Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

The United States correctional system is at a crossroads. Over the last 15 years, considerable attention has been drawn to the way that the US criminal justice system has been dealing with criminal offenders since the later part of the twentieth century. While disturbing, it is a well-known truth that the US incarceration rate is the highest among late modern democratic countries and the United States has 25 % of the world’s prisoners (Lacey, 2010). What is especially noteworthy is the growth in scale of punishment since the late 1970s. From 1920 to the mid-1970s, the incarceration rate was stable at around 100 per 100,000 people, but from 1980 to 2008, the US incarceration rate increased from 221 to 726 per 100,000 people (Western and Petit, 2010). By the end of 2010, about 1 in every 104 adults was in the custody of state or federal prisons or local jails; 1 in 33 is under some type of correctional control (Glaze (2011)). Academics, policymakers, and practitioners have argued that this level of incarceration is unsupportable from both philosophical and humanitarian perspectives and is economically unsustainable. The pressing question is how do we plan to address the emphasis on mass incarceration? This is perhaps one of the greatest challenges confronting our contemporary society.

What is unique about today’s correctional system is the massive size—over two million people incarcerated in prison and jail on any given day and another 5+ million on community supervision. And the two are not mutually exclusive. Failures on community supervision contribute to the size of the institutional population, and the size of the institutional population places demands on the need to expand community supervision. Yet, both institutional corrections and community corrections are stuck at the same place—the current array of institutional and community correctional programming is limited due to available resources, philosophies around the purpose of punishment, and historical attempts to remake and reshape the correctional landscape. That is, during the late 1980s and early 1990s when the war on drugs was waging strong and there was a surge in drug offenders with lengthier sentences, an attempt was made to remake the correctional landscape. The policy talk at that time (1990s) was focused on intermediate sanctions or the correctional interventions and programs that occurred between probation and prison. Morris and Tonry (1990), in their famous treatise Between Prison and Probation: Intermediate Punishment in a Rational Sentencing System, wrote:

Our plea is for neither increased leniency nor increased severity; our program, if implemented, would tend toward increased reliance on punishments more severe than probation and less severe than protracted imprisonment. At present, too many criminals are in prison, and too few are the subjects of enforced controls in the community. We are both too lenient and too severe; too lenient with many on probation who should be subject to tighter controls in the community, and too severe with many in prison and jail who would present no serious threat to community safety if they were under control in the community. (p. 3)

Morris and Tonry envisioned a community punishment system that had programming which would occur between standard probation (face-to-face contacts) and prison (secured institutional setting). They discussed fines, community service orders, house arrest, three types of probation (intensive supervision, residential conditions, and treatment conditions), intermittent imprisonment, restitution and compensation, fees for service, electronic monitoring, and forfeiture. The integration of these correctional interventions within the existing sanction and treatment structures faced significant setbacks. First, some were tried and tested, and it became apparent that the “public community” (including legislators, stakeholders, citizens, correctional and probation agencies, and offenders) was not ready for this form of punishment. For example, the concept of day fines was tried with a number of implementation barriers that impeded progress toward institutionalizing them in the United States (see Hillsman, 1990). Second, individual evaluations and more contemporary meta-analyses and systematic reviews have found that some of these interventions do not reduce recidivism. Such is the case for control-oriented intensive supervision (MacKenzie, 2006). If an intervention does not improve recidivism rates, then it begs the question as to whether we should routinely employ this intervention. Third, with insufficient resources, some of these innovations are partially (or even barely) implemented which dilutes their potential effectiveness. This is the case for electronic monitoring, probation conditions, probation with treatment, day reporting programs, some treatments such as cognitive behavioral therapy or therapeutic communities, and other ideals. Collectively, the systematic and organizational resistance coupled with insufficient attention to program fidelity created hesitations to move forward to implement a continuum of punishments that expanded from probation from prison.

The work of Morris and Tonry laid a foundation that many jurisdictions struggle to realize. Today there are new innovations developed during the 1990–2000s that are gaining support in the field and worthy of including in this system of punishments. The first is the growing use of drug treatment and problem-solving courts in the United States. These courts integrate treatment with control conditions to create the type of community controls that Morris and Tonry envisioned. Second, there are a host of new technological innovations that are front and center in terms of the potential to exact controls on offender behaviors. These include drug testing, GPS, electronic monitoring, and now smart phone applications that allow for daily diaries, journaling, and location monitoring (Pattavina, 2009). Technological advances will continue to influence the development of new approaches to support and monitor offenders in the community.

