Keywords

In this chapter, we review the literature surrounding restorative justice as a response to crime. While originally introduced in the form of an alternative dispute resolution process, only loosely incorporated into the formal criminal justice system, restorative justice is now a vital feature of justice processes in many jurisdictions. Here, we offer a thorough introduction to the theoretical foundations and empirical realities associated with restorative procedures. Our goal is to provide a wide-ranging review of what is currently known about restorative interventions in the psychological literature and what remains unclear.

We begin with an in-depth introduction to restorative justice, comparing it with the formal criminal justice system in terms of philosophical foundations, structural applications, and outcomes. From here, we move to a review of recent empirical evidence. We reflect on the findings of program evaluations (in particular, those related to stakeholders’ perceptions and recidivism reduction), evaluating the empirical support for reintegrative shaming and procedural justice as central theoretical explanations for the effects seen in restorative justice. In this section of the chapter, we summarize what is well-known about restorative justice in the psychological literature, highlighting the importance of continuing to better understand factors driving the effectiveness of restorative interventions.

Finally, we emphasize the need to investigate public perceptions of appropriate justice; we argue these are key to identifying the conditions under which restorative interventions will be evaluated as an acceptable justice response and, therefore, they will greatly affect the long-term viability of restorative justice. We dedicate the latter part of the chapter to addressing an underdeveloped area in relation to restorative justice: dual-process models of cognitive processing . Ultimately, we argue that the success of restorative justice is as dependent on public perceptions of the appropriateness of restorative interventions as it is on demonstrations of the effectiveness of such interventions in evaluative research. We draw attention to the discrepancy between legal and lay notions of justice, on the basis that legal notions of justice revolve around deliberative cognitive processing , whereas lay notions of justice stem from heuristic cognitive processing and are often dominated by a retributive impulse . We speculate about the effect of this discrepancy on the long-term viability of restorative practices, arguing that retributive motivations present an inherent challenge to the widespread use of restorative justice, but that psychological research on dual-process models will be useful in overcoming this challenge.

We conclude by suggesting more focused research directions that would advance the current state of knowledge in restorative justice, highlighting the value of experimental methods in building support for restorative interventions. We argue that such research would better equip scholars to overcome the challenges presented by the retributive impulse in ways that encourage the long-term viability of restorative justice.

What Is Restorative Justice?

The formal criminal justice system is characterized by distinctive features, including an adversarial process, adjudication by unbiased third parties (such as judges and juries), and punitive outcomes (most notably, incarceration). The court-based model associated with this system is likely the most well-recognized means of responding to criminal behavior among citizens of Western democracies such as Australia , Canada , and the USA. Despite several decades of application alongside or as an alternative to the formal criminal justice system (Braithwaite, 1999; Strang & Sherman, 2015), restorative justice is undoubtedly less well recognized as a means of responding to crime. While both the traditional and restorative justice models prioritize the provision of an efficient and effective response to crime as their ultimate purpose, there are notable differences between the two in terms of philosophical foundations, structural applications, and outcomes.

Philosophically, the formal criminal justice system and restorative justice differ in the extent to which they prioritize retribution. The two models are founded on distinct justice motivations that differentially prioritize retribution (i.e., punishment) and restoration (i.e., reconciliation of harm caused by a transgression; Wenzel, Okimoto, Feather, & Platow, 2008). Formal models of criminal justice in Western democracies have historically been grounded in the retributive philosophy. This orientation is linked closely with the dominance of rational choice sensibilities in Western nations. Central to the rational choice position are matters of individual culpability and calculated decision-making . From this perspective, consequences are recognized as very important to deterring secondary offending behavior generally, and primary offending behavior specifically (Zehr, 1997). A retributive philosophy naturally complements rational choice sensibilities, prioritizing punishment as the most effective response to criminal behavior. Punishment reprimands the individual criminal, while also conveying a message of the consequences of engaging in criminal behavior to would-be offenders in the broader population. Formal models of criminal justice embody the punitive orientation that accompanies a retributive philosophy of justice. However, retribution is but one justice motivation, and an alternative motivation—that of restoration—is generally not well represented in responses to crime that prioritize retribution (Braithwaite, 1989, 2000).

The fundamental orientation of restorative justice models is identifying and resolving the harm produced by criminal behavior, rather than simply punishing the offender. This involves recognizing that criminal acts produce harm for a range of parties, and that a central objective of responding to crime should be reconciling the harm that has been caused for all stakeholders. Restoration of the victim(s) and community to their pre-transgression states is a key priority of restorative justice, but comprehensive restoration also involves engaging with the harms suffered by the offender (Bergseth & Bouffard, 2007). Returning the offender to his pre-transgression state, in addition to providing him with opportunities for self-improvement, is key to a restorative philosophy. The philosophical starting points of justice models prioritizing retributive versus restorative motivations therefore vary considerably, highlighting distinct objectives. Their fundamentally different frames of reference lead to fundamentally different processes and outcomes.

The retributive philosophy guiding the formal criminal justice system in Western democracies lays the foundation for an adversarial process, in which the focus is the behavior and culpability of the accused. The procedure is characterized by pitting two parties (the defense and the state) against each other, who do not agree on a shared understanding of the accused’s behavior. The very language foundations of this model (e.g., guilty/not guilty, defense/prosecution) are indicative of the level of confrontation inherent in this approach; undoubtedly, there will be parties who are dissatisfied with the outcome. To minimize the financial costs associated with the process of the formal criminal justice system, plea agreements—which require offenders to admit to a crime (typically of lesser severity than the crime with which they were initially charged)—are often used to establish a form of negotiated consensus at the expense of the state’s desired conviction. However, when a shared understanding of offender culpability cannot be reached between the defense and prosecution, the full adversarial process is engaged. The offender is presumed innocent and represented by a defense attorney, while the state carries the burden of proof and is represented by a prosecuting attorney. Both parties have the opportunity to present relevant evidence to a designated third party (i.e., a judge, panel of judges, or jury) tasked with reaching a determination regarding the accused’s guilt. Victims are most typically incorporated into this process as witnesses, having very limited opportunity to share information that is not specifically asked of them or which is not directly relevant to the accused’s culpability—amounting to very restricted opportunities to have a voice in the process. All parties to the procedure have clearly defined roles in the process, which creates very formal, structured interactions, generally with the victim and accused not directly interacting at all, and with less directly affected victims (e.g., family members of the primary victim/s) having very little to no involvement in the procedure. If a determination of guilt is reached, the potential outcomes are focused on offender punishment and tend to be narrowly constructed (e.g., probation, incarceration, fines); these outcomes are capable of being administered to a broad offender population but do not necessarily speak well to individual offenders’ needs and largely fail to consider the needs of other stakeholders. Ultimately, the consequence imposed is intended to be the result of a sober calculation of proportional punishment that draws from existing formal guidelines and legal precedents (Bazemore & Umbreit, 1995).

In contrast, restorative procedures are relatively informal when compared with court procedures. There is not a single formula for conducting a restorative procedure as there is with court. Restorative practices have been recognized as taking a variety of forms, with victim-offender mediation, family-group conferencing , and peacemaking circles being recognized as some of the most prevalent strategies (Bouffard, Cooper, & Bergseth, 2016; Latimer, Dowden, & Muise, 2005; Zehr, 2002).

In general, restorative justice procedures can be characterized by a broad set of basic features which are fundamentally different from court. For instance, although restorative approaches tend to have a “leader” guiding the procedure—in the same way that a judge might be described as the leader in court-based procedures—this leadership is much less rigid and formal. Facilitators (as they are called in restorative conferencing) guide participants through the procedure (e.g., encouraging reflection on specific topics such as harm incurred and ideal restoration, as well as ensuring that all participants have adequate opportunities for voice), but they do not govern the procedure in the same way that a judge does (Wenzel et al., 2008). Instead, conference participants have substantial ownership over the workings of the process. Restorative justice procedures require the participation of an offender who acknowledges commission of an offence—very different from the role the accused adopts in court-based procedures. Additional participants in restorative justice ideally include victim(s), support persons for both offenders and victims, and, in some cases, community members more generally. Conference participants are substantially more engaged in restorative justice than court participants are in court procedures. For instance, while the state represents the victim in the formal criminal justice system, victims are given much greater latitude to choose the manner in which they represent themselves and play an active, decisive role in the restorative model (Zehr, 1997). Restorative justice advocates argue that this does not imply the offender is disenfranchised relative to the victim in restorative procedures (Braithwaite, 2000). Instead, the shared starting point that restorative justice begins with (i.e., acknowledgment by all parties of the harm caused and an effort to resolve that harm) is intended to level perceptions of hierarchy in the procedure, between participants as well as facilitators. In these ways, responses to crime founded on both restorative and retributive philosophies result in consequences for criminal behavior, but there is immense procedural variability between the two.

