6
Conclusions, Recommendation, and Research Agenda

Two major conclusions emerge from our review of research on parole and desistance from crime. The first is that desistance from crime varies widely among parolees. Released prisoners with lengthy criminal records and who have been to prison several times before have very high recidivism rates—over 80 percent are rearrested within three years of release from prison. In contrast, less than half of first-time releasees and older releasees are rearrested within three years of their release (Langan and Levin, 2002; Rosenfeld et al., 2005; Solomon, Kachnowski, and Bhati, 2005). Indeed, when it comes to desistance or recidivism, there is no such thing as the “average” parolee. In a word, the parolee population is heterogeneous. It follows that the types of services, sanctions, and supervision strategies effective in increasing desistance among some groups of parolees may not be effective for other groups.

The second conclusion that emerges from our review of research on recidivism and desistance concerns intervention effects. We define “intervention” broadly to include both the routine functions of the criminal justice system (e.g., parole supervision) and the characteristics of discrete programs and treatments (e.g., drug abuse treatment). With some exceptions, the characteristics of interventions, including parole supervision itself, that are effective in increasing parolees’ desistance from crime are unknown. This is not the same as saying that “nothing works” in reducing recidivism or increasing desistance (Farabee, 2005; Martinson, 1974); existing research knowledge is too thin to support that strong conclusion. In Petersilia’s (2004) review of the prisoner reentry programs, she estimates that less than



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 72
6 Conclusions, Recommendation, and Research Agenda Two major conclusions emerge from our review of research on parole and desistance from crime. The first is that desistance from crime varies widely among parolees. Released prisoners with lengthy criminal records and who have been to prison several times before have very high recidivism rates—over 80 percent are rearrested within three years of release from prison. In contrast, less than half of first-time releasees and older releasees are rearrested within three years of their release (Langan and Levin, 2002; Rosenfeld et al., 2005; Solomon, Kachnowski, and Bhati, 2005). Indeed, when it comes to desistance or recidivism, there is no such thing as the “average” parolee. In a word, the parolee population is heterogeneous. It follows that the types of services, sanctions, and supervision strategies ef- fective in increasing desistance among some groups of parolees may not be effective for other groups. The second conclusion that emerges from our review of research on re- cidivism and desistance concerns intervention effects. We define “interven- tion” broadly to include both the routine functions of the criminal justice system (e.g., parole supervision) and the characteristics of discrete programs and treatments (e.g., drug abuse treatment). With some exceptions, the characteristics of interventions, including parole supervision itself, that are effective in increasing parolees’ desistance from crime are unknown. This is not the same as saying that “nothing works” in reducing recidivism or increasing desistance (Farabee, 2005; Martinson, 1974); existing research knowledge is too thin to support that strong conclusion. In Petersilia’s (2004) review of the prisoner reentry programs, she estimates that less than 

OCR for page 72
 CONCLUSIONS, RECOMMENDATION, AND RESEARCH AGENDA 1 percent of all prisoner reentry programs implemented in the United States in the last decade have been subject to a formal evaluation, and the vast majority of those did not use a randomized experimental research design. She writes that “using this ‘body’ of research to conclude anything about which reentry programs ‘work’ or ‘don’t work’ seems misguided” (2004, p. 7). A major impediment to knowledge about “what works” in increasing desistance is poor program implementation. Without proper implementa- tion, as well as careful evaluation, one cannot determine whether a given program succeeds or fails in its conception, design, or operation. These two themes of parolee heterogeneity and intervention effects frame our summary of what is known and what needs to be learned about the characteristics of parolees and of the programs and interventions in- tended to increase their desistance from crime. In addition to our summary of the research findings on parolees and desistance programs and our proposed agenda for future research on parolees and their desistance from crime, we offer a policy recommendation that is driven by the research findings. WHAT IS KNOWN ABOUT PAROLEES AND DESISTANCE We use a comparatively permissive criterion for classifying a given research finding on parolees and programs as established or settled “knowl- edge.” We include a research result in the category of “known” if it has been replicated across several studies and is not (or not any longer) subject to widespread dispute in the research community. A more restrictive crite- rion, for example, that any of the studies producing the result must meet the rigorous requirements of experimental science, would yield a much leaner knowledge base on the characteristics of parolees and effective programs. The difference is analogous to that between the “preponder- ance of evidence” and “beyond a reasonable doubt” evidentiary standard in jurisprudence. We adopt the former for organizing the extant research on parolees and desistance; however, for future research, we propose that, when feasible, it should be conducted and interpreted according to more rigorous standards of proof. The need for more rigorous research methods in evaluating both pre- release and postrelease programs is beyond dispute, but the use of random designs does raise ethical questions in an environment that combines in- tervention and social control objectives. For example, a positive drug test typically triggers a sanction in most jurisdictions. Are treatments more re- strictive or likelier to result in official sanctions than baseline parole condi- tions? These kinds of issues need to be thought out carefully in the design of experimental research. However, research suggests that ethical randomized designs are possible, especially where there is a standard program that can