Morris and Tonry struggled with a system for determining the appropriate sentencing or punishment level for an individual. They outlined the concept of interchangeability that was based on equity among certain punishments in terms of their level and type of controls but allowed the punishment to be tailored to the individual’s situation. Hence, punishments could be “equivalent” in terms of severity, while substantively different. Similar to other sentencing schemes, the focus on assignment was based on severity of crime and criminal history, the two components of most sentencing guidelines. At the same time that Morris and Tonry were articulating this scheme, another set of scholars was advancing new concepts about offender management issues in corrections. Andrews and his colleagues offered a classification and programming scheme that focused more on the dynamic factors that affect offender outcomes. In their review of the literature, they proposed that correctional programming should be determined by the offender’s risk (criminal history) and needs (factors that affect their continued involvement in the criminal behavior (see Andrews, Bonta, and Hoge 1990). This was further developed into a framework referred to as risk–needs–responsivity (RNR), which focused attention on placement decisions based on the factors designed to control the risk of the offender to the community and attending to the factors that are most likely to reduce the likelihood of further involvement in the justice system. Figure 1.1 below combines the two models—Intermediate Punishment and RNR—into a vision for a correctional landscape that would best serve to reduce recidivism at the individual level and to build a correctional system that is responsive to the various needs of offenders. The model is based on the premise that recidivism reduction requires tailoring programming and placements to minimize risk but also using the least restrictive environment to achieve this goal. This book describes the development of a simulation models that allow jurisdictions and individual actors to put into place an empirically driven framework for making correctional placement assignments.

Fig. 1.1
figure 00011

Prototype risk–need model applied to various correctional settings

Full size image

The Predicament Arising from the Correctional Population Surge

The size and shape of the US correctional population has drawn serious attention since the recent economic recession. Political scientists and criminologists share the perspective that the dramatic growth of the prison population was largely achieved by policy changes that include the adoption of laws sending more drug and property offenders to prison (rather than jail or probation), lengthening prison sentences for various crimes, and requiring prisoners to serve larger portions of their sentences before being released (i.e., mandatory minimums) (Simon, 2010). Taken together these policy changes reflect a punitive or punishment model that emphasizes retributive justice and incapacitation as a means to promote public safety (Auerhahn, 2003). These initiatives appear to come at a high cost with only a modest return on investment—that is, if we use a utilitarian calculus of the costs and benefits of this experiment, we must assess the outcomes from a different perspective. Weisburg and Petersilia (2010) report that the growth in state imprisonment rates since 1985 accounted for no more than 25 % of the decline in serious crime during the 1990s. Western (2008) is more skeptical and reports that prison populations accounted for 10 % of the drop in serious crime. The modest decline in serious crime from 1993 to 2001 was achieved by the $53 billion in additional correctional spending and added half a million new prisoners (Western, 2008).

The cost of correctional expansion warrants more than just a discussion of outcomes couched in cost–benefit terms. The policies supporting the growth in incarceration also led to a growing disproportionate number of minority people from distressed communities being sent to prison (Clear, Waring, & Scully, 2005; Lynch & Sabol, 2001). Research has consistently shown that the policies governing punishment have resulted in the incarceration of a disproportionate number of black males. Blacks are 7 times more likely to be incarcerated than whites, and large racial disparities can be seen for all age groups and at different levels of education. One in nine black men in their twenties is now in prison or jail (Western, 2008). The same appears to be occurring to Hispanic males, but not to the same degree. Finally, the concentration of people with lower socioeconomic status in the justice system, regardless of race or ethnicity, continues to be problematic.

Prison conditions have also worsened in the wake of the growing carceral population. Overcrowding has led to unsafe and unhealthy living environments for inmates. Research indicates that because of housing needs, space for recreation and work and rehabilitation programs are eliminated to allow all useable space to be converted to dormitories for additional prisoners. Correctional workers in crowded facilities experience more job-related stress and fear of inmates (Martin, Lichtenstein, Jenkot, & Forde, 2012). Crowding has become so severe that the Supreme Court declared California prison crowding unconstitutional and is forcing the state to radically change the way it houses criminal offenders. Judges have recently ordered the state to reduce its inmate population as a way to improve medical care (Biskupic, 2011).

