Abstract
Introduction
A dominant ideology in today’s world is that of risk – before taking action, each individual actor should weigh up the risks of that action and perform what might be seen as an actuarial calculation of the costs and benefits of doing it (Beck, 1992). Associated with this key idea of risk is that of prevention, that it is possible and indeed desirable to prevent harm occurring by taking the right decisions to address specific ‘risk factors’. In relation to crime, this view of acts and the responsibility of those charged with taking decisions on those who may offend or have offended has become a key part of the work of criminal justice personnel, such as the police, prosecutors, penal authorities and the judiciary (Kemshall, 2003). It is also, as we shall see, a key part of the work of those providing and facilitating restorative justice, but the ways in which such risk decisions are approached, the power balance of those involved in them, and the relevant factors seem to be different in restorative justice. The aim of this article is to explore those differences and to reflect back on criminal justice as to whether there are different potential paths to risk assessment relevant to criminal justice as well.
Broadly speaking, in relation to criminal offending, there are two distinct areas of risk calculation: (1) decision-making for those who commit criminal behaviour and (2) professional assessment of the risks of further offending. For those weighing up the first of these – whether or not to offend – the theory of rational choice has set out a model of such decisions, whereby benefits such as attaining sought-after outcomes (e.g., money, power or status), are seen as being weighed against possible negative outcomes, such as the chance of being caught and receiving sanctions (Cornish and Clarke, 2014). There have been critiques of these ideas of risk, choice and decision-making. In particular, it has been queried as to how conscious these processes of decision-making are, at least for those thinking about offending, with violent offences seen as potentially less subject to such rational decision-making (Felson, 2015). Rational choice theory has also been critiqued as putting undue emphasis on the salience of potential punishment as deterring offending, and as potentially tautological, in that it may promote seeing a crime having occurred as necessarily the result of its benefits outweighing its disadvantages for the person committing it (Steinmetz and Pratt, 2024). Equally, rational choice theory sees these decisions being based primarily at the individual level, so it is individual offenders or criminal justice decision makers who are seen as key.
Ideas of risk, following the work of Beck (1992), have similarly been placed predominately at the individual level. Though of course there is often interaction between individual, group and societal factors, risk from such a theoretical viewpoint is seen as how the individual weighs up and interprets all these elements. Some critics of rational choice theory have stressed, however, that societal factors, such as poverty, deprivation and income distribution, constrain people’s opportunities, choices, and perceptions of their options. Proponents of rational choice theory have countered that what matters in individuals’ decision-making is their own perceptions of their circumstances, including their social context, rather than some abstract idea of costs and benefits. They would say that individuals have both relatively stable decision-making traits but also take into account the particular circumstances of each crime (Nagin and Paternoster, 1993).
Beck himself, probably the most influential theorist in relation to perceptions of risk, saw society as moving increasingly towards an individualisation of how each person sees risk because of the ‘gradual loosening of the structures and networks of tradition’ (Mythen, 2004). Beck (1992) commented: Individualization in this sense means that each person’s biography is removed from given determinations and placed in his or her own hands, open and dependent upon decisions. The proportion of life opportunities which are fundamentally closed to decision-making is decreasing and the proportion of the biography which is open and must be constructed personally is increasing. Individualization of life situations and processes thus means that biographies become
There are, however, consequences to this individualisation. For the individual, it is a double-edged sword: ‘One blade cuts greater choice and autonomy, the other carries the burden of continual decision and responsibility’ (Mythen, 2004: 119). In relation to crime and criminal justice, this is one of the factors which have been driving a greater perceived responsibilisation in Western criminal justice, whereby those who have offended against the criminal law are seen as bearing individual responsibility and so fault and guilt for the harm done (see, e.g., Barry, 2013; Kelly, 2001).
