) study revealed that overall in 64.3 percent of the studies the treatment group did better (in most cases this refers to a reduction in recidivism) than the control group. The mean effect size for the studies was .172 which was comparable to previous meta-analyses of more highly selected sets of studies. One way to understand this effect size is to translate it into a comparison to a baseline of 50 percent. This effect size is equivalent to an average reduction in recidivism from 50 to 45 percent. That is, considering all treatment program studies combined, 45 percent of those who received treatment would be expected to recidivate in comparison to 50 percent of the non-treated control group.
In more detailed analyses, Lipsey worked to identify those characteristics most important in determining differences among treatment and control groups. Table 7 shows the estimated recidivism rates for the treated group if 50 percent of the control group had recidivated. He examined both methodological aspects of the study (sample size, equivalence of groups, attrition, outcome measures, etc.) and treatment aspects (subjects, dosage, treatment modality, philosophy). As expected, differences in study methodology were associated with effect sizes. More important for purposes here are the findings for the treatment effects (once the methodology effects are controlled). As shown in Table 7, the more effective programs were predicted to reduce recidivism substantially. For instance, in comparison to 50 percent recidivism rate for the control group, only 32 to 38 percent of the juveniles who were given employment, multi-modal and behavior programs were estimated to recidivate.
Overall, the results of the meta-analysis indicate that more effective programs were:
o Judged to provide larger amounts of meaningful contact (treatment integrity) and were longer in duration (more dosage);
o Provided by the researcher or in situations where the researcher was influential in the treatment setting;
o Behavioral, skill-oriented and multi-modal treatment.
There was also some evidence that more effective programs targeted higher risk juveniles but this effect was small and nonsignificant. On the other hand, treatment in public facilities, custodial institutions, and in the juvenile justice system were less effective than other alternatives. This suggests that treatment provided in community settings may be more effective. However, Lipsey cautions that this conclusion is confounded with dosage (intensity) and needs a more refined breakdown before definite conclusions can be drawn.
It is interesting that effective programs were those that were either provided by the researcher or where the researcher was influential in the treatment setting. This may indicate that treatment delivered or administered by the researcher was better implemented than typical programs.
Table 9-7 shows his results for individual treatment modalities. Lipsey cautions the reader to interpret the individual categories carefully because crude descriptions of treatment programs in the studies as well as the multiple elements in some programs made coding extremely difficult. He suggests instead that the reader examine the broader patterns of the rankings of treatment modalities. From this perspective, the more structured and focused treatment (e.g., behavioral, skill-oriented)7 and multi-modal treatments8 seem to be more effective than the less structured and focused approaches (e.g., counseling). Interestingly, while programs that emphasized employment are near the top in effectiveness, vocational treatment programs were associated with increased recidivism for the treated group. The reason for this is difficult to determine. Possibly the employment programs were more directly related to skills needed to find and keep jobs while the vocational training programs were school based and less directly associated with obtaining employment.
The best treatment types show delinquency effects of meaningful practical magnitude, in the range of 10 to 20 percentage points reduction in recidivism. On the other hand, there is no evidence that programs emphasizing deterrence treatments are effective and, in fact, such programs were estimated to increase recidivism (e.g., 62 percent of those who received a deterrence program were estimated to recidivate in comparison to 50 percent of the controls.
In comparing his results to the earlier findings by Andrews et al. (1990), Lipsey asserts that with few exceptions the largest effect sizes occurred for treatment that would be classified by Andrews et al. as clinically relevant. Similarly, as found by Andrews et al., deterrence treatments were associated with negative effects (e.g., an increase in recidivism). Few studies of interventions deal exclusively with the most serious or most violent juvenile offenders so, at this point, little can be said about the effectiveness of programs for these offenders.
Table 9-7. Effect size estimates for different treatment modalities after controlling for study methodology for juvenile delinquency treatment programs (Lipsey, 1992).