What happens upon release from prison illuminates the cumulative social impact of incarceration. Each year, over 739,000 inmates return home from prison. Those coming home from prison face significant challenges. The problem of prisoner reentry has been well documented by leading scholars in the field (Petersilia, 2005; Travis, Solomon, & Waul, 2001; Travis & Visher, 2005). Most return to neighborhoods of concentrated disadvantage where support services are lacking. Men with prison records are often out of work (Visher & Kachnowski, 2007). The jobs they do find pay little and do not offer the benefits and earning potential necessary to support the socially valued roles of husband and provider (Uggen, Wakefield, & Western, 2005; Western, 2008). Petersilia (2005) provides a general profile of soon to be released prisoners based on inmate survey data. She found that 41 % reported that they did not have a high school diploma and 33 % were unemployed the month before arrest. Family disruption and substance abuse were also problems for many inmates. Approximately 27 % were divorced and 59 % reported using drugs in month before committing their crime. Nine percent had an overnight admission for a mental condition.

The prevalence of these life circumstances among inmates reflects the wide range of cognitive and behavioral deficits that will continue to challenge them when they are released back into the community. Yet, the correctional system seems unresponsive to these issues. Petersilia (2005) describes the current state of knowledge regarding offender needs and prison treatment programs. She laments that prison administrators are not able to inform researchers regarding the number of prisoners who need different types of programs or the extent to which offenders participate in programs. Even in cases where counts are available, details about the duration and intensity of programs are often lacking and programs likely to be evaluated attesting to greater concerns about the quality of the treatment programming (Gendreau, 1996; Lowenkamp & Latessa, 2005; Welsh & Zajac, 2004). Her analysis of survey data shows that less than half of those in need of drug and alcohol treatment had enrolled in a relevant treatment program. Moreover, she reports that those most in need of certain programs are not always the most likely to participate in them. Given that these programs operate on a volunteer basis, the participating inmates may reflect the pool of inmates that are lower in need but “savvy” in that they realize that program participation both consumes idle time in prison and appears to suggest the offender is preparing for release. These offenders consume valuable space in limited programs, leaving behind those that are of greater need. She found that participation in educational and vocational programs was similar across low, moderate, and high need levels. Phelps (2011) in a recent article reviewing the availability of programming in correctional institutions found that the type of programming has shifted from treatment to more “reentry” services that emphasize life skills. Typical rehabilitation programs are far and few between given the growth in more life skill building programs which the meta-analysis literature finds to be of little importance to the recidivism reduction efforts.

A survey of treatment programs in prisons by Taxman, Perdoni, and Harrison (2007) identified major gaps in the availability and delivery of treatment services to offenders. Moreover, the services that are available are generally of low to moderate quality (Friedmann, Taxman, & Henderson, 2007). They also found that many correctional facilities attempt to provide services but that the resources available limit the size of the programs. Thus, program capacity amounts to a small percentage of their daily population. The services available tend to be more oriented toward educational awareness and minimal counseling, as opposed to intensive clinical and treatment services, reinforcing the shift noted by Phelps (2011).

Faced with limited opportunities for a productive life after prison, many offenders are likely to eventually recidivate. The 1994 recidivism study by Langan and Levin (2002) found that within 3 years of release from prison, 68 % were rearrested, 47 % are subsequently convicted, and 25 % returned to prison for a new crime. These rates seem impenetrable given the rather consistent finding across studies (see Pew, 2011). High recidivism rates suggest that people released from prison appear unprepared for life on the outside and that they are being entrenched in the wheels of justice. Many return to prison numerous times in a process Lynch and Sabol (2001) refer to as churning where many offenders cycle in and out of prison serving short sentences, getting released, and returning a few months later on another charge or parole violation only to be released again in a few months. Some states are responding to this problem by eliminating harsh sentencing practices that lead to crowding (Mauer, 2011). Other states are reducing the use of incarceration for non-criminal technical violations occurring during the period of supervised release after a prison/jail term. Another set of states are exploring the state of correctional programming for offenders in efforts to promote offender change and rehabilitation.

The attention brought to the problem of reentry by leading scholars has helped to change the discourse on correctional programming by reasserting the importance of rehabilitation. The notion of rehabilitation is slightly different here in that the treatment programs are being discussed in the context of altering recidivism. The linkage between other philosophies of punishment—just deserts, incapacitation, retribution, and deterrence—and recidivism has been shown to be weak (see Cullen & Jonson, 2012) giving rise to a concern that in the utilitarian assessment of whether the surge in correctional populations has been fruitful for societal gains, the costs (fiscal, humanitarian, increased recidivism, etc.) outweigh the benefits.