OCR for page 72
 PAROLE, DESISTANCE FROM CRIME, AND COMMUNITY INTEGRATION serve as a control for other randomized treatment groups, that is, in a situ- ation where no one would get less than the standard postrelease treatment (see, for example, National Research Council, 2001). Heterogeneity in the Parole Population Recidivism rates, defined as the probability that parolees are rearrested or returned to prison, are significantly different for different groups of parolees. They are lower for women than for men; lower for older than younger parolees; lower for people with relatively short criminal records; and lower for violent offenders than for property or drug offenders (Langan and Levin, 2002; Petersilia, 2003). Black parolees have higher recidivism rates than white parolees for violent and property crime, but not for drug crimes (Rosenfeld et al., 2005; see, also, Langan and Levin, 2002; Solomon et al., 2005). We note, how- ever, that the race difference in recidivism is smaller than the race difference in overall arrest or imprisonment rates. Parolees released from prison for the first time have lower recidivism rates than those who have been released in the past and then returned to prison. This finding holds even when sex, age, race, criminal record, offense type, and other characteristics of parolees are controlled (Rosenfeld et al., 2005; Tonry, 2004). The cause of this difference has not been established, however. Selection may play a major role; past failure at reentry predicts future failure. It also is possible that parole authorities and the police supervise and watch “two-time losers” more closely or are less willing to overlook any violations of their parole contracts. The finding that past im- prisonment predicts future rearrest and imprisonment is consistent with the idea that the prison experience itself is criminogenic, but, recidivism does not appear to be related to the length of time an individual spends in prison (Rosenfeld et al., 2005). Another possibility is that people who have been imprisoned multiple times possess unmeasured traits or deficits that impede desistance. At present, the simple conclusion one can draw from what is known is that past recidivism predicts future recidivism. One of the most significant findings that emerges from our work is that the peak rates for recidivism occur in the days and weeks immedi- ately following release. Arrest rates decline over time after release from prison, especially for property and drug crimes. Moreover, death rates for new releasees—within the first days and weeks—are much higher than for matched demographic groups in the general population. This is a new re- search finding, the importance of which is underscored by the fact that the causes of death for parolees and inmates are different. For the state prison population, the leading causes of death are disease related: cardiovascular disease, cancer, liver diseases, and AIDS-related illnesses. For the releasee

OCR for page 72
 CONCLUSIONS, RECOMMENDATION, AND RESEARCH AGENDA population in the state of Washington, by contrast, the four leading causes of death were drug overdose, cardiovascular disease, homicide, and suicide. Two of these, homicide and drug overdose, are directly related to risky be- haviors, and all may be preventable if close attention and intensive services are given to these releasees at the time of release. Parolees are characterized by a range of deficits in a number of areas. Large fractions of them have educational and cognitive deficits, substance abuse and mental health problems, inadequate housing, and difficulties in finding and keeping a job (Petersilia, 2003; Travis, 2005). A clear need ex- ists for appropriate support services and treatment for people reentering the community from prison with these deficits. Adequate research on specific elements such as the nature, timing, and dosage of services has not been conducted. Research on this population and on the effects of such inter- ventions is the only way to establish whether the absence or inadequacy of services for released prisoners is causally related to recidivism. Formal and Informal Controls The limited research that has been done shows that formal parole su- pervision has only a small effect on recidivism. However, we again must caution against drawing a broad conclusion on the basis of existing re- search, which is methodologically weak and masks large differences in local variations in supervision and services received by parolees. The effect of parole on recidivism appears to be a function of selection of prisoners for release rather than supervision in the community. Controlling for sex, race, age, criminal history, and other factors, parolees released through a discretionary process have a lower recidivism rate than those subject to mandatory release—even though both groups experience generally the same conditions of supervision in the community (Rosenfeld et al., 2005). The ef- fects of parole supervision, however, differ for different groups of parolees. Parole supervision appears to reduce the recidivism rates of parolees who are comparatively low risk (e.g., women and parolees with shorter criminal records), but has little effect on the recidivism rate of higher risk parolees (Solomon, 2005, 2006). Informal social controls, such as marriage and work, are more effec- tive than formal social controls, such as parole supervision and rearrest, in increasing desistance from crime in ways that are generally similar across crime types. Comparatively strong evidence exists regarding the causal ef- fect on criminal behavior of informal social control, especially marriage. Married men are less likely than unmarried men to commit crimes, and recent research has extended this finding to women (King et al., 2007; Griffin and Armstrong, 2003) The effect of marriage on criminal behavior persists even when the traits that predispose men to marry are controlled