In criminal justice decision-making, the second area of risk calculation mentioned above, the idea of needing to consider risk has become increasingly dominant in decisions about prosecution, sentencing and where to place people in the penal system (and whether to release them) (Kemshall, 2003; O’Mahony, 2009; O’Malley, 2004). However, it has not been accompanied, we would argue, by a realisation of its implications in relation to individualisation – despite the highly individual-focussed character of ‘responsibilisation’ in contemporary Western criminal justice. Instead, possibly primarily as a matter of institutional convenience, risk is being measured through the individual being seen as a member of a group of people with particular characteristics, all of whom will behave in similar ways. This group-based risk assessment can be seen, for example, in the calculation of OGRS (Offender Group Reconviction Scale) scores for individuals in relation to their likelihood of reoffending in probation, prison and youth justice practice (Howard et al., 2009). The OGRS score comes from the calculation of the reoffending rates of a group of offenders and is ‘the standard method of predicting reoffending in the Probation Service of England and Wales’ (p. 1). The professional assessment of risk of reoffending is therefore undertaken at a different level from the individual decision-making regarding criminal behaviour. In this article, we argue that this professional model of risk is flawed and does not accord with what we know about how to achieve lower risk of harm. We look to restorative justice as portraying a different view of risk assessment in the criminal justice context, which might profitably be adopted in criminal justice decision-making.
First, though, we need to explore what kinds of risk and outcomes are seen as particularly valuable in the criminal justice context, before examining the risks in restorative justice and comparing methods of risk assessment in the two domains.
Risk in criminal justice
In criminal justice in England & Wales (and well beyond), risk and risk management are now dominant paradigms. So, for example, the Probation Service (2023) defines itself as a ‘statutory criminal justice service that supervises high-risk offenders released into the community’, where ‘Our priority is to protect the public by the effective rehabilitation of offenders’. Public protection is based on risk assessment and management, particularly the risk of reoffending. In England & Wales, as in many other countries, risk assessment is based on an RNR model (risk-need-responsivity: Andrews et al., 1990), in which effective rehabilitation is based on assessing those who have offended in relation to their risk of reoffending, and their ‘criminogenic needs’ (needs that are thought to be functionally related to criminal behaviour) and then interventions targeted according to those needs and the individual’s learning style and abilities (responsivity).
Even in relation to interventions based on public health perspectives, such as drug treatment or violence programmes, the type of programme used is determined according to a utilitarian view of looking at the whole population and judging initiatives in relation to the risk posed by the behaviour. Essentially, we see taking a public health perspective as looking at the effect of an initiative on the whole population, that is, to provide the maximum benefit for the largest number of people. It can also mean emphasising preventive aspects rather than post-incident, reactive perspectives, but we are here emphasising its group-based nature. For this, what should be done with the individual is determined according to that individual’s perceived future risk of engaging in harmful acts and contributing to a negative balance in relation to the harm, based on their membership of certain categories, coded and systematised as ‘risk’ and ‘protective’ factors. The public health approach to, say, violence prevention embodies four steps: defining and monitoring the problem; identifying risk and protective factors; developing and testing prevention strategies; and assuring widespread adoption of the approach (Centers for Disease Control and Prevention, 2023).
When applied to those who have offended, this is a negative view which measures them according to their likelihood of engaging in harmful behaviour, whether to themselves or others. It is concerned with ‘danger or peril’ . . . ‘some unhappy event which may happen to someone’ (Ewald, 1991: 199). That danger or peril is normally measured from what has happened in the past, with its potential occurrence in the future; it is the art of prediction. Indeed, as Beck (2000) says, it is risk measurement: those ‘practices and methods by which the future consequences of individual and institutional decisions are controlled in the present’ (p. xii).
The negativity can be seen in relation to the gaze the risk measurer directs towards the person being measured. It is essentially
The frowning, negative, risk-oriented criminal justice gaze is not the only way we could look at those who have offended. Another possibility would be to acknowledge the uncertainty involved in prediction and say we simply do not have reliable quantification for risks of, say, reoffending over a long time frame. There is, for example, a rather ignoble history of criminology trying to predict dangerousness in the last century, where it became clear that though committing criminal offences in the past was a significant guide to committing more such offences in the future, the interventions used and even the language in reports on individuals for, say, parole boards, themselves tended to lead to judgements of dangerousness, without necessarily any further evidence. Maguire et al. (1984) found that any mention of dangerousness in reports created a ‘label’ of not necessarily being safe, which affected parole board decisions to release (or not) those on life sentences.