Discussion among policymakers has begun to focus on what the corrections system is doing to help offenders prepare for life on the outside and what support was available to communities. The Council of State Governments created the Reentry Policy with a mission to develop a collaborative report recommending policies intended to improve outcomes for returning prisoners, their families, and communities (Travis & Visher, 2005). Then in 2008, President Bush signed the Second Chance Act, which funds literacy programs, drug treatment, and other services for prisoners and ex-prisoners. The Second Chance Act can be viewed as one achievement in the broader movement for improved prisoner reentry policy (Western, 2008) and lays the groundwork for a revitalization of the principled correctional and community correctional programming to address the unintended consequences of the surge in using the correctional system.

The Challenges Before Us in Creating a Continuum of Recidivism Reduction Programming

Despite the growing support for offender rehabilitation in the public discourse on crime and punishment, financial support for programs and services has been slow to materialize. The Department of Justice proposed a 100 million budget allocation for the Second Chance Act which amounts to barely .14 % of the 70 billion spent on corrections each year (Gottschalk, 2010). Moreover, even though Attorney General Holder recently stated in a speech that the administration would not focus on incarceration as the sole means to protect the public, the 2010 and 2011 budgets increased allocations for law enforcement and new construction. Clearly, there continues to be some political reluctance to fully support programs that promote offender rehabilitation. So the challenge remains for researchers and practitioners to find ways to improve our correctional system in ways that promote offender change without compromising public safety.

Perhaps the reluctance of policymakers to more fully support rehabilitative programming stems from the conflict among researchers within the corrections field regarding the impact of rehabilitation and treatment programs. It is well known among the corrections community that the value of a rehabilitative approach toward criminal offenders has been significantly challenged by research reviews claiming that evaluation studies of treatment programs largely failed to demonstrate successful offender outcomes (Martinson, 1974; more recently Farabee, 2005). By focusing on the issues of whether a program is “effective” (i.e., reduced recidivism, null results compared to the control condition), the discussion has been that rehabilitation programs do not “work.” More recently some scholars contend that the effect size of the impact is small and may not be worth the investment (Clear & Austin, 2009). In response, advocates of rehabilitation such as Palmer (1992) suggest that we should interpret the work of Martinson and other detractors as a reminder that success is hard to come by and that correctional intervention has accomplished a great deal. The small effects can be increased by improving the quality of programming (Lowenkamp, Latessa, & Smith, 2005). The future of rehabilitation and correctional programming may have a more nuanced focus if we accept that it may not be possible to change all offenders, but we can devote our attention to addressing known criminogenic factors including substance dependence, social networks that include antisocial peers, and other targeted factors that affect offending activities. To this end, the system should encompass the principle of ensuring that appropriate offenders are placed in appropriate programs instead of putting people in the “first available” program. As we plan for the future of correctional interventions, it might be wise to focus our efforts on who to target for deterrence, for incapacitation, and for retribution and who then we should target for rehabilitation-type programs. This will not be an easy undertaking because it requires addressing very complex questions that involve issues of cost, feasibility, justice, and public safety (Feeney introduction to Palmer, 1992) as well as more recent questions of responsivity, efficacy of interventions, and treatment matching.

Despite the criticisms launched against the potential legitimacy of correctional programming (due to program quality issues, see Cullen & Jonston, 2010; Lowenkamp et al., 2006), the search for the most effective ways to promote offender change has continued among a committed group of advocates, practitioners, and academics (Cullen, 2005). Researchers working in this area have generated an important body of literature devoted to advancing models of offender risk assessment and linking needs to appropriate offender treatment programs. This is an evolving area of work as more attention is paid to the question of “what works for whom.” Even in the meta-analysis literature, a focus on moderator analyses to identify the patterns has emerged as scientists focus more attention on expanding our knowledge of maximizing our placement practices with evidence-based decision-making principles. While few studies have been able to isolate such patterns, researchers are committing to the use of moderators to better understand individual-level factors that account for positive (or negative) outcomes. In recent years, the RNR model (see below) has emerged as a dominant framework that emphasizes the importance of matching risk and need assessment with appropriate services that are consistent with the behaviors that drive their criminal activity (Ward & Maruna, 2007). It mirrors the movement in other fields—namely, substance abuse and mental health—where placement criteria have evolved to augment clinical decision-making processes to integrate evidence with clinical science to assist practitioners to improve individual-level outcomes.