While the consequences prioritized in the formal criminal justice system emphasize punishment, this does not mean that the consequences of retributive oriented justice models are inherently cruel, harsh, or unfair; rather, retributive models of criminal justice aim to swiftly, proportionally, and fairly administer a punitive consequence to a particular offender while also conveying a message of general deterrence to the public more broadly (Bazemore & Umbreit, 1995). While addressing punitive concerns, models of justice prioritizing a retributive orientation tend to be criticized for being inattentive to a broader range of post-crime concerns, such as restoration of harm. In this regard, a restorative orientation to justice has proven particularly valuable, as a complementary as well as an alternative model to the formal criminal justice system.

Consequences are conceptualized differently in restorative justice, with the terminology of sanctions preferred to punishments, underscoring that restorative outcomes are not primarily punitive in orientation. Sanctions are the tangible outcomes of restorative procedures for offenders. The principles of specific and general deterrence are still communicated through sanctions; however, rather than prioritizing punishment, sanctions prioritize the restoration of stakeholders both materially and in terms of relationships following a transgression (Van Ness, 1993). A fundamental difference between retributive and restorative philosophies of justice is observable through sanctions. Specifically, restorative justice emphasizes the redemption of the offender; while an act worthy of condemnation has been committed, the individual offender is socially recognized as redeemable. Sanctions are still directed at the offender in restorative interventions—as punishments are in retributive models of justice—but restorative justice sanctions strive to assist the offender in returning specific victims, the community more broadly, and themselves, to their pre-transgression states. Just as participants are granted greater ownership over the procedure in restorative justice, they are also granted considerably more control over the crafting of outcomes (Braithwaite, 2000; Latimer et al., 2005). Consequently, restorative sanctions tend to be much more responsive to the needs of particular situations and stakeholders than the consequences of the formal criminal justice system (Bazemore & Umbreit, 1995). While there is considerable flexibility in the specific sanctions developed in any given scenario, sanctions are not inherently inconsistent. Two underlying motivations are intended to guide the development of all sanctions: stakeholder restoration and the personal growth of offenders (Zehr, 1997). Through these more diversified outcomes, restorative justice is argued to more reflexively and successfully respond to the needs of victims, offenders, and communities.

More specific examples of sanctions in restorative justice would include personal growth opportunities (e.g., completion of programs or tasks), victim compensation (e.g., personal service or financial remuneration), and community compensation (e.g., community service or fines), which would be selected based on offender, victim, and community needs. Take the instance of a case where a 16-year-old youth (John) is engaged in a restorative conference because he has vandalized a neighbor’s property. John (offender), the neighbor (victim), John’s parents (supporters), and the local neighborhood watch leader (community member) are all present for the conference along with the trained conference facilitator. During the conference, John reveals that he has a difficult time coping with anger and committed the vandalism as a random act of destruction in frustration after his parents told him they could not afford to give him money to go out to the movies with his friends. All of the conference participants agree that John has behaved inappropriately, but that steps can be taken to rectify the situation. The neighbor requests that he be compensated for the $100 of damage that John caused, but recognizes John’s parents’ strained financial situation and suggests that it would be best for John to be responsible for the reimbursement. John agrees, but says he has no money to compensate his neighbor and does not know how to get it. The neighbor suggests a local office that offers classes in résumé building as well as an employment service that helps connect potential employees with employers. John accepts these suggestions and three sanctions are established; John is expected to: (1) complete a résumé within 2 weeks at the local office, (2) submit his résumé to the employment office and apply to at least three available part-time job opportunities within 3 weeks, and (3) reimburse his neighbor financially or through personal assistance if he cannot secure the money within 3 months. Furthermore, John’s parents draw attention to the need for John to better control his emotions . The neighborhood watch leader describes an emotion management program in the community that helps youth develop patience, stress management, and accomplishment through gardening. John is interested in this program and the rest of the conference participants agree that John and the community would benefit from him learning to better control his emotions. As such, a fourth sanction, that John must join a program oriented towards emotion management and attend regular sessions for a minimum of 2 months, is also established. As the example illustrates, the victim, offender, and community more broadly participated in the determination of meaningful sanctions that aim to restore the victim to his pre-transgression state alongside restoring, or even improving, the pre-transgression states of the offender and community more generally.

The preceding review has now established an understanding of the philosophy, procedure, and outcomes of restorative justice, relative to the more dominant retributive justice model. Restorative justice has been described as offering a response to crime that is less punitive and stigmatizing than retributive justice, providing the offender with greater opportunities for personal growth and community reintegration while also responding to the harms of crime more broadly (i.e., victim and community concerns; Bazemore & Umbreit, 1995). However, the conceptual overview provided thus far largely centers on the intentions of the restorative justice approach, engaging very little with the empirical realities associated with the use of restorative procedures. Although restorative procedures conceptually promote a more holistic, empathetic response to wrongdoing, we cannot assume that these approaches to crime are inherently better or, indeed, that they do not cause harm to stakeholders (Strang & Sherman, 2015). Evidence-based policy is key to the development and administration of responses to crime; reflecting on the findings of program evaluations, theoretical explanations of effects, and links between theory and practice are all central in this regard.

The heterogeneity of restorative justice programs is both a strength and a challenge of restorative responses to crime. Flexibility in program design and administration allows individual programs to be geared towards community needs, as well as allowing individual conferences to be responsive to the realities of specific victims and offenders. Embracing diverse tactics is a quality of restorative justice that contributes to its success in some ways; however, this strength also presents a challenge to evaluating and replicating the achievements of restorative programs. This has been, and will surely remain, a reality that makes it difficult to be confident in reaching definitive conclusions in evaluative research. Despite this caveat, a number of high quality evaluations and meta-analyses have illuminated the value of restorative responses to crime. Through these empirical projects, there is an accumulation of evidence supportive of some of the central optimistic promises associated with restorative justice; in particular, its capacity for stakeholder restoration and satisfaction, as well as reductions in recidivism (Braithwaite, 1999). These findings are demonstrated across a number of research projects engaging a variety of methodological approaches and exploring a diverse range of restorative procedures, but they are generally consistent with regard to stakeholders’ perceptions and rates of recidivism.

Stakeholder Perceptions

Restorative justice has long been argued to foster more positive sentiments among both victims and offenders than the formal criminal justice system (Braithwaite, 1999). These positive outcomes are observable in stakeholders’ perceptions of each other as well as of the justice process more generally. Most notably, both victims and offenders report greater perceptions of satisfaction with restorative relative to retributive procedures. This effect has a substantial history of empirical validation in a number of high quality meta-analyses (e.g., Latimer et al., 2005; Mazerolle, Antrobus, Bennett, & Tyler, 2013; Sherman & Strang, 2007). More recent program evaluations continue to support these findings. For instance, Bouffard et al. (2016) examined a variety of restorative interventions for youth (incorporating varied degrees of direct contact between offenders and victims) and concluded that “participants in RJ programs often report high degrees of satisfaction with the intervention” (p. 14). Researchers tend to attribute these effects on satisfaction to the highly engaging nature of restorative justice relative to court-based procedures. In particular, opportunities for voice during restorative procedures help establish a shared sense of process control between parties, while the joint crafting of sanctions promotes the understanding that outcomes are a collaborative effort between parties. The ability of all parties to actively contribute to both the process and outcomes of the procedure enhances perceptions of fairness and satisfaction from the perspective of both victims and offenders.

Restorative justice practices are also associated with other positive perceptions for stakeholders. For instance, offenders participating in restorative processes report higher levels of engagement in the proceedings as well as enhanced perceptions of ethical treatment relative to court-based procedures (Barnes, Hyatt, Angel, Strang, & Sherman, 2013). Much of the research has focused on victims’ experiences, producing findings that participation in restorative conferencing leaves victims feeling less fearful of offenders (Strang, 2002), less angry with offenders Sherman & Strang, 2007), and less likely to experience the symptoms of post-traumatic stress disorder (Angel et al., 2014). Overall, a substantial body of evidence suggests that restorative procedures leave victims and offenders feeling more satisfied with their encounter, and more positive in general, than does the formal criminal justice system. While stakeholder perceptions have long been of interest to researchers, rates of recidivism are the most broadly recognized measure of the success of restorative interventions.