OCR for page 72
6 PAROLE, DESISTANCE FROM CRIME, AND COMMUNITY INTEGRATION (Laub and Sampson, 2003; Sampson et al., 2006; King et al., 2007). The marriage findings for females are more ambiguous. Whether the marriage effect on criminal behavior applies specifically to parolees is not known. Intervention Effects Several kinds of intervention programs have been carried out and studied enough for some conclusions to be drawn, although, as noted above, the quality of implementation in these programs is often wanting. The research does show that the effectiveness of interventions to increase desistance from crime depends heavily on implementation characteristics, including staff quality and training, program length and intensity, and or- ganizational readiness. Moreover, few successful interventions have been “manualized.” The effects of in-prison programs on recidivism are rather small. In- prison programs have larger effects on recidivism when coupled with postrelease community-based programs. Among psychological therapeutic approaches for reducing criminal behavior, cognitive-behavioral therapeutic approaches are more effective than other approaches in reducing recidi- vism. For substance abuse, treatment appears to reduce criminal behavior, at least during the period a person is in treatment. However, it is not clear whether this result applies to parolees specifically. It is clear that treatment for substance abuse is more effective in reducing recidivism in combina- tion with criminal justice supervision than either treatment or supervision alone. Criminal offenders under legal pressure to undergo substance abuse treatment have higher attendance rates and remain in treatment longer than those entering treatment voluntarily (National Institute on Drug Abuse, 2006). RESEARCH AGENDA The research literature raises more questions than it answers about the characteristics of parolees and the effects of interventions on desistance. For that reason, our list of questions to guide future research is longer than our list of research findings. The topics are not in order of priority. As with our description of what is known, we divide the proposed research agenda into questions about the heterogeneity of parolees and those on the effects of interventions in reducing recidivism and increasing desistance from crime. We have not aimed for exhaustiveness in framing the agenda, but rather have emphasized those questions that arise most directly from the existing research. Nonetheless, answers to these questions would greatly enhance knowledge about desistance from crime and the charac- teristics of interventions that increase desistance. In addition, continuing

OCR for page 72
 CONCLUSIONS, RECOMMENDATION, AND RESEARCH AGENDA research is needed on how to develop and standardize measures of desis- tance, a more complex concept than recidivism. The committee believes that research on community supervision and desistance from crime should constitute a major research priority of the National Institute of Justice and of private organizations that fund criminal justice research. Understanding Parole Heterogeneity Early Failure Early failure is a high research priority if recidivism is to be reduced and desistance supported and encouraged. Does the fact that much recidivism occurs in the first days after release mean that people predisposed to fail usually fail quickly or that those days are especially risky for all released prisoners? How do parolees who fail early differ from those who fail later? Is motivation to succeed a key factor and if so, what kinds of programs and policies could support such motivation? More data are needed on the individual characteristics of persons who fail on parole. Recidivism Rates Recidivism rates can be seen as one measure of the failure to desist from crime. What drives this failure? Are the higher recidivism rates of parolees with multiple imprisonments a function of selection (a predisposition to fail among those who have failed before), the consequence of the criminogenic effects of imprisonment, the consequences of community characteristics, or the result of undetected individual traits? Special Populations What kinds of services will best meet the needs of specific groups of parolees in the future, such as women or elderly releasees? Why are racial and ethnic disparities in the rearrest and reincarceration of parolees differ- ent from (lower than) disparities in initial police contacts, arrests, convic- tions, and prison sentences? Informal Social Controls How do the effects of informal social control differ over the course of criminal careers? Do the known effects of marriage on desistance specifi- cally hold for parolees? What forms of informal control are most effective with younger parolees with high expected recidivism rates and low mar- riage rates? What drives the low rates of formal marriage among this popu-