According to Andrews and Bonta (2006, 2010), the RNR model integrates the psychology of criminal conduct into an understanding of how to reduce recidivism by targeting the unique individual factors that affect involvement in criminal behavior. The model proposes that correctional interventions should be structured according to three core rehabilitation principles: risk, need and responsivity. The risk principle specifies that offenders should be grouped by the criminal justice history that represents their “threat level” that a person may pose to society. Measures of static (historical) risk include age, criminal history, age at first arrest, number of prior probation violations, and other historical facts about individuals. The higher the level of risk, the greater the dosage or intensity of the treatment and controls should be. The need principle holds that the treatment of offenders should be targeted to specific dynamic risk factors (i.e., criminogenic needs) that are predictive of criminal behavior and that are amendable to change. Amenable to change infers the broad set of factors that contribute to offending behavior, but it does not include the demographic (i.e., age, gender, cultural) that may help explain recidivism rates but that there is little a person can do about these factors. These key dynamic risk factors are antisocial values and attitudes, antisocial peers and associates, criminal subculture, low self-control, substance use disorders, and dysfunctional family environment. The number and type of criminogenic need also drive the targeting decision in that offenders that exhibit more than one factor should benefit from more intensive services, treatments, and controls. The criminogenic need category focuses on presenting factors that can be addressed with proper attention. The third principle is responsivity, which involves the proper matching of correctional interventions in ways that consider contextual factors such as the ability and learning style of the offenders, the number of non-criminogenic destabilizing factors (i.e., mental health disorders, lower literacy rates, negative work history), and strengths such as stabilizing factors of a strong support network, good work and educational experiences, and positive social skills. Gender, age, and culture also affect responsivity since some programs, treatments, or approaches may be more beneficial to different demographic factors such as the gender-specific treatment, developmentally appropriate treatment, and culture competency in approaches.

When implemented correctly, the concept of service matching that is guided by principles of RNR is considered best practices for corrections (Taxman & Marlowe, 2006) and has been shown to significantly reduce recidivism in certain settings (Andrews, Zinger, Hoge, Bonta, Gendreau, and Cullen 1990). Research has also shown that nonadherence to RNR principles in service delivery, however, is not only ineffective but also detrimental to offender treatment outcomes (Lowenkamp & Latessa, 2005). Not treating offenders or placing offenders in inappropriate treatments can increase the risk of recidivism. Moreover, research suggests that program caliber is an important consideration when considering treatment delivery. Attending to implementation and quality is an important factor affecting the spread and utilization of treatment programming. There is a need to go beyond merely looking at the program components to assessing the quality of the delivery. Friedmann et al. (2007) reported that the attention to evidence-based practices is low in correctional programming including that treatment programs with the same name and identical treatment manuals vary in their overall program effectiveness from jurisdiction to jurisdiction (Latessa, Smith, Schweitzer, & Brusman Lovins, 2009). Taxman and Bouffard (2003) observed that there is great variation in substance abuse counseling regardless of the known program components.

A sizeable and growing body of literature devoted to each of the three RNR principles exists. For example, there is a growing body of literature on the development and application of relevant risk and need assessment instruments (see Pattavina & Taxman, 2007). A variety of risk assessment tools are available and many correctional agencies advancing in their use of such tools. While the risk tools may vary in content, they have the collective purpose of determining who is at higher risk of reoffending and identify the deficits and strengths of each inmate. This information is used to determine appropriate program planning. We are also learning more about which RNR-based programs are most successful at promoting offender change. Practitioners that have incorporated RNR elements into cognitive-based treatment plans and evidence-based reviews appear more satisfied that these programs significantly reduce recidivism. Evidence-based assessments of programs are necessary for determining which programs are most appropriate and which programs should be expanded or eliminated. Applying evidence-based research findings to the search for “best practices,” evidence-based practices, or promising strategies benefits offenders, promotes public safety, and may be more cost-effective.

The commitment to evidence-based approaches serves to increase the demand for more rigorous research designs necessary to assess valid program evaluations and answer the question of what works for whom. The growing interest in determining “what works” has led to support for initiatives that promote evidence-based research. Examples include the Washington State Institute for Public Policy, sponsored by a state legislature to conduct program evaluation research; Crime Solutions, supported by the US Department of Justice (www.crimesolutions.org); and the Campbell Collaboration responsible for sponsoring a variety of evidence-based reviews on effective crime reduction strategies (Mauer, 2011).