Recidivism

The measurement of recidivism is quite challenging. One key issue is establishing an operational time period over which recidivism will be assessed; however, clarity and transparency in operationalization have allowed researchers to demonstrate relatively consistent findings. For instance, Sherman, Strang, Mayo-Wilson, Woods and Ariel (2014) adopted a 2-year operational definition of recidivism and found that offenders who completed restorative procedures were less likely to reoffend than offenders who completed non-restorative procedures. Many program evaluations have produced similar findings, demonstrating a reduction in recidivism for offenders engaged in restorative justice within relatively short follow-up periods (e.g., Bergseth & Bouffard, 2007, 2012; Bradshaw, Roseborough, & Umbreit, 2006; Braithwaite, 2007). A number of meta-analyses support similar conclusions across offender types. For instance, restorative procedures perform at least as well as, if not better than, retributive justice responses with regard to recidivism rates among both juvenile and adult offenders (e.g., Latimer et al., 2005; MacKenzie & Farrington, 2015; Sherman & Strang, 2007). Most recently, a meta-analysis of 21 studies exploring the effect of restorative procedures on juvenile recidivism rates concluded that restorative programs have a beneficial effect for youth, as evidenced through longer desistance periods between reoffending than youths directed to traditional court procedures (Wong, Bouchard, Gravel, Bouchard, & Morselli, 2016). Particularly interesting is that even restorative procedures that involve very minimal or indirect contact between victims and offenders were more effective at reducing recidivism than court-based procedures (suggesting that the success of restorative procedures might not hinge on the direct interaction between offenders and victims—allowing greater opportunities for restorative procedures to be engaged). Overall, these evaluations provide substantial evidence in support of the association between offender participation in restorative procedures and recidivism reduction.

Limitations of existing research: While the findings highlighted above have concentrated on the optimistic outcomes of restorative justice, empirical work also demonstrates limitations of existing research. For instance, returning to the importance of the operational definition of recidivism adopted, the observed effects of restorative relative to retributive models dissipate when a longer time frame for recidivism is considered. In particular, Bergseth and Bouffard (2007) found that when recidivism was assessed over a 4-year follow-up period, differences between restorative and court-based interventions were no longer significant.

Likewise, Tyler, Sherman, Strang, Barnes, and Woods (2007) demonstrate that differences between restorative and court-based procedures are not statistically significant so long as key mechanisms (i.e., reintegrative shaming and procedural justice) are incorporated into the design of each; this casts doubt on longstanding assumptions that retributive and restorative orientations produce fundamentally different outcomes. Instead, this finding suggests that procedural treatment makes all the difference in outcome effects. While a substantial body of accumulated research suggests that restorative justice is, in many ways, superior to the formal criminal justice system in responding to crime, it is important to remain critical of the actual nature of these effects.

A particularly important consideration is the existence of a self-selection bias , which might strongly skew the results of restorative justice research. Victims and offenders who take part in restorative procedures choose to do so, suggesting that these persons could be fundamentally different than victims and offenders engaged with court-based procedures. The voluntary nature of restorative procedures coupled with the fundamentally different starting point of offender acknowledgment of harm might set the stage for participants in restorative procedures to interpret justice quite differently from those in court-based procedures (Latimer et al., 2005). In particular, stakeholders in restorative justice might simply be more motivated to reach a resolution to the criminal act, providing baseline conditions that promote the realization of the intended effects of restorative procedures (Braithwaite, 2016). As opposed to the formal criminal justice system, which is primarily engaged with imposing corrections upon offenders, restorative justice strives to seize opportunities to encourage an existing desire for desistance in offenders (Robinson & Shapland, 2008).

It is imperative that scholars scrutinize the methodology of research that demonstrates the effectiveness of restorative relative to court-based procedures. Doing so will allow researchers to more comprehensively understand and, ultimately, to better predict the outcomes of specific justice processes for specific persons. Theoretical explanations are essential to developing these better understandings , and two dominate the restorative justice literature: reintegrative shaming and procedural justice.

Explaining the Effects of Restorative Justice: Theoretical Foundations

Alongside research evaluating restorative justice programming, a considerable body of work has been dedicated to explaining the effects observed. Given that both restorative and retributive justice aim to produce meaningful consequences to crime, the focus for most researchers has been the procedural differences associated with their application. While court-based procedures can be somewhat dismissive of the perspectives of both victims and offenders, restorative programs have been recognized as more adequately attending to stakeholders’ perspectives through the theoretical principles of reintegrative shaming and procedural justice. At their core, both of these theories advocate for the importance of process, maintaining that critical outcomes of justice responses (e.g., participant perceptions of satisfaction and legitimacy, as well as reoffending behavior) are highly dependent on the means used to administer justice.

Reintegrative Shaming

Although the formal criminal justice system operates on principles of proportional retributive justice , criminological theory suggests that this model might be detrimental to the offender. In particular, it has been argued that offenders experience harmful stigmatization as a result of encounters with the criminal justice system—particularly those which result in the administration of punishment—through a process of labelling . In other words, to be labelled as a “criminal” prompts society to reject offenders, identifying them as different and treating them accordingly (Becker, 1963; Links, Cullen, Frank, & Wozniak, 1987).

The origins of reintegrative shaming theory lie in the recognition that retributive responses to crime can be detrimental to offenders specifically, and to society more generally. In particular, a harmful form of stigmatization results from determinations of guilt levied in the formal criminal justice system, and this stigmatization can make reintegration into mainstream society difficult by way of a labelling effect (Braithwaite, 2000; Braithwaite & Mugford, 1994; Maruna, LeBel, Mitchell, & Naples, 2004). A function of stigmatization in retributive justice models is to connect the commission of a criminal act with feelings of shame—feelings which are seen as essential to foster future desistance from crime. However, the effects of shame are argued to be dependent on the structure of its administration, with stigmatizing shaming actually increasing subsequent offending (Braithwaite, 1989, 2000). Shame that is stigmatizing is grounded in messages of degradation, humiliation, and a lack of forgiveness; the offender is recognized as a bad person as indicated by his bad behavior. Consequently, the offender’s social reintegration is made more challenging as the individual is likely to perceive, as well as actually experience, exclusion on the basis of the criminal label (Braithwaite, 2000; Braithwaite & Mugford, 1994; Maruna et al., 2004). By contrast, restorative procedures recognize the value of shame but avoid engaging it through stigmatization by employing a model of reintegrative shaming.

Reintegrative shaming seeks to communicate shame for the harmful act while maintaining a position that is respectful to the offender and demonstrative of a willingness to forgive him (Braithwaite, 1989, 2000; Braithwaite & Mugford, 1994). As opposed to producing damaging consequences, shame incurred through a reintegrative process aims to encourage self-improvement, relationship restoration with those harmed by their actions, and inclusion with rather than exclusion from the community. Reintegrative shaming requires a different perspective from the offender—namely, that she is willing to accept responsibility for her behavior and acknowledge it as harmful—but also involves a fundamentally different structure of shaming, during which the offender (along with other stakeholders) is empowered with process control (Braithwaite & Mugford, 1994; Tyler et al., 2007). Reintegrative shaming is not an isolated component of the restorative justice conference; rather, it is an underlying philosophy that guides the procedure. This structure of shame is argued to work because it draws together people respected by the offender to disapprove of her behavior constructively, allowing the offender to recognize that, while her action was wrong, she is still valued (Braithwaite, 2000). Reintegrative shaming is a core foundation of the practice of restorative justice.

Restorative justice procedures endeavor to foster future desistance from crime by constructively conveying the harmful implications of a specific criminal act. This process ideally engages feelings of shame, but not stigmatizing shame. Instead, the various consequences of a crime (e.g., emotional, physical, material) are conveyed by stakeholders civilly in a way that is intended to genuinely compel the offender to avoid repeating the behavior in question. In these procedures, personal denunciation is ideally avoided; constructive disapproval of the offender’s behavior, rather than the offender on the whole, is the objective (Braithwaite, 1989). Communicating the harmful consequences of behavior is intended to prompt feelings of shame on the part of the offender, but the communication of this information also ideally conveys care and support for the offender (Braithwaite, 2000). In these ways, restorative procedures embody the essence of reintegrative shaming and, as such, have been hypothesized to produce positive outcomes for offenders (Johnstone, 2002).