Treatment Modality Estimated Recidivism
Treated Group/Control Group
Employment 32/50
Multi-modal 38/50
Behavioral 38/50
Institutional, other 40/50
Skill-oriented 40/50
Community residential 42/50
Any other juvenile justice 43/50
Probation/Parole, release 45/50
Probation/Parole, reduced caseload 46/50
Probation/Parole, restitution 46/50
Individual counseling 46/50
Group counseling 47/50
Probation/Parole, other enhancement 47/50
Family counseling 49/50
Vocational 59/50
Deterrence programs 62/50
7.2 Juvenile Residential Programs
One type of program particularly popular during the late 1970s and early 1980s was the wilderness or outward bound-type programs. These programs emphasize physical challenge and demand that individuals excel beyond what they feel they can do. Winterdyk and Roesch report that they found well over one hundred wilderness programs for treating delinquent youths in North America in the early 1980s. Outcome evaluations have been extremely rare (Gendreau and Ross 1987). Recently, several other wilderness-type programs have been studied. The results are shown in Table 8. All of these programs consider themselves wilderness programs. Perhaps the most frequently cited study of this type of program in the VisionQuest study by Greenwood and Turner (1987). They examined the behavior of the juveniles during the six to 18 months after release from the program (controlling for prior arrests). Youth from VisionQuest had fewer rearrests than youth who had served time in a probation camp or who had refused to accept the VisionQuest placement and were placed in other programs. While the results appear positive, as noted on the table the research methodology makes it impossible to draw conclusions regarding the program's effectiveness.
In a more recent study, Deschenes, Greenwood and Marshall (1996) examined the Nokomis Challenge Program in the Michigan Department of Social Services. Nokomis was designed as an intensive treatment program for low to medium risk juveniles. The focus of the program was on relapse prevention. Male youth were expected to spend less time in the residential facility but a longer time in community treatment when compared with youth in the training schools. Findings (see Table 9-8) indicated that the Nokomis youth had more felony arrests after release than did the comparison (significant). It is important to note that the examination of the implementation of the program revealed that the aftercare phase of the program failed to provide many of the expected treatment programs. There was limited substance abuse treatment and control group youth had more family counseling than the treatment group.
Castellano and Soderstrom completed a study of the Spectrum program in Illinois. This wilderness program was modeled after outward bound. The thirty day course focuses on teaching wilderness survival and group living skills to pre-delinquent and delinquent juveniles. A comparison of recidivism rates indicated that 75 percent of the Spectrum participants were rearrested in the follow-up period compared with 62.6 percent of the matched comparison group (nonsignificant).
In a random assignment study, RAND researchers examined the effectiveness of the Paint Creek Youth Center (PCYC) in southern Ohio (Greenwood and Turner 1993). The program targeted youth convicted of serious felonies who were required to spend an average of almost a year in residential treatment. While the program was located in a rural setting, it would not be classified as a wilderness or challenge program because these activities were not a major component of the program. The distinguishing features of the PCYC were: small size, problem oriented focus, cognitive/behavioral methods, family group therapy and intensive community reintegration and aftercare. Youth were randomly assigned to either the PCYC or regular training schools. Their behavior in the community after release was compared. The design was weakened because a relatively large number of the youth (25 percent) were removed from the PCYC and sent to the training schools to serve the remainder of their term. Furthermore, 27 percent of the remaining youth did not complete all three phases of the residential program. Official records of recidivism indicated that 50.7 percent of the PCYC youth (including those who were removed) and 61.3 percent of the control group had been arrested during a one-year follow-up. The difference was nonsignificant. The small numbers of offenders in the study limits the power to detect differences between groups. This along with the loss of 25 percent of the PCYC youth makes it difficult to draw any definitive conclusions from the research.
Overall, these studies of juvenile residential programs had very mixed results. Although several of the studies were well designed, problems with the small number of subjects, attrition and program implementation limit the conclusions that can be drawn about the effectiveness of the programs in preventing crime. The one program that included both a strong research design and a reduction in recidivism, although this difference was not significant, was Paint Creek. Interestingly, this program followed many of the principles proposed by Andrews et al. (1990). High risk youth were targeted for participation an the intensive program that used a cognitive/behavioral mode of treatment. However, problems with the research design severely limited the potential for detecting differences even if the program had indeed been effective. Most notably, the focus of the program was not on wilderness or challenge activities.
The other programs reviewed in this section either targeted individuals who were lower risks for recidivism (Nokomos, Spectrum), were of short duration (Spectrum), were less behavioral in treatment philosophy, or focused on non-criminogenic factors such as physical challenge (Spectrum). Thus, from the perspective of the research on rehabilitation (see section on rehabilitation and the Andrews et al. 1990 study), we would not expect them to be effective in reducing future criminal behavior.