Another approach would be to view individuals as having the potential to change, and to desist from offending. This is similar to the ‘good lives’ approach to offending (GLM: see, e.g., Ward, 2010; Ward and Maruna, 2007) and also those perspectives from desistance theory which emphasise the role that agency plays as well as individuals’ attempts to lead a ‘good life’ (see Paternoster et al., 2015). Andrews et al. (2011) have compared the good lives model with a straight RNR approach and conclude that GLM does not add much to an in-depth use of RNR that builds on strengths and rewards noncriminal alternatives to risk factors favouring criminal activity. Partly this is an issue of implementation – that those using RNR have not always emphasised the need for this full and rounded appreciation of the individual, so not always considered strength in depth. In addition, in desistance research, it has been found that those aspects promoting desistance, both relating to the individual (such as agency) and to the social context (such as partners and significant others emphasising keeping away from offending) are not simply the opposite of those leading to more offending (Stouthamer-Loeber et al., 2004). Hence, it is important to consider the strengths of the individual offender positively and fully, rather than just note the absence of criminogenic risk. As we will go on to discuss, restorative justice offers an approach for achieving this, with commonalities with the GLM, as strength-based approaches intended to restore dignity and enable agency to rebuild trust and facilitate desistance from crime (Walgrave et al., 2019).
Indeed, it is now clear that many of those who commit offences, as they age and potentially mature, decide to try to change their lives and desist from offending (Laub and Sampson, 2003). They themselves can see the looming spectre of what Paternoster and Bushway (2009) have called their ‘feared self’: someone continuing to be drawn into crime, losing the people they value (partners, children, etc.) and spending more time in prison. Even those with more persistent offending behaviour normally desist, as the age-crime curve shows (Bottoms and Shapland, 2011). Turning away from crime makes that feared self-less likely to occur and could provide (as Paternoster and Bushway (2009) argue) an initial impetus to desistance, or motivation to continue to try to desist. More punitive criminal justice systems, which focus primarily upon risk, as does the system in England & Wales, can stymie those positive impulses to change to non-offending, because if people with a significant history of offending, despite long crime-free gaps, commit another offence, their long records prompt courts to send them to prison for longer and longer periods, thereby negating any social and human capital which may support desistance (Shapland, 2022a).
The concern is that, in this concentration upon criminogenic risk, we may have created a new version of Martinson’s (1974) view of interventions with people who commit offences. At that time, looking at the dispiriting results of evaluations, Martinson concluded that ‘nothing works’, as his view came to be called. In fact, his own view was that interventions would work for some people some of the time, and we needed to be far more specific. Today, the pessimism of the criminogenic risk doctrine seems to be leading to a view that nothing will change in terms of preventing reoffending
There is, therefore, in our view a real need to reconsider the consequences of an unvarnished risk doctrine. Should we instead be thinking of considering risk with uncertainty and the individual’s own potential promise, that those who offend can change? This would create a new version of risk which would not only take on board the ‘good lives’ research but also the insights from virtue ethics, that those who commit offences have agency, are often motivated to act in line with ethical virtues, and are generally trying to live what they would see as a ‘good life’ (Bottoms and Shapland, 2014).
Risk assessment in restorative justice
Our eyes have been opened to the need for a new conceptualisation of risk in criminal justice through exploring the nature of risk and mitigation of risk in restorative justice processes. Commonly, restorative justice practices have been compared against criminal justice (see, for example, Walgrave, 2007). In this instance, however, we think that criminal justice might well have much to learn in its conception of risk from restorative justice.