Empirical investigations have validated reintegrative shaming as a promising technique for responding to crime. Barnes et al. (2013) suggest that incorporating a reintegrative shaming philosophy into restorative procedures enhances the overall effectiveness of the response to crime, particularly with regard to participants’ perceptions of satisfaction with the justice process. A key aspect of this satisfaction from the perspective of the offender is the extent to which procedures utilizing reintegrative shaming articulate respect for the offender. Procedures that promote the communication of respect for offenders foster positive relationship development for conference stakeholders (Ahmed, Harris, Braithwaite, & Braithwaite, 2001). Tyler et al.’s (2007) work provides a further test of this relationship. Adopting the starting point that reintegrative shaming strengthens the offender’s relational ties to significant others, the authors hypothesized that offenders taking part in restorative procedures would be less likely to reoffend than those in court-based procedures because offenders in restorative programs would be more concerned with negatively affecting the relational bonds they had established as a result of the process. Using longitudinal data drawn from a portion of the Reintegrative Shaming Experiments (RISE) program in Australia that focused on offenders involved in drunk driving incidents, recidivism rates for offenders who were directed to court-based procedures were compared against those who were diverted to restorative programming after 4 years. The results provided evidence in favor of the use of reintegrative shaming practices; specifically, that the use of reintegrative shaming techniques in justice responses reduced recidivism rates (Tyler et al., 2007). Regardless of whether an offender was assigned to restorative or court-based procedures, if the procedure incorporated reintegrative shaming, lower rates of reoffending were observed than if the procedure did not incorporate reintegrative shaming.

It might be the case that reintegrative shaming promotes the development of social bonds that encourage abstaining from crime. Alternatively, the effectiveness of reintegrative shaming in reducing offending might be due to the enhanced perceptions of respect and fairness reported by participants in such procedures (Mazerolle et al., 2013). The role of fairness perceptions in restorative justice can be better understood by turning to the procedural justice literature.

Procedural Justice

While reintegrative shaming theory is directly linked to the restorative justice literature, procedural justice theory has a broader history. In the legal context, procedural justice is concerned with evaluations of the application of law; specifically, the extent to which a procedure, as opposed to an outcome, is perceived as fair and satisfying (emphasizing the subjective interpretations that define perceptions of justice; Tyler, 1989). Procedural justice theory rests on the assertion that perceptions of the extent to which an outcome (e.g., consequences for criminal behavior) is evaluated as fair and satisfying rest heavily on the procedure used to determine the outcome. This is demonstrated through a wealth of psychological literature exploring the relationships between processes and outcomes (e.g., Lind & Tyler, 1988; Thibaut & Walker, 1975; Tyler, 1989).

Relational models of procedural justice contend that procedural features such as “voice” (i.e., the opportunity to express one’s opinion) are influential because they imply that the participant is a valued member of the group overseeing the procedure (i.e., she is asked to provide input because she has an important contribution to offer; Lind, Tyler, & Huo, 1997). Relational interpretations of procedural justice provide an explanation for a consistent finding in the literature: Even when individuals know that the voice they are permitted during a procedure will have no effect on the outcome reached, ratings of fairness and satisfaction associated with the procedure and outcome are still greater than when there is no opportunity for voice incorporated into the procedure. In fact, even post-decision opportunities for voice lead to higher fairness evaluations than no voice (Lind, Kanfer, & Earley, 1990). This finding necessitates an explanation grounded in non-instrumental, or relational, concerns. Fundamentally, relational models posit that treatment is a way of conveying messages of status in social groups. When treatment by authorities demonstrates attention to relational concerns , subordinates’ feelings of in-group membership and value to the group are fostered, which in turn increases their fairness perceptions. Conversely, when treatment demonstrates disregard for relational concerns, this can lead to perceptions of exclusion, and consequently, decreased fairness perceptions (Bradford, 2014; Lind & Tyler, 1988).

Voice is not the only relational concern; it is one indicator among several process-oriented concerns that affect the extent to which a procedure, as well as its outcome, is determined to be fair and satisfactory (Lind et al., 1990, 1997; Platow et al., 2013). Although multiple relational models have now been developed to explain procedural justice effects (e.g., the Relational Model of Authority, Tyler & Lind, 1992; the Group Engagement Model, Tyler & Blader, 2000), these iterations all maintain the same core tenets outlined in the first model developed, the Group Value Model (Lind & Tyler, 1988), which still serves as the dominant relational explanation in the procedural justice literature.

Lind and Tyler (1988) posited that the extent to which a procedure appeals to central relational concerns dictates evaluations of fairness and satisfaction with not only the procedure, but its outcomes as well. In particular, three relational concerns were identified: (1) neutral and consistent treatment, (2) trust in administrator benevolence, and (3) interactions demonstrative of respect and dignity. All three variables independently affect procedural justice perceptions , and demonstrate procedural concerns that go beyond desires to wield control over outcomes (Tyler, 1989).

Perceptions of respect are derived from interpersonal interactions that are perceived as polite, dignifying and considerate of personal rights (Tyler, 1989, 1994; Tyler & Lind, 1992). Ultimately, disrespectful treatment conveys to the individual that he is a person of low status within the group in question, and can also imply the social standing of groups in relation to one another (Heuer & Stroessner, 2011; Tyler, 1989). Likewise, relational models of procedural justice have always incorporated “trust,” understood as the extent to which the decision-making authority is perceived as trustworthy. This aspect of procedural justice concerns involves evaluating the perceived intentions of the administering authority, specifically, the extent to which the authority is perceived as reasonable (Tyler, 1989, 1994). Perceptions of trust are based on evaluations of the benevolence of the authority’s treatment (Tyler & Lind, 1992), and are particularly influential in shaping perceptions of legitimacy because the inference of a benevolent disposition fosters the belief that an authority can be trusted in the long term (Hough, Jackson, Bradford, Myhill, & Quinton, 2010; Tyler, 1994). Finally, “neutrality” broadly refers to the “even-handedness” of a procedure (Lind & Tyler, 1988; Tyler, 1989, 1994). This concern is based on the extent to which treatment by an administering authority is perceived as unbiased, objective and administered with equality (Tyler, 1994; Tyler & Lind, 1992). Together, these relational variables shape procedural justice judgments .

The consequences of procedural justice perceptions underscore the need for procedural concerns to be taken very seriously in legal contexts. A plethora of research has demonstrated that when people feel that the relational concerns described above are acknowledged and implemented in a procedure, the procedure’s outcome is likely to be described as more acceptable (Mazerolle et al., 2013; Tyler & Degoey, 1995), more satisfying (Mazerolle, Bennett, Antrobus, & Eggins, 2012; Mossholder, Bennett, & Martin, 1998), and more legitimate (Sunshine & Tyler, 2003; Tyler, Degoey, & Smith, 1996; Tyler & Wakslak, 2004). Likewise, the administrator of the procedure also tends to be evaluated as more legitimate, and these legitimacy perceptions are of particular significance in legal contexts. Legitimacy engenders behavioral compliance to the specific outcome in question as well as to the future instructions of the administrator of the procedure more generally (Gibson, 1989; Sunshine & Tyler, 2003). In other words, responses to criminal behavior that are perceived by the offender as fair and legitimate are more likely to promote future desistance from criminality. For these reasons, procedural justice theory in general, and the Group Value Model in particular, have been explored as offering explanatory power in relation to the effects of restorative procedures.

Restorative justice models cater well to relational concerns . In particular, they allow the opportunity for respect, trust, and neutrality to feature prominently in the procedure (Morrison, 2006; Tyler, 2006). The underlying philosophy of the model—disapproving of criminal acts but recognizing offenders as worthy of redemption—lays the foundation for inclusive treatment that helps convey to offenders that they will be treated without bias, with respect, and that they can trust in the process. Although they are being sanctioned, this treatment demonstrates care for the offender rather than simply control, and recognizes the offender as a valued member of the community, which serves to affirm value consensus and shared group membership (Gromet & Darley, 2009; Okimoto, Wenzel, & Feather, 2009; Wenzel et al., 2008). Finally, restorative procedures bestow a good deal of process control on participants, establishing substantial opportunities for voice (Braithwaite, 1998; Tyler, 2006; Zehr, 1997), which, in turn, enhances procedural justice evaluations (Lind et al., 1990; Tyler et al., 1996). Cumulatively, procedures adhering to a restorative justice philosophy should be guided by foundations that are likely to enhance perceptions of procedural justice, and that are attentive to relational concerns in particular.

Translating Theory into Practice: When Is Restorative Justice Effective?