OCR for page 72
 PAROLE, DESISTANCE FROM CRIME, AND COMMUNITY INTEGRATION lation and what could be done to increase rates of formal marriage? What dynamics underlie the difference in outcomes of marriage and cohabitation in relation to crime? How can parole take advantage of and leverage “natu- rally occurring” guardians in the community, such as spouses, parents, neighbors, and employers? Community Effects on Parolees What are the effects of neighborhood or community conditions—such as the presence of high crime rates or drug markets or the availability, or lack thereof, of social and treatment services—on parolees? Should prison and parole authorities consider relocating released prisoners away from communities with high levels of crime or other characteristics that impede desistance? Should such relocation strategies be considered only for released prisoners who do not have strong family or other social ties to their “home” communities? What would be the effects of relocation? Parolee Effects on Communities What are the effects of parolees on the crime rates of the communities (neighborhoods and cities) to which they return after release from prison? There are just two studies on the effects of released prisoners on state ar- rest rates (Rosenfeld et al., 2005; Raphael and Stoll, 2004). Do the find- ings also apply to local communities? Does the impact of parolees on local crime rates (if any) differ by crime type (i.e., violent crime, property crime, drug crime)? Intervention Effects Interventions can only be effective if they are affordable and can be implemented competently. Moreover, most postrelease interventions are viewed as adding costs to an already expensive system: thus, the cost of implementing reentry programs that build on research or even of conduct- ing the research itself creates a formidable obstacle to the kinds of changes we propose. Yet a number of studies have shown that these costs are far lower than the incarceration costs that are currently being incurred by parolees who have been returned to prison and are expected in the future (Aos et al., 2001; Castellano and Riker, 2001). A new report from the Pew Foundation (2007) forecasts that state and federal prison populations will grow by more than 192,000 inmates: these prisoners alone could cost as much as $27.5 billion in new operating and construction costs. The new costs will not be evenly distributed across states. For example, 18 percent of all parolees in the United States are in California (Zhang et

OCR for page 72
 CONCLUSIONS, RECOMMENDATION, AND RESEARCH AGENDA al., 2006). Moreover, a higher proportion of inmates in the future are likely to be female or elderly; both groups have special needs and higher costs. Enhanced penalties for certain kinds of offenders promise to raise costs as well. The committee believes that developing strong reentry programs that lower reoffending, rearrest, and reincarceration rates is critical to lower- ing these costs. Research on interventions should include cost effectiveness studies and should suggest how cost savings in terms of reduced prison costs could be realized. Effects of Prerelease Planning What kinds of reentry issues and problems are considered when devel- oping the prerelease plan? How can needs be prioritized to prevent early failure or death? How well are prerelease plans followed by the parole of- ficer and releasee over the course of the parole period? How do prerelease plans interact with availability and accessibility of services? Effects of Parole Prior efforts to improve parole programs have neglected one of the core functions—the role of the parole or probation supervision officer. Solomon and her colleagues (2005) have shown that formal parole supervision has limited effects: Why, then, are the recidivism effects of parole greater for some groups than for others? What is the role of agency culture and parole officers’ orientation and training? What is the contribution of a commu- nity’s capacity (e.g., program availability, resources) to foster desistance? What kinds of parole officer training are required when implementing new approaches or managing individualized re-entry plans for releasees? Designing Interventions Given the heterogeneity of the parolee population, what can be done to ensure that parolees are appropriately matched with specific interventions? Are there general programs (e.g., education and literacy programs, family- focused supervision) that are effective with all parolees? How effective in reducing recidivism and increasing desistance from crime are “triage” ap- proaches that concentrate services and treatments on lower risk parolees and intensify supervision for higher risk released prisoners? How effective are strength-based approaches in increasing desistance from crime?

OCR for page 72
0 PAROLE, DESISTANCE FROM CRIME, AND COMMUNITY INTEGRATION Timing Interventions Recent research on how the timing of surveillance, supervision, and ser- vices affect recidivism and desistance needs to be replicated and extended. Which interventions are most effective when introduced immediately after release from prison? Would it be better to begin certain reentry services and treatments prior to release from prison? If so, at what point during imprisonment should such interventions begin (e.g., a year before release, 6 months, 3 months)? How should they be connected to interventions in the community? Comprehensive Approaches Do comprehensive, multilevel strategies (involving community or orga- nizational change) produce significant reductions in recidivism? What types of community or organizational change are most effective? Effects of Restrictions on Releasees How do policies that restrict the access of released prisoners to public housing and other forms of public assistance—including treatment services, educational benefits, and other resources—affect desistance from crime? Role of Technology How effective are technological innovations, such as computerized re- porting, electronic monitoring, and global positioning system (GPS) moni- toring, in improving compliance with parole requirements and desistance from crime? For which offenders (e.g., sex offenders, gang members) is the technology warranted? Are such innovations cost-effective when compared with the traditional practices they supplement or replace? Sanctions for Parole Violation Do stringent special conditions in parole contracts cause parolees to fail? Are low-level sanctions, such as short stays in jail, for violating the conditions of parole effective in reducing the commission of new crimes? Should such sanctions be graduated in severity for subsequent violations, or are constant sanctions as effective as graduated sanctions in maintaining compliance with parole requirements and desistance from crime? What are the costs and benefits of alternative policies for technical violation on the overall justice system, crime, and criminal desistance?