We first need to define the aims and values of restorative justice. There are many different processes termed restorative justice, including mediation (a direct meeting between the victim and offender, with a trained facilitator), indirect or shuttle mediation, conferencing (with the victim, offender and facilitator being joined by supporters) and sentencing circles or panels (often with community members also present). They have developed, often in parallel, in different countries, informed by particular theoretical perspectives (Shapland et al., 2011). Hence, the aims of restorative justice programmes vary considerably. Some focus on particular outcomes, such as reducing harm, healing and promoting reconciliation (Zehr, 1990). Others see restorative justice as a means of resolving conflicts in communities (e.g., Christie, 1977). Yet, others include among the outcomes reducing reoffending, as does Braithwaite (1989). Some statements of aims focus not on ultimate outcomes but on the process of restorative justice, such as that of Marshall (1999)’: Restorative justice is a process whereby parties with a stake in a specific offence collectively resolve how to deal with the aftermath of the offence and its implications for the future’ (p. 5). Other definitions of restorative justice embody both outcomes and specification of process, such as that of Dignan and Marsh (2001). This article will be describing the views of restorative justice facilitators from several different countries across Europe, from schemes with different aims.
The values of restorative justice schemes, however, are far more similar to each other than are their expressed aims, and are also linked to the relevant international instruments of the Council of Europe and European Union,
3
as well as to the academic literature (Braithwaite and Mugford, 1994). They include that restorative justice processes should be voluntary for all participants, who should be able to make an informed choice as to whether to participate or continue to participate; that they should be safe and participants should feel safe; that the person who has caused harm should admit responsibility for causing that harm; that communication should be facilitated by a neutral, trained facilitator; that communication should be respectful; and that the preferred restorative justice process should be appropriate to the offence concerned and to the circumstances of the participants. This last point is particularly relevant here because it implies an acknowledgement that every person – and every incident of crime, harm and/or conflict – is different. So, for example, restorative justice, as defined by the Scottish Government (2017), is a process of independent, facilitated contact, which supports constructive dialogue between a victim and a person who has harmed (whether this be an adult, a child, a young person or a representative of a corporate or other body) arising from an offence or alleged offence.
These common values of restorative justice, as we shall see, drive the forms of risk assessment being used and allow comparison between schemes which may have rather differently expressed aims.
Before starting a process of restorative justice, the international instruments require that there be a risk assessment carried out by a trained facilitator. In the course of the development of restorative justice in Scotland, where it was envisaged that restorative justice provision was to be expanded rapidly (Scottish Government, 2019), we realised that there was little research on this risk assessment or its parameters and processes, or the mitigatory elements introduced by the facilitator if risks were found to exist. 4 We therefore undertook some exploratory research on risk assessment and mitigatory processes for restorative justice, funded by the Scottish Government.
The research on mitigation and risk
The research involved interviewing experienced facilitators from 11 European jurisdictions as to their typical practice and procedures for assessing potential risks involved in undertaking restorative justice for different offences, together with what they had found to be potentially helpful mitigations for risks present for individual cases. A total of 30 facilitators were interviewed, from Scotland, England and Wales, Northern Ireland, Belgium, Denmark, Norway, Finland, the Republic of Ireland, France, Austria and Estonia. 5 We interviewed facilitators from a number of countries, because the dominant forms of restorative justice varied (e.g., whether mediation or conferencing is predominately used), as did its legal basis (whether it was set out in legislation or not), whether or not restorative justice was a mainstream element for young people or adults who offend), and the types of organisation delivering restorative justice (government and/or third sector organisations), as well as the culture of the criminal justice system. Hence it was possible that the process of risk assessment would also vary.
This sample of experienced facilitator respondents cannot be seen as representative of all restorative justice providers or facilitators in these countries. It is not possible to acquire a representative sample because there is no defined population of either providers or facilitators in most countries. Restorative justice is primarily provided by agencies working in the voluntary sector, with referrals from criminal justice personnel, though in some countries facilitators may be employed by criminal justice agencies or social work bodies (as in Northern Ireland and Scandinavia). This must therefore be regarded as exploratory research. We did, however, take care to approach potential respondents from the main providers in each country. Respondents were approached personally, generally from previous research contacts of the authors. Ethical approval for the research was given by The University of Sheffield School of Law Research Ethics Committee.