Reintegrative shaming and procedural justice offer sound theoretical explanations for the success and effectiveness of restorative justice models. Before proceeding further, however, we must note that the notion of “restorative justice models” suggests a dichotomous understanding of restorative justice: That justice interventions either are, or are not, restorative, and, depending on that label, are, or are not, successful. Some restorative justice scholars have argued that their findings are better interpreted through the realization that responses to crime exist along a continuum of “restorativeness” (Bolitho, 2012). The manner in which justice responses are enacted affects the positioning of any intervention on the restorativeness continuum as well as the attributes of success being prioritized. For instance, the restorativeness of a justice intervention is enhanced by the procedural inclusion of practices indicative of reintegrative shaming and stakeholder relationship restoration (Ahmed & Braithwaite, 2012); the provision and nature of opportunities for empowerment, community restoration and remorse (Braithwaite, 2002); and “other values such as storytelling, respectful listening, victim and support attendance, and apology” (Bolitho, 2012, p. 61). In terms of effectiveness, empirical findings support the prediction that justice interventions falling along the restorative end of the continuum reduce subsequent offending (Hipple, Gruenewald, & McGarrell, 2014).

It is worth noting that, while identifying and incorporating effective procedural aspects in restorative interventions is important, rigidly prescriptive procedures that reduce the flexibility associated with restorative responses to crime should be avoided. Procedural flexibility is a hallmark of restorative justice that cannot be eliminated (Ahmed & Braithwaite, 2012; Bolitho, 2012); it is this flexibility that will allow restorative interventions to remain relevant and satisfying to the greatest possible range of offenders, victims, and offence types. The valuable notion that restorativeness is not measured dichotomously, but instead, along a continuum is highly relevant here. In this sense, a procedure is not “restorative” as a result of adherence to a strict procedure; rather, a variety of flexible qualities (such as stakeholder participation and collaborative decision-making ) contribute to the relative restorativeness of procedures, and flexibility helps ensure that the intervention appropriately addresses the transgression and the needs of the stakeholders in question. Although restorativeness can be encouraged by striving for the inclusion of general components (e.g., empowerment, community restoration, remorse), it is not ideal to mandate them (Bolitho, 2012). Our objective is not to advocate for a singular restorative solution.

To demarcate justice interventions as either restorative or not fails to take into consideration the variety of components that would be undesirable to specifically mandate in a restorative justice procedure but that, nonetheless, might influence the overall restorativeness of the intervention. For instance, victim presence is assumed to be essential to a procedure being identified as restorative. However, interestingly, both experimental work (Saulnier & Sivasubramaniam, 2015a) and field work (Bouffard et al., 2016) have provided evidence that it might be viable to use less intensive forms of victim presence (i.e., presence of a victim representative or indirect mediation) and still attain positive outcomes for offenders in terms of their subjective evaluations of the procedure as well as reoffending behavior. Conceptualizing victim presence in alternative forms is an example of a procedural aspect of restorative interventions that complicates a singular notion of restorative justice, drawing attention to the value of a continuum of restorativeness.

Several other procedural features problematize a singular understanding of restorative justice. For instance, restorative interventions generally encourage communications conveying apology from the offender to the victim, and forgiveness from the victim to the offender. While the inclusion of these characteristics is highly flexible in every restorative procedure, empirical evidence has associated positive outcomes with these features. For instance, the issuance of apology has been associated with subjective benefits for conference participants in field research , including greater outcome satisfaction for victims (Dhami, 2012). Likewise, experimental laboratory work has demonstrated that offers of apology improve victims’ perceptions of offenders and diminish victims’ punishment recommendations (Jehle, Miller, & Maskaly, 2012). However, that does not mean mandating an offer of apology in a procedure would make it restorative. For instance, while Jehle et al. (2012) found that offenders who offered apologies were reacted to more favorably by victims than offenders who did not offer apologies, they also found that victims were sensitive to whether apologies were coerced or voluntary, with voluntary apologies producing the most favorable reactions to offenders. Further, experimental work has suggested that coercing offenders to offer an apology negatively affects the quality of the apology offered as well as offenders’ subjective evaluations of the procedure (Saulnier & Sivasubramaniam, 2015a).

Victims’ attitudes towards offenders will also determine the restorativeness of a procedure, and these attitudes are closely tied to offenders’ behavior and the structure of the decision-making process. For instance, offers of apology by offenders foster victims’ feelings of forgiveness and reduce desire for revenge in some circumstances (Fehr, Gelfand, & Nag, 2010; Jehle et al., 2012). Empirical work has demonstrated that the means of justice—whether the orientation of a procedure is fundamentally about punishment or consensus-seeking—strongly drives the achievement of forgiveness, with more restorative procedures more effectively engaging forgiveness (Wenzel & Okimoto, 2014). However, again, classifying a procedure as restorative on the basis of whether forgiveness was achieved is not appropriate. A restorative procedure that failed to foster the communication of forgiveness is not inherently non-restorative; rather, a procedure that incorporated the communication of forgiveness simply speaks to achieving some aspects of restorativeness.

In sum, restorative procedures must be recognized as multifaceted; even “core” aspects of restorative justice such as victim presence, apology, and forgiveness are better understood as positioning justice interventions along a continuum of restorativeness, rather than simply restorative or not. According to this logic, evaluations of the success of restorative justice would be more valid if they focused on evaluating the effectiveness of various, core procedural features that locate a process at one end or another of the restorativeness spectrum, rather than comparing the effectiveness of “restorative” versus “non-restorative” procedures, per se. Further attention should be devoted to identifying the procedural elements of restorative interventions that translate theory into effective action and evaluating the effectiveness of those procedural elements. However, the effectiveness of restorative justice in producing socially desirable outcomes in response to crime does not inherently align with success in terms of public support for restorative responses to crime.

Effectiveness Versus Acceptability: The Restorative Justice Dilemma

The notion of “success” is already complicated in justice interventions by the variety of ways in which effectiveness can be operationalized and the relative weight assigned to those outcomes (e.g., producing positive subjective perceptions of the experience for offenders and victims, reducing recidivism , punishing offenders). Restorative justice tends to fare better than the formal criminal justice system in terms of its effects on stakeholders’ subjective perceptions, such as respectful treatment (Braithwaite, 2002), procedural fairness (Umbreit, Coates, & Vos, 2001), procedural and outcome satisfaction (Latimer et al., 2005), process finality (Strang & Sherman, 2006), accountability (Regehr & Gutheil, 2002), value consensus between parties (Okimoto & Wenzel, 2009), remorse, and empathy (Choi & Severson, 2009). While these potential outcomes of restorative interventions are key to understanding the effectiveness and, thereby, success of restorative justice, equally important to the viability of such procedures are public perceptions of the appropriateness of restorative responses to crime.

Developing a better understanding of how support for various justice procedures is constructed, produced, or maintained requires recognizing dominant lay philosophical justice orientations. Retributive motivations are generally recognized as being the dominant philosophy engaged by laypersons in response to crime (Carlsmith & Darley, 2008; Cullen, Fisher, & Applegate, 2000; van Prooijen, 2010). Empirical research suggests that retributive motivations are simply a default standard among average members of the public responding to observed transgressions in Western nations. For instance, van Prooijen (2010) concluded that unbiased third party decision makers default to retributive motivations in response to crime, on the basis of two findings. First, that participants assigned greater financial penalties to hypothetical offenders when the payment was described as a punishment for their crime rather than as compensation for the victim; and, second, that participants were more attentive to information pertinent to offender punishment relative to victim compensation when asked to relay key details of a vignette describing a hypothetical justice intervention.

Researchers have explored the factors that might drive this general tendency towards retributive motives. Evidence suggests that certain characteristics associated with the offence can affect justice motivations. For example, offences committed by youth as well as transgressions that are non-violent are associated with greater desires for restorative oriented responses (Cullen et al., 2000). Alternatively, relationship bias—specifically, greater emotional proximity to a victim—produces the opposite effect, enhancing punitive oriented responses (van Prooijen, 2010).

Third party perceptions of appropriate justice responses are also associated with more symbolic concerns, such as perceptions of group membership, transgression meaning, and offender deservingness (Feather, 2006; Wenzel et al., 2008; Wenzel, Okimoto, & Cameron, 2012). Perceptions of group membership refer to an identification that a target person is an in- versus out-group member relative to oneself. Although the exact criteria used to make this judgment vary considerably depending on context (e.g., race , religious orientation, family membership), research suggests that when third parties to a transgression identify the offender as an in-group member, they are more likely to adopt a restorative response to crime (Wenzel et al., 2008). Alternatively, identifying the offender as an out-group member promotes retributive responses. Similarly, the symbolic meaning of the transgression itself influences perceptions of appropriate justice. Retributive motivations tend to be engaged when a transgression is seen as a violation of status/power expectations, whereas restorative motivations are more likely to be engaged when a transgression is seen as a violation of values (Wenzel et al., 2012).