Table 9-8. Studies of youth residential programs showing scientific methods score and findings.
Study Scientific Findings
Methods Score
Greenwood and Turner 2 VisionQuest (39%) fewer arrests than
(1987) YCC Control (71%), S.
Deschenes et al (1996) 3 Nokomis group (48%) had more arrests
than control (23%), S.
Greenwood and Turner 3 Paint Creek youth had fewer
(1993) official arrests (51%) than control
youth (61%), NS.
Paint Creek youth self-reported more
serious offenses (75%) than control
(62%), NS.
Castellano and Soderstrom 2 Spectrum youth did not differ from
(1992) control youth in recidivism, NS.
Note: NS=nonsignificant, S=significant
7.3 Community Supervision and Aftercare for Juveniles
Approximately 55 percent of adjudicated juveniles are given probation (Butts, Snyder, and Finnegan 1994). Furthermore, those knowledgeable about juvenile corrections increasingly argue for aftercare and transitional services for juveniles following a period of incarceration. Both of the recent meta-analyses (e.g., Andrews et al. 1990; Lipsey 1992) suggest there will be greater reductions in recidivism if treatment is provided in community settings instead of in institutions. However, national surveys of intensive supervision and aftercare programs for juveniles completed during the 1980s revealed that few programs had been evaluated (Armstrong, 1988; Krisberg et al., 1989). Additionally, the evaluations that had been completed were severely limited in scientific rigor. Two exceptions to this were the New Pride Study sponsored by NIJ and OJJDP (NIJ, 1985) and The Violent Juvenile Offender Study implemented by OJJDP (Fagan et al., 1988). However, in neither study did the group who received the additional aftercare or supervision have significantly lower recidivism rates.
Most recent studies of community programs have focused on the increased surveillance and restraint aspects of the programs and not the enhanced services. While some of the programs enhance services, the research is designed to compare the increased surveillance and restraint with or without increased services. For this reason, the research is included in this section on community restraints instead of the rehabilitation section. The treatment and restraint components cannot be untangled, and since the research designs focus on surveillance the outcomes are more indicative of the effectiveness of restraints than rehabilitation. Additionally, when the treatment integrity is examined, few differences are found between the experimental program and the control in either the services delivered or the impact on risk factors. For example, in the Greenwood et al. (1993) study described below researchers did not find that the aftercare program they studied had an effect on the targeted risk factors. This section examines some of the recent studies of increased restraints over juveniles in the community (see Table 9-9).
Using a random assignment design, Land, McCall and Williams (1990) examined the North Carolina Court Counselors' Intensive Protective Supervision Project (ISP). The majority of the subjects were status offenders who entered the program as runaways or truants. The program was designed to enhance both the surveillance and services provided to the juveniles. As shown in Table 9-9, the results indicated that youth with no prior delinquent offenses had fewer delinquent offenses compared to the control group (11.9 percent to 27.5 percent) but the ISP youth with prior delinquent offenses had more delinquent offenses (57.1 percent compared to 33.3 percent). However, there were only a small number of youth with prior delinquencies. Whether the results were the effect of surveillance or services could not be distinguished in the research design.
In another study of youth in the community, Wiebush (1993) compared the performance of youth on intensive supervision (ISP) with comparison groups of youth on probation and parole. During the 18 month follow-up, a larger percent of the ISP youth received felony complaints (50.6 percent) than the probationers (37.9) but fewer felonies than the parolees (56.6 percent). Similarly, a larger percent of the ISP group were adjudicated (76.5) when compared to the probationers (61.6 percent) but fewer than the parolees (77.6). The results were not significant. Furthermore, it is difficult to draw conclusions because the groups were not randomly assigned and the groups differed prior to the treatment.
Sontheimer and Goodstein (1993) examined whether a juvenile intensive supervision program (ISP) in Pennsylvania had an impact on juvenile's propensity to reoffend (a rehabilitative or deterrent effect) or whether the restraints provided by the officers limited the opportunity juveniles had to reoffend. The program was an intensive aftercare program for serious juvenile offenders. Probationer officers supervising juveniles in the aftercare program were required to have frequent contacts with the juveniles and significant others; however, other than these additional contacts, there was no statement of the mission or philosophy of the program. Significantly fewer of the experimental group were rearrested (50 percent) than the controls (74 percent) and their mean number of rearrests were fewer (1.02 compared to 2.07 for the controls). The researchers interpret their findings as support for the restraining effect of ISP and not necessarily a reduction in the criminal propensity of the juveniles.