OCR for page 72
 CONCLUSIONS, RECOMMENDATION, AND RESEARCH AGENDA Incentives for Parolees How can incentives be used along with negative sanctions to ensure that released prisoners comply with parole requirements and to encour- age desistance from crime? What types of incentives (e.g., shortening the length of parole, relaxation of certain requirements) are most effective? What is the benefit of state issued “certificates of rehabilitation” in foster- ing desistance? System Incentives What types of incentives are most effective in improving the morale and performance of parole officers and system response to released prisoners? Would a regime that ties organizational rewards to improved monitoring, service delivery, and compliance with parole requirements spur organiza- tional innovation? Can such a system increase desistance from crime by parolees in comparison with traditional parole procedures and practices? Measurement and Methods Measurement Issues How valid are arrests, technical violations, and other recidivism indica- tors as measures of desistance from crime among parolees? How well do violations of the technical requirements of community supervision predict the commission of new crimes? To what degree do recidivism measures, such as arrests, confound criminal offending with the system response to offending? Are conventional recidivism indicators more valid for some groups of parolees than others? Methods As this review unmistakably demonstrates, the application of scien- tifically rigorous methods in research and evaluation on community su- pervision has not been the norm and is only now beginning to emerge. Inadequate implementation of program principles and procedures appears to be a significant obstacle in the way of program effectiveness or of finding out whether a program might have benefits for participants. A major limita- tion of current program evaluation results is the failure to account fully for self-selection bias. Random assignment of persons to treatment and control conditions remains rare in research on the reentry process. What kinds of experimental evaluation and cost effectiveness studies could be designed and implemented to address and improve this situation? What methods are

OCR for page 72
 PAROLE, DESISTANCE FROM CRIME, AND COMMUNITY INTEGRATION most appropriate and how can barriers to using them (i.e. implementation, security, or ethics issues) be addressed and overcome? RECOMMENDATION The new work that confirms long-standing research findings on the high rates of recidivism and the risk of death in the first weeks and months after release from prison lead the committee to make a recommendation regarding policies and programs for parolees and other releasees. The committee recommends that parole authorities and administrators of both in-prison and postrelease programs redesign their activities and programs to provide major support to parolees and other releasees at the time of release. These interventions should be subjected to rigorous evaluation. Given the paucity of rigorous evidence about the effectiveness of many intervention programs or the motivation underlying individual change, the committee can offer only limited advice about what specific form some of these programs should take. Cognitive-behavioral approaches have strong scientific support and the committee believes that they should be widely implemented and continually evaluated, especially taking account of pro- gram implementation issues. Drug treatment coupled with frequent testing for drug use also shows evidence of lowering recidivism. Several other pro- grams and approaches show promise in reducing violations of community supervision requirements, arrests for new crimes, and drug use. Included here are programs that focus on individual change and motivation, and comprehensive, multiservice employment and training initiatives. “Nothing works” is no longer a defensible conclusion from assess- ments of program effects on reentry outcomes. When a person leaves prison it is clear that he or she has needs an immediate place to live, a person such as a case manager to facilitate the immediate transition from prison to the community, and a program to guide postrelease life. However, we cannot identify with confidence other best practices for reducing recidivism and enhancing desistance among people returning to local communities from prison. Because so few reentry service programs are accompanied by rigor- ous evaluations, a scientific review panel, such as this committee, has very little to draw on with confidence (see National Research Council, 1979, for a history of this problem). Yet there is a great deal of experiential and practitioner knowledge with regard to the apparent efficacy of various pro- grams (Wilkinson, 2004). The challenge over the next decade, as prisoner reentry, parole, and desistance from crime become even more important, is to subject these promising practices to rigorously designed evaluations.