It is not sufficient to rely only on the literature that addresses the respective elements of the RNR model separately as a way to determine the overall significance for correctional practice. What makes the RNR especially appealing is the focus on the interconnectedness among the three principles needed to achieve the most successful outcomes and the ability to provide more rationality to sentencing schemes and program placement criteria. While we continue to move forward with producing quality research studies with respect to program matching, quality, and effectiveness, it is equally important that we begin to examine the implementation implications of this model within an operational context (Ward & Maruna, 2007; Taxman and Belenko, 2012). A continuing need exists to expand the research base to assess how the connections among RNR dimensions operate in real-world correctional settings. Research shows that the current distribution of treatment services to offenders in prison, jail, and community corrections is inconsistent with the needs of the offender populations, as discussed in Chaps. 2 and 6 in this book. Significant improvements cannot be made unless this gap is closed. There is thus a pressing need to help jurisdictions develop guidelines as to how to allocate offenders into appropriate services. The list continues to grow of the facets of how to fine-tune the correctional system to integrate the rudiments of RNR-based evidence-based practices.

About This Book

We wrote this book to articulate an approach to implementing RNR into practice in justice, correctional, and health organizations that serve people involved in the criminal justice system. In the chapters that follow, the authors present research from projects designed to collectively inform the comprehensive development of a model that is grounded in principles of the RNR model and attempts to make connections among the principles in a way that maximizes matching that will produce reductions in recidivism. Ultimately the model can also include the cost-effective possibilities. The work we present will establish validated estimates of key parameters that describe the national corrections population and appropriate treatment services. These parameters will be used to build simulation models designed to examine how varying levels of RNR implementation affect offender recidivism. At the national level, simulation results can be used to inform the debate on the integration of RNR as an emergent framework into practice. At the local level, simulation inputs are translated into an expert system designed to assist in the day-to-day decisions correctional staff make about the best program options for offenders available in specific jurisdictional settings. Much of this book is about building the RNR model to incorporate the major research findings and then demonstrating how the model can work as a static model and a discrete-event model. The components of the RNR Simulation Tool expert system are described in this book.

Taxman and colleagues will present a case study for the issues related to treatment gaps in Chap. 2. The purpose of this chapter is to establish clearly the issues that confront the RNR model—that is, a methodology for examining how to assess current level of programming, a range of programming available, and the gap between need and programming. It also shows the assumptions that are plausible and needed in an RNR model. The case study focuses on substance abuse treatment to establish some of the key components of models. But it makes the case that provides the approach for a broader range of criminogenic needs.

How to build a useful simulation model is addressed in Chap. 3. Greasley introduces the reader to the stages of simulation model development and makes recommendations on how to build and validate models. Simulation models are often built in stages and rely heavily on model conceptualization and process mapping. He outlines the technical features associated with constructing a working model and discuss methods used to test alternatives.

In Chap. 4, Taxman, Caudy, Pattavina, Byrne, and Durso present the empirical basis for the RNR model that will be used as the basis for simulation models presented in the subsequent book chapters. They will identify important assumptions relevant to the RNR model in measuring risk and needs, as well as the issues related to responsivity. Their interpretation of RNR model will allow us to transcend the “what works” mantra to the more focused question of what works for what kind of offender and under what circumstances (Brennan, 2012). The assessment and treatment needs that derive from their summary of the RNR framework will be used as a basis for measurement of RNR concepts in the chapters that follow.

The RNR framework presented in Chap. 4 serves as the guide for the design of contemporary data-driven techniques that will be used in subsequent chapters to create and validate measures of offender risk and criminogenic needs, build the link between risk and needs and appropriate treatment groups, and identify the programs that work for each treatment group. The offender risk/need profile distributions and matched treatment options will then be used as inputs for a nationally based simulation model that examines recidivism outcomes for offenders through RNR adoption in a prison setting. The distributions will also be adjusted to reflect locally based offender inputs for use in an expert system to guide local jurisdictions in implementing RNR-based program model. A specific simulation model designed to estimate the cost-effectiveness and public safety outcomes from programs that divert special populations from jail to community alternatives is also presented. The offender risk and need profiles developed and validated in Chap. 5 will be used by the authors in Chap. 8, to inform the creation of a synthetic data set designed to reflect the profile distributions and associated recidivism estimates of an inmate population.