The aim was to obtain a sample of practitioners who had considerable experience of facilitating restorative justice over many years and who therefore would be able to draw on their experience of working on a wide variety of cases potentially posing a variety of risks, and who had employed and would be able to comment on different possible mitigatory strategies for those risks. Details of the sample are given in Shapland et al. (2022). Thirteen facilitators had worked with both adults and children who had harmed, nine with primarily adults who had harmed, and eight with primarily children who had harmed. Twenty-three facilitators had responsibility for supervising others. We wanted to know whether they had dealt with more serious or complicated cases, so specifically asked whether they had dealt with cases of homicide (in 15 interviews they had), cases of sexual assault (in 20 interviews they had) and cases of domestic violence (in 16 interviews they had). The sample is therefore one of experienced facilitators, and the results cannot be taken to represent the practice or views of more junior practitioners.
Despite our expectations, the results showed very similar processes undertaken by facilitators in different countries and from different providers. Risk assessment for restorative justice was about what risks would be present for that particular set of participants and that case, and whether the risks identified enabled the facilitator to move forward with the restorative justice process the participants desired. It was not a one-off assessment but a continuing process, depending on the information received and the views of the potential participants. Some facilitators did use checklists which prompted them to cover certain areas, as indicated below, but there was no actuarial instrument used. This was not due to any aversion to evidence-based practice, but simply because actuarial instruments assume it is possible to apply group-based measures to an individual’s risk assessment. On the contrary, the facilitators were adamant that risk assessment for restorative justice had to be assessed on an individual basis, with many different types of risk covered, and with the active participation of the potential participants. Risk had to be evaluated in relation to all the potential participants in whatever type of restorative justice process was being considered and would be agreed by the participants (including where applicable, any supporters, advocates or legal representatives), not simply the person responsible for the harm or the victim/person harmed. It could not be a top-down, entirely professional-led process.
Key attributes of assessing risk for participation in restorative justice
The aim of assessing risk in relation to restorative justice is to minimise the risk to all potential participants of proceeding further to set up the restorative justice process, considering possible different mitigations which could be put in place. It is therefore not a future outcome-predicting mechanism, as to, for example, whether victim needs would be met, or reoffending be reduced, but gauging the risk of using different procedures in terms of the safety of the participants. The risk assessment done has therefore to be a true risk assessment of the particular situation, rather than an attempted prediction of future outcomes. This makes it different from much risk assessment in criminal justice, in terms of the focus on the immediate procedural next step, rather than any eventual outcomes, and in the concentration on what individuals (as opposed to agencies) wished to achieve. The key attributes of doing this risk assessment, for all the expert facilitators interviewed were the following. 6
First, the facilitators saw risks as relating to individual cases and individual potential participants, not types of people or types of case (or types of offence). Some types of case or offence might be overall more risky than others (e.g., domestic abuse and cases between family members are generally riskier than say burglary where the parties do not know each other), but other elements may reduce or enhance those risks.
Second, the interviewees said it was impossible to predict the risk in a particular case before engaging with the individuals involved. The reasons are, first, that different people experience ‘risk’ factors and deal with them in very different ways. What may be an indication of risk to, say, a professional working with the victim may not be seen in the same way by the victim herself or himself. Alternatively, something perceived as minor by one person may be a real trigger for another, depending on their previous experience and other personal circumstances. Equally, the seriousness of the harm experienced in a particular case may not be seen as equivalent to criminal justice system or legal ratings of the harm of that category of offence, as shown in seriousness scales such as the Cambridge Crime Harm Index, which derives its measures of harm from the lengths of sentences in terms of days of imprisonment given by judges (Sherman et al., 2016). By definition, that index reflects the views of judges, not those of the particular individuals, such as victims, involved in the case. It is clear that offences can have very different effects on different victims (Shapland and Hall, 2007).
Although the risk of proceeding could not be calculated from the views of those who were not potential participants in the restorative justice process, it was none the less important to obtain information from professionals who were already working with the potential participants, both in terms of other surrounding circumstances and also to see whether they would be able to offer relevant support to the participant. So it was important to contact relevant social workers, medical professionals (particularly in the case of mental health difficulties) and other key workers, as well as, of course, the legal guardian/parents of young people, where applicable.