There were problems with the implementation of the program studied by Sontheimer and Goodstein. Contacts were substantially less than the mandated number and there was a large turnover of staff. This turnover would be expected to create turmoil for participants and uneven staff training and accountability. This combined with the failure to clearly define the mission of the program led the researchers to question whether the results were indicative of problems in the implementation of the treatment components of the program and not what could be achieved in such programs.
Minor and Elrod (1990,1992) examined the impact of an enhanced treatment program for juveniles on intensive and moderate levels of supervision. While there were no significant differences between the groups, juveniles in the enhanced treatment ISP program had more criminal offense complaints than the juveniles on ISP but without the enhanced treatment. Follow-up analyses also indicated that the intervention did have an effect on those who had more lengthy criminal backgrounds (e.g., the higher risk group). The major problem with this research was the number of subjects was so small there was no power to detect any difference that might have existed.
Greenwood, Deschenes and Adams (1993) studied the Skillman aftercare program in Michigan and Pennsylvania. The programs were designed to provide treatment components, hence the term "aftercare", along with intensive supervision. Subjects were randomly assigned to either the aftercare ISP or the control. Results indicated no significant difference between the experimental and control groups in the proportion of the youth who were rearrested, or who self-reported either offending or drug use in a one year follow-up period. However, an examination of what the programs provided for the youth indicated that in comparison to the control group the aftercare group: did not participate more in education or work activities; had little family support; and did not associate less with delinquent peers. Thus, despite the fact that the program was designed to promote changes in these risk factors there was little evidence of such change. Consistent with the previous meta-analyses of rehabilitation, it appears that the program did not have the required "treatment integrity" to bring about the changes in the risk (criminogenic) factors associated with criminal behavior.
The above studies compared the ISP programs in the community to other community alternatives. The following studies were designed to compare the recidivism of those who spent time on community supervision with others who had spent time in training schools.
Barton and Butts (1990) evaluated an in-home ISP program compared to commitment to traditional training schools in a random assignment study. They found that the ISP groups had a higher mean number of charges but the mean seriousness of the charges was greater for the control group. These differences were not significant when time in the community was controlled in the statistical analysis.
Gottfredson and Barton (1993) used a nonequivalent comparison group design to compare the effect of the closing of a juvenile training facility to the performance of juveniles who were then managed in the community. They found that the juveniles who had spent time in the institution had significantly lower recidivism rates than the comparison group. It is important to note that the comparison group was not intensively supervised in the community and there is little information about what services they may have received in the community.
In summary, most of the results reveal no significant difference between the experimental condition and the controls. In part, this reflects the small number of subjects in the studies so there is little power to detect any differences that might exist between the groups. The two studies by Land et al. (1990) and Sontheimer and Goodstein (1993) did find lower recidivism rates for the experimental groups. In both cases it appears that the experimental group received more services than the comparison. Again, this suggests the importance of meeting the rehabilitative needs of such offenders. This may also be why the institutionalized juveniles in the Gottfredson and Barton (1993) study had lower recidivism rates -- because of the services and rehabilitation they received when they were in the institution. Whether or not the juvenile is in a facility or on ISP may not be as important as whether appropriate rehabilitation programs are a part of the correctional option.
Table 9-9. Studies of juvenile community supervision and recidivism showing scientific methods score and findings.
Study Scientific Findings
Methods Score
Land et al (1990) 5 ISP youth (mostly status offenders)
with no prior delinquent offenses
had fewer delinquent offenses (12%)
than control group (28%), S.
ISP youth with prior delinquent
offenses had more delinquent
offenses (57%) than control group
(33%), NS.
Weibush (1993) 3 ISP youth had more felony
complaints (51%) than probationers
(38%) but fewer than parolees
(57%), NS.
ISP youth had more adjudications
(77%) than probationers (62%) but
fewer than parolees (78%), NS.
Sontheimer and Goodstein 5 ISP juveniles had fewer rearrests
(1993) (50%) than parolees (74%), S.
Minor and Elrod (1990) 2 ISP group had more self-reported
criminal and status offenses, NS.
Elrod and Minor (1992) 2 ISP group had fewer status offenses
but more criminal offenses (68%)
than control group (67%), NS.