Finally, and fundamentally, justice motivations are driven by the treatment the offender is judged to deserve. Perceptions of deservingness are a product of the relative accord between actions and outcomes (Feather, 1996, 2006; Lerner, 1980); in other words, in the context of criminal behavior, a variety of characteristics associated with the criminal act come to influence perceptions of the way in which the offender should be treated. Evaluations of intent, harm caused, and provocation are all key criteria used to evaluate offender behavior (Saulnier, Lutchman, & Sivasubramaniam, 2012). When offender behavior is recognized as particularly egregious (e.g., deliberate, harmful, and/or unprovoked), third parties do not see the offender as deserving respectful treatment (Heuer, Blumenthal, Douglas, & Weinblatt, 1999). Given that conveying respect for the offender is a core aspect of restorative justice, this finding suggests that members of the public might not see many offenders as deserving of a restorative intervention. Justice responses of a fundamentally retributive or restorative orientation prioritize different justice goals (Gromet & Darley, 2009), but neither response will be recognized as adequately satisfying the goal of justice when the offender is not seen as deserving the treatment received (regardless of the reason for that lack of congruence).

As noted earlier, differences in outcome measures between restorative conferences and court-based procedures were negligible, so long as the approach that was employed incorporated reintegrative shaming and demonstrated attention to procedural justice concerns (Tyler et al., 2007). Recognizing that the theoretical foundations of restorative justice have a place in primarily retributive procedures prompts reflection on whether there is a place for punishment in primarily restorative procedures. Philosophical starting points of restorative and retributive responses to crime suggest that the models are driven by fundamentally conflicting goals; however, it could be the case that striking a balance between the two is the best way to appeal to public justice motivations while also producing socially desirable outcomes. For example, empirical research demonstrates that the incorporation of opportunities for retributive outcomes in restorative procedures enhances third party perceptions of the appropriateness of restorative interventions as a response to serious crime (Gromet & Darley, 2006). Here, again, it is important to remember that restorative justice is not best understood dichotomously, but instead, along a continuum of restorativeness. Including the possibility for punishment does not inherently negate the restorative potential of a justice response (Gromet, 2012). Conceptualizing justice responses as simultaneously capable of serving restorative functions while also appealing to retributive motivations through the possibility of punishment substantially broadens the scope of restorative interventions by appealing to a wider, lay audience.

Public support for restorative justice is based on a variety of factors independent of the effectiveness of restorative justice. Advancing the success of restorative justice involves not only implementing restorative interventions that are empirically demonstrated to be successful on various important outcome measures; it also involves engaging with public perceptions of acceptability. This requires a more complex approach to making restorative justice viable in the long term . One way to achieve this viability is to accommodate widespread public notions of deservingness and retribution within restorative procedures through the inclusion of opportunities for offender punishment. Another way to achieve this viability is to investigate those public notions of retributive justice with the goal of amending them to be more amenable to restorative interventions. This requires being attentive to people’s preferences for “just deserts” (Carlsmith & Darley, 2008) and the cognitive processes engaged during justice-oriented decision-making .

Cognitive Processing, Justice Reasoning, and Restorative Justice

Although outcomes such as stakeholder satisfaction and offender recidivism should be the key factors guiding the implementation of justice interventions , public perceptions of appropriate justice might, in fact, be among the most influential factors affecting the widespread application and long-term viability of restorative justice. Justice motivations shape satisfaction with justice interventions (Gromet, 2012; Okimoto & Wenzel, 2011; Wenzel et al., 2012). However, research has demonstrated that the justice motive comes in different forms, capable of producing quite different understandings of just outcomes.

The Justice Motive

We have already described the importance of procedural concerns as drivers of justice perceptions , but it is also important to consider research on distributive justice in order to understand perceptions of just outcomes among the public more broadly. In particular, the cognitive processing of responses to criminal behavior is divided into two streams; one relying on intuitive, heuristic processing , and the other on systematic, deliberative processing (Chaiken & Trope, 1999; Lerner, 2003; Sivasubramaniam, forthcoming). Though punishment is often central to both lay and legal notions of justice, the reasoning driving punishment decisions tends to be initiated through distinct cognitive processing mechanisms (heuristic versus deliberative processing, respectively; Chaiken & Trope, 1999; Gigerenzer & Gaissmaier, 2011; Lerner, 2003) that produce different renditions of justice.

Dual-process theories of cognition describe the processes underlying immediate, intuitive reactions to scenarios versus judgments based on some extended deliberation . Heuristic processing is an initial reaction, involving a more automated response (Chaiken & Trope, 1999; Lerner, 2003). In the justice context, this typically involves identifying a person directly responsible for causing a harm and desiring a consequence (most frequently punishment) for that wrongdoing. Systematic processing involves a more thorough review of available evidence. In the justice context, this typically involves overriding the reaction produced through heuristic processing to arrive at a more tempered outcome than that which was immediately desired. A fundamental difference between these systems is the extent to which they are automatically versus consciously engaged. Decisions reached as a result of heuristic processing are the product of scripted associations that reside within the individual (though these can be influenced by external factors such as cultural norms); whereas decisions reached as a result of systematic processing are the product of a conscious decision-makingeffort that involves considering a wider variety of variables, as well as determining the varied importance of those variables (Richetin, Perugini, Adjali, & Hurling, 2007). This fundamental variation in the rendering of judgments is essential to understanding how and why perceptions of appropriate justice are established, the ways in which legal and lay notions of justice vary, and finally, how support for restorative justice might be affected.

Legal notions of justice: Although the desire to restore a sense of justice following the commission of an offence is a goal of both the legal system and lay people, the processes the two groups employ prompt what actually constitutes a just outcome to be construed quite differently. Legal notions of justice are the product of complex and deliberative processing of information guided by procedural law (Krasnostein & Freiberg, 2013). This is a highly systematic means of determining just outcomes that fosters careful, sober and reflective second thought, actively striving to avoid the inclusion of emotional impulses. In practice, this involves a legal professional carefully considering a number of key factors that guide the legal production of just outcomes, such as ensuring adherence to sentencing principles and statutory law (i.e., abiding by legislated definitions of criminal offences including minimum and maximum consequences), as well as considering case-specific aggravating or mitigating factors alongside more generally prescriptive legal precedents. These factors are considered independently and in combination by persons specially educated and trained for this complex decision-making task: normally, judges.

The means by which sanctions are determined in restorative procedures are quite different from those employed in the formal criminal justice system, but they do share the fundamental feature of clearly involving deliberative, systematic reasoning. For example, participants in restorative interventions are encouraged to be reflective and think outside of their isolated position in the conflict when determining the harms that have been caused and how they should best be resolved, and they are guided by experienced facilitators in this process.

Lay notions of justice: The deliberative decision-making invoked in legal responses to crime involves a process of reasoning and reflection; by contrast, the intuitive decision-making process—which tends to be the first, if not the only, way in which lay notions of justice are rendered—is more akin to perception (i.e., not engaging with reasoning and reflection but, instead, more compulsory reactions; Haidt, 2001). The process is simple: A person learns of an offence and heuristic processing takes over, quickly establishing a sense of who is at fault in the interest of determining how the injustice can be rectified, with little regard for other relevant data such as contextual information. Without active attention to engaging in systematic processing , it is the heuristic system—and therefore, in the context of criminal behavior, retributive impulses —that dominate.

Heuristics function by actively ignoring substantial portions of relevant information (Gigerenzer & Gaissmaier, 2011), and while the judgments produced through this process might be accessible, the process itself is not (Haidt, 2001). Intuitive processing produces an outcome judgment by way of a process that is entirely opaque. Perhaps most troubling about judgments reached through heuristic processing is that perceptions associated with intuitive judgments are particularly difficult to challenge or reverse. These are visceral reactions, and although people tend to not be able to explain them well, they do tend to be highly invested in their accuracy (Haidt, Koller, & Dias, 1993; Haidt & Hersh, 2001).