Although the views of other professionals were seen as important in making decisions about restorative justice, facilitators’ own judgements and, particularly, the views and needs of the potential participants were seen as more important. All decisions on risk and mitigation should, they thought, be discussed with the potential participants individually and their views sought (though the final decision has to be that of the facilitator – and where the restorative justice is set out in legislation, it is the facilitator who is given that responsibility).
Importantly, the risk assessment decisions always involved comparing the potential risks in going ahead with the restorative justice process, with the risks of
Risk assessment for restorative justice involves considering risks, but also considering, simultaneously, potential mitigations for those risks. Risk and mitigation consideration are done together, not separately (as would tend to happen in criminal justice in relation to criminogenic risks and strengths, or risky factors and protective factors). This was seen by the facilitators, many of whom had previously worked in criminal justice settings (Shapland et al., 2022), as a much more holistic and whole person process, that included consideration of the roles and needs of other people who might be involved in the process, rather than the ‘totting up’ of individual factors, such as the process used in some assessments in criminal justice (such as OASyS used by the prison and probation services in England & Wales: Moore (2015)).
Many participants, in the interviewees’ experience, wanted supporters present, particularly for face-to-face meetings. These could be family, friends or colleagues, but might be other professionals involved with the participant. This was seen by the facilitators as useful, but care was needed to avoid power imbalances or inappropriate interventions from supporters. To this end, the facilitators insisted on meeting all potential participants before the meeting – ‘surprise appearances’ were seen as extremely problematic.
Risk assessment for restorative justice is a process, not a one-off calculation. Hence, as the restorative justice procedure develops, risk assessment needs to keep pace with this, with assessment being revisited and mitigation reassessed as the case develops. The corollary of this is that facilitators must not be too quick to ‘jump’ to considering the final format for the restorative justice (e.g., a letter of apology or a face-to-face meeting), but rather be thinking of the next, rather than the final, step. The dynamic nature of risk assessment means that facilitators are prepared to pause or call a ‘timeout’ during a restorative justice encounter if problems arise to check whether and how the process may be safely continued.
Mitigations for risks
Just as the risk assessment needs to be aligned with the particular context of the case and the participants’ views and wishes, so mitigations for risks encountered also need to be specific to that situation and those people. The range of risks which may be identified is very large – the potential for conflict; participants’ mental health; threats to easy communication; physical aspects of the venue and so forth. If we think of the venue for a face-to-face meeting, for example, it needs to be accessible to participants, with both entry and exits being considered (and/or entry and exit times staggered) so that participants do not inadvertently run into each other. Ideally, there should be a break-out room, so that if anyone becomes upset or annoyed, there can be a time out. All of these aspects would be key to any community meeting, but for a restorative justice meeting, it is also important that the meeting place is not redolent with unfortunate previous memories for any particular participant. Police stations or courts are likely to have such links for those who have caused harm, for example. Sometimes, the venue is set because one party is serving a custodial sentence. In that case, victims can find going into a prison to be scary and distressing, because of the necessary rules and the security provisions (Shapland et al., 2011). It will also be necessary to organise for each participant to be escorted through the prison.
If we think about potential risks to communication, these can include where participants speak different languages (when interpreters can be needed), or if others may be disconcerted if one participant has a particular ‘tic’ (such as nervous laughter), or if there are mental health problems. There are potential mitigations for all these. Barriers to effective communication can be discussed with other participants beforehand by the facilitator so that effects are minimised and people are prepared. Many mental health conditions vary over time, so that the time course for the restorative justice process can be suited to when it would be best for that participant. We have discussed the many forms mitigation can take in more detail in Shapland et al. (2022).