Barton and Butts (1990) 5 ISP juveniles had more charges but
control group had more serious
charges, NS.
Greenwood et al (1993) 5 Detroit: Aftercare group (22%) had
more arrests than controls (18%),
NS.
Pittsburgh: Aftercare group
had fewer arrests 49%) compared to
controls (48%), NS.
Gottfredson and Barton 4 Institutionalized juveniles had
(1993) fewer arrests than
non-institutionalized, S.
Note: NS=nonsignificant, S=significant
8. DRUG-INVOLVED OFFENDERS
8.1 Treatment in Prison
Advocates of treatment and rehabilitation have perhaps made the strongest arguments in favor of increased treatment for substance abusing offenders. The need for treatment is demonstrated by the large body of research indicating the relationship between criminal activity and use of alcohol and other drugs (Chaiken 1986; Chaiken and Chaiken 1982; Inciardi 1979; Johnson and Wish 1986; Nurco, Kinlock and Hanlon, 1990; Speckart and Anglin, 1986). Furthermore, the National Institute of Justice (NIJ)'s Drug Use Forecasting (DUF) program consistently finds high rates of illicit drug use among arrestees in the 24 participating cities. In 1995, between 47 and 78 percent of the men and 44 to 85 percent of the women test positive for use of illegal drugs.
Documentation of this high level of use and the strong association between drug use and crime clearly indicates the critical need for treatment for these offenders. However, the recent focus of criminal justice policies on incapacitation and deterrence did not easily provide the necessary funds. In 1987, approximately 11 percent of prison inmates were involved in some type of drug treatment (Chaiken 1989). Although the numbers are sizeable (51,500), the majority of inmates with substance abuse problems still do not receive treatment while in prison. In 1991, 48 percent of state prisoners and 43 percent of the Federal prisoners reported that they had been in a drug program since admission to prison (BJS 1994). Yet the intensity and quality of these treatment programs is unknown.
Despite the fact that many drug-involved offenders are not treated while they are under the control of the criminal justice system, a growing body of research indicates that treatment for substance-involved offenders can effectively reduce substance use and criminal recidivism (Gerstein and Harwood 1992). Effectiveness of drug treatment is directly related to the length of time an individual remains in treatment. This is true for various treatment modalities. Furthermore, the treatment is effective whether the offender enters voluntarily or under some form of coercion (Anglin and Hser 1990a, b; Anglin and Maugh 1992; Falkin, Wexler, and Lipton 1992; Leukefeld and Tims 1992; Travis, Wetherington, Feucht, and Visher 1996). From this perspective, the criminal justice system presents an ideal opportunity to require offenders to participate and remain in treatment.
Some of the most promising evaluations of drug treatment for criminal justice have focused on the effectiveness of prison-based therapeutic communities (TCs) that operate as 24-hour live-in facilities within the prison. We examined evaluations of five such programs (Wexler et al. 1992; Martin et al. 1995; Wexler et al. 1995; Field 1989; Eisenberg and Fabelo 1996). As shown in Table 9-10, the studies were judged to be of sufficient rigor to draw conclusions about the effectiveness of the treatment programs. In all studies, the researchers found that the graduates of the programs had lower recidivism rates than those who spent less time in the programs. They concluded that the programs were effective in reducing recidivism of the offenders (See Table 9-10).
Although these studies are widely cited as evidence of the success of drug treatment, there are some concerns about the research methodology. In particular, the programs have high attrition rates (large numbers of offenders leave the programs before completing). Many of the research designs do not take this attrition into account in the data analysis. As a result, it is difficult to conclude that the programs are effective in reducing recidivism. The alternative conclusion is that the programs identify those who are ready to change and these will be the individuals who will be successful in the community. That is, the research design does not permit one to draw conclusions about the effect of the program because offenders who complete the TC program may be very different from those who do not, a difference which could have existed prior to the program. More work needs to be done to examine program attrition and how programs can be carefully matched to the needs of offenders so that larger numbers will complete the program. It may also be possible to use the power of the criminal justice system to coerce offenders to remain in programs.
In summary research examining the effectiveness of drug treatment shows:
o Drug treatment is effective is reducing the recidivism of offenders;
o Offenders coerced into treatment by the criminal justice system do as well as those who enter voluntarily.
Dropouts from treatment