Furthermore, intuitive reactions to descriptions of harm—particularly direct acts of harm between people—tend to frame justice responses primarily in terms of punishment (Carlsmith & Darley, 2008; Vidmar & Miller, 1980). Heuristic processing involves the production of a simple script in which emotions drive cognitive reactions to offending, prompting a tendency towards anger and the prioritizing of punishment. Several explanations exist for this effect: ideological preferences, instrumental motivations, and relational motivations (Gerber & Jackson, 2016). The first, ideological preferences, simply suggests that different people view the world differently; in particular, that persons who rank highly on scales of authoritarianism and conservatism are more likely to support punitive justice responses (Gerber & Jackson, 2013, 2016; Tyler & Boeckmann, 1997). The second, instrumental concerns , suggests that the fear of victimization drives more punitive responses as a strategy to reduce future exposure to harm by incapacitating known offenders as well as generally deterring others from engaging in crime (King & Maruna, 2009; Tyler & Boeckmann, 1997). Finally, the third explanation is grounded in relational concerns associated with interpersonal bonds. This explanation suggests that more punitive responses are a strategy of maintaining moral boundaries in response to community breakdowns (Carlsmith, Darley, & Robinson, 2002; Darley, Carlsmith, & Robinson, 2000; Tyler, 2006). Through this lens, the offender is fundamentally seen as someone who deserves to be punished. This sentiment is much more in line with the retributive orientation and stigmatizing structure of shame common to the formal criminal justice system than to the broader scope of harm response and the reintegrative structure of shame common to restorative procedures.

While explanations for the retributive impulse vary, the impulse is recognized as widespread and fundamental. It appears to be a product of a basic human response that is unlikely to vary considerably across persons on the basis of characteristics such as demographic variables. However, there is a need to continue to explore factors associated with variations in this impulse. For instance, actual victims of crime are more satisfied with responses to crime that are less punitive than more removed observers, demonstrating that we tend to react more punitively to observed, rather than personally experienced, harm (FeldmanHall, Sokol-Hessner, Van Bavel, & Phelps, 2014). Developing a greater understanding of the individual and situational variables that influence the retributive impulse is necessary to continue advancing understanding of this impulse, which can be usefully applied to improving subjective and objective outcomes of responses to crime.

While heuristic processing tends to be fundamental in establishing lay notions of justice, it must also be acknowledged that members of the general public are capable of engaging in deliberative processing , particularly when the emotions that initially compelled intuitive responses dissipate. Deliberative processing simply involves a more thoughtful evaluation of information relevant to the decision-making process (such as deservingness, culpability, and considering multiple forms of recourse), all in the interest of establishing the most appropriate justice response possible. Importantly, engaging deliberative processing can lead to different outcome judgments than those initial responses produced through heuristic processing (Haidt, 2001). In the context of evaluating criminal behavior, engagement solely with intuitive reasoning tends to be associated with a retributive impulse to punish the offender. Engaging with more reasoned, deliberative processing fosters a sober-second thought to retributive impulses , making it more likely that responses to criminal behavior will consider a broader range of socially desirable outcomes; notably, adherence to the principle of due process, but potentially more restorative responses to crime as well.

Unfortunately, the cognitive resources required to move beyond simple heuristic processing mean that transitioning to deliberative processing does not automatically follow initial, intuitive responses. Characteristics unique to the individual (e.g., personal cognitive capacities or tendencies to avoid deliberative reasoning) as well as the situation (e.g., limited time or resources) will influence one’s ability and likelihood of engaging in deliberative processing (Sivasubramaniam, forthcoming). As such, while judges responsible for formal decision-making in legal contexts are required to engage in deliberative processing that overcomes their initial heuristic response, there are no guarantees that laypersons will find themselves in a situation, or with the skills, to do the same.

The dominance of the retributive impulse among lay people might seem somewhat inconsequential; after all, it is judges and not laypersons who determine the consequences levied on offenders (though laypersons do determine consequences in some situations, such as capital punishment and some serious felonies in the USA). However, while a reasoned, deliberate approach to decision-making dominates responses to crime in the formal criminal justice system, public opinion is still highly influential in shaping the parameters within which legal decision makers operate (e.g., in shaping the legislation governing judges’ determinations; for a broader discussion of the importance and pitfalls of relying on community sentiment in lawmaking, see Miller & Chamberlain, 2015). To the extent that public opinion is driven by intuitive, heuristic (and therefore retributive) approaches, formal legal responses will tend to adopt a fundamentally retributive orientation. This retributive impulse among lay people also poses a serious and specific challenge to advocates of restorative justice: restorative responses to crime adopt a much more divergent position from lay notions of justice than the formal criminal justice system and, therefore, will be less likely to receive the public approval necessary to become a widely deployed justice response. However, advocates of restorative justice might find productive paths in psychological research addressing ways to either override or amend the retributive impulse .

Restorative Justice and the Retributive Impulse

The key question we are left considering is: “Can restorative justice serve a sense of justice if it does not centralize the imposition of punishment on the offender?” (Sivasubramaniam, forthcoming, p. 151). We believe the answer is “yes,” but that achieving this sense of justice requires careful attention to the relationship between cognitive processing mechanisms and the retributive impulse . Support for restorative justice can be increased in two distinct ways that take advantage of this relationship. The first path would be to encourage deliberative rather than heuristic processing ; and the second would be to amend the scripts contained in people’s justice-related heuristics, so that when heuristic processing is engaged, restorative (rather than retributive) responses to transgressions are primed.

First, increases in the use of deliberative processing by the public can be encouraged. Research suggests that people will default to heuristic processing unless encouraged to engage in deliberative processing (Simon, 1967). Therefore, ways to prompt deliberative processing of justice-related information need to be explored. In a very general sense, this is something that should be advocated for as central to the learning process; in other words, children, from a very young age, should learn about decision-making in emotionally complex situations through school-based curricula that differentiate between “fast” and “slow” thinking (Kahneman, 2012) and provide a series of steps, akin to the scientific method, for encouraging deliberative processing when faced with such situations. In the more specific sense of encouraging systematic reasoning once a person is exposed to a justice scenario, research programs that address the question of how to generate deliberative processing in emotional, justice-related scenarios that would normally evoke intuitive, heuristic responding need to be developed.

Without centralizing punishment, it might be assumed that restorative justice cannot achieve a sense of justice that satisfies lay notions of justice. However, research demonstrates that a variety of factors influence support for restorative justice, indicating that if deliberative cognitive processing can be engaged, then a desire for punishment is only one factor among several that people will consider when establishing their support for restorative interventions. When systematic processing is engaged and a wider variety of relevant information is taken into consideration, restorative justice tends to be acknowledged as meeting a wider array of desirable justice outcomes than justice responses with a primarily retributive orientation. In particular, other dimensions of success, such as greater opportunities for stakeholder inclusion (Moore, 2012), enhanced victim restoration and satisfaction (Bazemore, 1998; Latimer et al., 2005; Strang et al., 2006), and more successful offender reintegration into the community as well as reduced recidivism rates (Ahmed & Braithwaite, 2012; Braithwaite, 2002), become more central in determining people’s support for restorative interventions. The existing evidence certainly suggests that victims’ and observers’ perceptions of the appropriateness of restorative responses to crime are improved when this wider array of factors is taken into consideration (Gromet, 2012), an outcome which can be partially credited to their own engagement with deliberative processing . However, it should be noted that systematic processing does not necessarily, or wholly, negate a desire for punishment; and research has demonstrated that restorative interventions tend to be evaluated more favorably when they include opportunities for retribution as a possibility (Carlsmith et al., 2002; Gromet & Darley, 2006). As such, designing restorative interventions that permit the possibility of retributive outcomes but foster deliberative processing could simultaneously appeal to retributive impulses , while not necessarily seeing the retributive impulses realized.

Second, heuristic processing can be harnessed to garner support for restorative justice. Specifically, better understanding the intuitive reasoning process associated with crime and justice responses might make it possible to alter the heuristic that tends to produce a retributive impulse. Essentially, this would involve reprogramming the basic scripts/associations that people hold so that heuristic processing no longer leads to the retributive impulse but, instead, leads to support for restorative responses to crime. To understand how to achieve this goal, we turn to the social psychology research on stereotypes .

In the social cognition literature (e.g., Schneider, Hastorf, & Ellsworth, 1979), stereotypes are employed to minimize the use of cognitive resources. In this sense, stereotypes are a form of heuristic processing reliant on simple associations to quickly reach decisions (Kahneman & Tversky, 1972). In the context of crime and justice responses, the simple association is: crime equals punishment. Punishment is an automatic association with crime and, so, it is overweighted in responses to justice produced through intuitive processing . As a result, we propose that the problem of heuristic processing producing retributive impulses can gradually be overcome by rescripting automatic responses to crime to: crime equals resolving harm.