Creating and validating offender risk and need measures is an essential first step in mapping the RNR process. An important component of the proper use of risk assessment instruments is the practice of validating the instrument for the particular sample on which it is to be applied. Offender populations vary across jurisdictions according to age, gender, race, ethnicity, and type of crimes. It is therefore a best practice to measure an instrument’s validity for a particular sample before using it in that setting to make treatment placement decisions. In Chap. 5, Ainsworth and colleagues present the construction of a nationally representative database that merges publically available data on offender risk, needs, and recidivism. This is followed by a discussion of static risk and criminogenic need factors and the various procedures that were used to create and validate risk and need scales created using synthetic data. The resulting risk and need profiles and distributions created from these data will be used as the standard inputs in subsequent chapters that examine program matching and RNR outcomes for offenders at the national level.

In Chap. 6, Crites and colleagues present a method used to incorporate responsivity concepts into treatment planning. She uses the risk and need profile parameter estimates developed in Chap. 3 to identify the appropriate program content and dosage that meet offender risk levels and needs. Six program levels are described representing a continuum of care using increasing intensities of programming targeting different levels of needs. For individuals, key contributors to program-level assignment are risk level based on criminal history, dependence on hard drugs, multiple criminogenic needs, and presence of multiple destabilizing factors (e.g., unstable housing, dysfunctional family, low education).

In addition to identifying the appropriate programming level for individuals, Crites chapter also describes a program classification tool that is designed to identify which program level a specific program or intervention fits into based on characteristics such as target, dosage (clinical hours), content, and staff credentials. Once individuals and programs have been classified, individuals can be match to programs within the appropriate program level to meet their needs. In response to the need for program fidelity, a special program assessment tool is developed that will allow jurisdictions to evaluate available programs along four dimensions including setting, duration, content, and caliber. This chapter concludes with a discussion of pilot tests of this model using data from state and local criminal justice agencies.

Connecting individual risk and needs (described in Chap. 5) with appropriate program levels (described in Chap. 6) is the first stage of the model. Next, the simulation model must determine which specific correctional interventions are most effective at reducing recidivism in each level. In Chap. 7, Caudy and colleagues discuss how evidence-based reviews and meta-analytic findings from the field of criminology can be used to inform simulation model that estimates the impact of adherence to the principles of the RNR model on recidivism. Because meta-analytic research syntheses provide summary effect sizes which reflect the numerous primary studies that have been done on a given topic, they are particularly well suited for informing policy and practice. This chapter illustrates the utility of evidence-based reviews and meta-analyses for identifying the most effective correctional treatment programs. Successful programs can be added to program-level inventories and used as resources for practitioners when selecting appropriate programs for their jurisdictions as well as assisting in the treatment matching process. Caudy and colleagues subsequently illustrate different approaches to measuring system outcomes and simulate the impact on recidivism. Ultimately the model should model should help us identify what impacts might we expect on recidivism if we are able to effectively transfer the RNR principles of effective treatment into actual correctional settings? This is an important concern, and structuring correctional treatment protocols to be consistent with RNR principles in real-world settings may be easier said than done. Research in this area is limited, but a recent study in a local prison setting conducted by Bourgon and Armstrong (2005) found that when properly implemented within an RNR framework, treatment significantly reduces recidivism. Caudy et al found that treating 4 offenders with RNR programs will prevent one recidivism event. This is in contrast to punishing 33 people in order to prevent one recidivism event.

Given that the offender risk and need profiles constructed in Chap. 5 are created from nationally based data sources, they are most useful for supporting RNR implementation on a national level. The profile distributions will require adjustments to reflect locally based populations to support state and local jurisdictions wishing to adhere to an RNR-based offender treatment protocol. In Chap. 8, Bhati and Taxman describe the design and use of a synthetic database for this purpose as the foundation for a simulation model. The model borrows the parameters discussed in Chaps. 5, 6, and 7. Synthetic databases have, at their core, theoretically possible attribute profiles. The profiles are weighted (or re-weighted) to reflect different aggregate properties. The properties may reflect such features as means, rates, variances, covariances, and correlations of various attributes. In effect, once constructed, the synthetic database can be analyzed in much the same way as a real sample from the population of interest. In other words, the synthetic databases can be customized to reflect the characteristics of a local jurisdiction, thereby making it more relevant for localized policy simulations. This chapter describes the methodology used in constructing and re-weighting synthetic databases and demonstrates the procedure with real data from several jurisdictions. This chapter provides an overview of the RNR Simulation Tool expert system and discusses its potential applications to the field. The tool is comprised of three portals that operate collectively to guide the application of the RNR principles in a variety of correctional settings. This innovative web-based simulation tool provides decision support for agencies at the individual, program, and jurisdictional level.