There is some direction in the literature for rescripting heuristic processing . Bordalo, Coffman, Gennaioli, and Shleifer (2016) present empirical findings demonstrating that perceptions of greater representativeness foster the creation of stereotypes (in other words, the more commonly associations are seen, the more likely stereotypes based on those associations will be formed). This presents a difficult problem, as this finding suggests that retributive legal justice responses beget punitive lay notions of justice—bearing in mind that fundamental changes to the criminal justice system (such as a transition away from a primarily retributive orientation) require public support, thus producing a catch-22. However, social and economic circumstances can be leveraged in this regard. For example, nations adopting highly punitive justice responses, such as the United States , now find themselves driven to urgently change their punitive incarceration policies out of financial necessity. The monetary burden of sustaining large numbers of prisoners for extended periods of time is simply too great for the economy to bear (Shannon, 2015). As such, there is a window of opportunity to put forward a restorative agenda that can appeal to the general public through lowered relative cost, and which promotes alternatives to simply punishing offenders through extended incarceration. Research has demonstrated that justice concerns are, in part, influenced by people’s perceptions that they will also be negatively affected by the decision (e.g., losing desired resources; Greenberg & Cohen, 1982; Lerner, 2003; Steensma & Vermunt, 1991). Therefore, cognitive rescripting that effectively associates administering punishment to others with harmful consequences for oneself (such as diminished resources) could be useful in diverting the association between criminal acts and retributive impulses. On a broader level, this would equate to an incremental, implicit change in the norms of society.

Advancing the Science: The Value of Psychological Research

Developing a more meaningful understanding of restorative justice necessitates continued research. In particular, investigations adopting a psychological approach would be highly valuable (Sivasubramaniam, 2012). Psychological research is highly useful for identifying variables relevant to the initiation, workings, and outcomes of restorative procedures, all of which are valuable in establishing when and why restorative justice is a viable response to crime. Psychological investigations advocate for the systematic exploration of variables associated with restorative justice, which is particularly suitable for better understanding how support for restorative interventions can be established and maintained. To this end, attention to cognitive processing is an essential and underdeveloped area.

There are a variety of methods that can be used to gain a more thorough understanding of the psychological mechanisms underlying restorative justice processes, but experimental work is particularly lacking in the restorative justice field. Experimental designs prioritize the isolation and manipulation of variables in a controlled setting, in the interest of determining their effects. This is an especially useful strategy for developing understandings of the existence, direction, and strength of relationships between variables in restorative procedures (Ahmed & Braithwaite, 2012; Dhami, 2012; Saulnier & Sivasubramaniam, 2015b). Experimental designs offer a means to test hypotheses while striving to eliminate the influence of confounding variables. As such, experimental methods are capable of generating knowledge of causal relationships between variables (Cosby, 1977; Salkind, 2006), which can be particularly compelling when results are replicated in both laboratory-based and field-based tests. Experimental investigations of justice-related cognitive processing mechanisms would be a valuable new direction for restorative justice researchers to pursue, offering considerable practical value in terms of better aligning restorative justice with lay notions of justice (either by amending the characteristics of restorative justice or by amending the architecture of lay notions of justice). In preparation for experimental field research , we suggest beginning with simple experimental laboratory work exploring cognitive processing that would allow relevant factors of interest surrounding restorative interventions to be controlled, isolated, and manipulated.

For instance, above we suggested exploring whether heuristic cognitive processing can be rescripted. A simple experimental laboratory design might invite participants to read a short vignette detailing a criminal transgression in the interest of subsequently assessing retributive impulses through questionnaire items, manipulating participants’ pre-experiment intuitive associations (and, thereby, their heuristic processing ). Specifically, participants could be assigned to one of three conditions (restorative, retributive, control) in which they are given details of, and encouraged to imagine, justice responses that are either primarily restorative or retributive, depending on their condition. (The control condition would provide a comparison to assess the general effect of priming.) Significant reductions in retributive impulses following exposure to the restorative condition would suggest that heuristic processing is capable of being reprogrammed in the short term (mirroring work suggesting that heuristic processing associated with stereotypes can be rescripted; Blair, Ma, & Lenton, 2001), and on the basis of such initial investigations, larger scale research programs could investigate the reprogramming of heuristic scripts in the longer term.

We also suggested exploring whether deliberative processing can be encouraged among lay people, to decrease retributive impulses. A simple, laboratory-based experimental paradigm could measure retributive impulses after manipulating the extent to which participants were permitted to “rush to judgment.” Imposing different time limits and other criteria associated with the participant’s decision-making process would allow for different styles of cognitive processing (i.e., heuristic versus deliberative) to be prompted. For example, participants could be randomly assigned to one of three conditions (heuristic, deliberative, control), given a short vignette to read detailing a criminal transgression, and asked to offer their decision about the most appropriate justice response from pre-crafted options (ranging from highly retributive to highly restorative in orientation). In this case, significant reductions in retributive impulses following exposure to the deliberative condition would suggest that actively encouraging deliberative processing is a means of influencing lay notions of justice to be less retributive.

In general, research on dual-process models of justice reasoning will allow us to better understand variations in support for justice interventions , and how support for restorative interventions can be increased. However, further work is also needed to improve understanding of the mechanisms at work within restorative procedures. While it is generally established that primarily restorative procedures are more successful than primarily retributive procedures across a variety of measures, advancing the current state of knowledge requires dissecting the success of restorative procedures at the operational level—what works, when, and for whom? A prime example of a procedural feature that requires further operational investigation is the issuance of an apology by the offender to the victim. Although apology has long been recognized as central to successful restorative procedures, empirical work exploring the actual effect of apology on desirable outcomes of restorative justice has been limited. The evidence that does exist suggests that apologies are associated with beneficial outcomes of restorative procedures, such as victim satisfaction (Dhami, 2012), perceptions of offenders and inclination towards punitive responses (Jehle et al., 2012), as well as offender perceptions of process finality (Saulnier & Sivasubramaniam, 2015a), but further research is needed to establish the conditions under which apology is a beneficial procedural feature.

Conclusion

In this chapter, we provided an up-to-date review of key information pertaining to the psychological study of restorative justice, closing with important future directions for researchers to pursue in this domain. We identified distinct philosophical foundations, structural applications, and outcomes of restorative justice models and the formal criminal justice system. We then discussed evidence-based policy as it pertains to restorative interventions—recognizing this as key to the development and administration of any successful response to crime. We reflected on the findings of program evaluations, theoretical explanations of effects, and linkages between theory and practice. We established the need to be clear about the multifaceted nature of restorative justice, as well as the multiple dimensions of success in such interventions. Crucially, we also noted that effectiveness in producing desirable outcomes in response to crime does not inherently align with success in terms of public perceptions of support for restorative responses to crime.

We discussed the crucial role of public perceptions of justice in shaping the viability of restorative justice, noting that the widespread retributive impulse dominates lay notions of justice. We argued that restorative justice can serve a sense of justice without prioritizing the punishment of offenders, but we reached two key conclusions in this regard: (1) that heuristic processing leads to retributive impulses, which does not promote a favorable response to restorative justice; and (2) that deliberative processing tends to temper retributive impulses, encouraging greater consideration of a wider variety of information, including more restorative responses to rectifying harm. Heuristic processing , therefore, should be associated with diminished support for restorative justice, and deliberative processing should be associated with enhanced support for restorative justice.

We considered two ways in which we might draw on the psychological literature to amend the retributive impulse : first, encouraging increases in the use of deliberative processing by the public in justice-related scenarios; and second, harnessing heuristic processing to garner support for restorative justice by disrupting the retribution-oriented script and replacing it with a restoration-oriented script. Finally, we advocated for continued research adopting a psychological orientation as key to advancing the science, noting experimental designs as particularly valuable in establishing a better understanding of causal effects and particularly well suited to investigating cognitive processing mechanisms and how they relate to the justice motive.

In conclusion, we note that we are undergoing an important, new stage in the development of restorative justice research. Early work in this field drew attention to the promises of restorative justice and laid the foundation for theoretical explanations of differences between restorative and retributive responses to crime. A second wave of research built on this groundwork through data-driven contributions offering empirical evidence of the distinctions between the mechanisms and outcomes of restorative interventions and court-based procedures. Researchers are now moving beyond replicating what is already fairly well established in the restorative literature, and are focusing on producing research that will continue to foster reform in the formal criminal justice system via more restorative procedures; however, regardless of the effectiveness of restorative justice, the success of restorative justice might be limited by its alignment (or lack of alignment) with the justice notions of the general public. The new wave of restorative justice research must not only improve the effectiveness of restorative justice; it must also enhance the success of restorative justice (in terms of public support for its expansion across the justice system and its long-term viability), through understanding, appealing to, and managing public notions of justice.