The preceding chapters have used empirical evidence supporting RNR to create links among the principles that model a delivery framework that can be used to guide implementation in correctional settings. This framework presents an opportunity to use simulation techniques to investigate the potential impacts of implementing RNR practice without requiring changes to existing system. Simulation generally refers to a computerized version of a model, which is run over time to study the outcome implications of defined interactions. For our purposes, simulation can be used to show the effects of RNR implementation in a virtual setting.

Simulation techniques have been used to model criminal justice system operations dating as far back as the 1970s (Nagel, 1977). Despite a long history in the criminal justice field, simulation modeling was not widely used due to large resources that were necessary to build and maintain complex integrated models, along with the lack of available data to validate model outcomes. Advances in computer technology and simulation software have made access to simulation model development easier, and the availability of archived criminal justice data sets has provided important resources that can be used to build and validate model inputs. Simulation models have become useful tools for investigating the impacts of various sentencing strategies on the corrections system (Auerhahn, 2003) forecasting prison populations (Austin, 1990) and more recently have been introduced to examining the long-term effects of drug addiction (Zarkin, Dunlap, Hicks, & Mamo, 2005). The application of simulation techniques to assess the RNR model is appropriate given that we are interested in understanding the impact of adopting RNR as a model of correctional treatment. Simulation allows us to create an operational computerized version of the RNR model and then explore various “what if” scenarios regarding RNR implementation and compare the outcomes without disrupting the existing system Chap. 10 describes the building and application of a discrete event simulation model to examine the impact of several treatment scenarios on recidivism at the national level. The results show that RNR programing substanially reduces the number of returns to prison.

Special populations present unique challenges for correctional treatment delivery. This may be particularly true for patients with serious mental illness. Some of these offenders would be better served by being placed in specifically designed treatment programs in the community rather than in jail where serves are lacking. Simulation modeling can be useful for investigating the impacts of jail diversion on these populations and on system outcomes such as cost and public safety. Chapter 9 provides details on a simulation model for projecting the costs and benefits of comprehensive and evidence-based services for mentally ill offenders. The development of the model had two main objectives: (1) to develop the model using operations research methods to simulate the impacts of jail diversion programs and (2) to test that model to obtain projections of the fiscal and client outcome implications of implementing a jail diversion program for the criminal justice system, the mental health system, and the total system expenditures in a community. The model results quickly allowed a comparison of diverted and not diverted groups on several key variables, including costs to mental health and substance abuse systems, costs to the criminal justice agencies, jail days, and individual outcomes (i.e., functional level improvement).

The Mental Health/Jail Diversion Simulation Model provides a tool for communities to use in the process of planning a jail diversion program with a fiscal impact assessment. The model addresses an important public policy consideration: specifically, whether and to what extent jail diversion achieves current and future cost savings. The model confirms the pattern of cost shifting from the criminal justice to the mental health system observed in prior studies. Moreover, the results of the simulations provide stakeholders responsible for designing jail diversion programs with insight into how eligibility criteria affect the pool of individuals who can be intercepted, as well as the overall fiscal impact of the interception itself.

In the last chapter of this book, Taxman, Caudy and Pattavina discuss the future of RNR modeling and simulation for the US correctional system. Whether the goal is to better understand how the criminal justice system works or to examine the possible outcomes of anticipated or planned changes in criminal policies or practices related to correctional treatment, the authors will draw upon the work in this book that demonstrates simulation models can be useful tools for building knowledge about the operation and improvement of the criminal justice system.

The particular focus of this book has been on correctional treatment and planning using an RNR treatment approach. The application and use of simulation tools hold much promise for the future of corrections because policymakers and practitioners are looking for improved means to manage correctional populations in ways that can help offenders lead productive lives. Moreover, academics have provided important treatment frameworks and evidence-based studies to inform the search for improved treatment options. Simulation can be used to test the effects of changes in treatment delivery on a national scale and to serve as a basis for an expert system designed to aid local practitioners. Despite the promise that simulation holds for advancing correctional goals, challenges remain. This chapter concludes with a discussion of the challenges and suggest opportunities for advancing simulation work in criminal justice.