When risk becomes too great
After considering potential mitigations, are there then cases which are just too risky to permit going ahead with restorative justice, even if potential participants wish to do so? The experienced facilitators we interviewed all considered that there were such cases, but that ‘too risky’ was relatively rare, potentially dependent on timing or other interventions (such as therapy), and should not be typified as a particular group of cases. In other words, the decision not to proceed needed to be taken on a case by case basis and could not be reduced to a particular kind of offence or type of person. This is consistent with the individualisation of risk assessment and of potential mitigations (and indeed with Beck’s (1992) original idea of individualization, quoted above). The decision not to proceed could not simply be because the offence was one of domestic abuse, or sexual assault, or involved a participant who found it hard to communicate. ‘Too risky’ cases were not seen by these experienced facilitators as
This makes it much harder for facilitators to judge the extent of risk than having a simple rule that, for example, a score of a particular amount on a questionnaire would bar someone from restorative justice. Working with intrinsically more risky cases puts more emphasis on preparation, consultation with potential participants, thought about venues and decisions on the type of restorative process being used. Difficult or more complex cases require additional care and preparation. Bringing in advice from experienced colleagues or discussing difficult cases at an ‘action learning set’ meeting can be helpful. ‘Action learning sets’ is the term used in the health service, for example, for a meeting specifically convened to solve problems but has been used in the Probation Service for a meeting of colleagues to discuss (current or past) cases which are or have provoked particular difficulties, as a learning and problem-solving tool, to disseminate experience through a team where members normally work individually on cases. 7 Several of our interviewees used this approach without necessarily calling them ‘action learning sets’.
The major kinds of risks that would make facilitators pause and seriously consider whether to proceed with the restorative justice process were:
If we consider this list of seriously risky obstacles to effective restorative justice proceedings, we can see that each of these risks threatens to undermine one of the core values of restorative justice. Each is of course dimensional – it can be present to a minor or more substantial extent. In relation to the last point, for example, everyone thinking about participating in restorative justice is probably seeking to influence others present, so that they can appreciate the pressures at the time of the offence, or the effects the offence has had. But it is when that wish to influence becomes a likely attempt to control or coerce that the restorative justice process itself is threatened. There is therefore a link between the kinds of risks that make restorative justice too risky and its core values. And yet, serious concerns at one point do not rule out the possibility of restorative justice permanently. Given that restorative justice has underpinning assumptions about the belief in people to change, to take responsibility for wrongs, and engage in constructive dialogue, even where a case is seen as ‘too risky’ at one point in time, there may be a future when the risks are seen as manageable where, for example, someone recovers from mental health or drug misuse issues, takes responsibility for their harmful behaviour, or is willing to engage in dialogue.
There may be a link here to how we might consider risk in criminal justice. ‘Too risky’ in restorative justice relates to breaches of core values of restorative justice, such as voluntariness, taking responsibility for harm caused, and allowing all participants to take part. This would mean that were we to apply the same principles to risk assessment in criminal justice, would the only elements of risk for criminal justice which would be dangerous be those undermining the core values of criminal justice? And, if so, what would those core values be for criminal justice? This is not a discussion which has often featured in relation to ‘improving’ the process of criminal justice. One aspect might be moderating the degree of coercion able to be applied to a participant. Unlike restorative justice, participating in criminal justice is not voluntary, whether for those accused or convicted of criminal behaviour (viz. powers of arrest and imprisonment if they do not attend court) or witnesses, including victims (witness summonses). Yet, the extent to which coercion can be applied is limited in principle by applicable human rights instruments (such as the United Nations (1984) Convention against Torture), as well as in practice (unwilling witnesses are not termed ‘hostile’ witnesses by the prosecution as a joke). Unlike in restorative justice, involvement in mainstream criminal justice processes can be compelled, but need not be sincere. Although restorative justice practitioners may intend to protect participants’ human rights in a general sense, our interviewees were more concerned with the specific values of restorative justice; principles such as due process, proportionality and rights to legal representation were developed specifically to deal with criminal trial processes, and therefore, their applicability to restorative justice is a matter of debate (Skelton and Sekhonyane, 2007).
In considering the applicability of our argument about risks as threatening principles in relation to criminal justice, it may be necessary to consider what are the core values for criminal justice and whether it is possible to move beyond the current group-based, actuarial measures described in the introduction to this article to a different form of risk assessment.
Risk, restorative justice and desistance
It is worth considering whether the attributes for conducting good risk assessment in restorative justice fit with common outcome measures and aims for restorative justice schemes. If the suggested ways of doing risk assessment cut across and against achieving good outcomes, that would not be very helpful. Restorative justice providers may have different aims for their schemes, depending on their theoretical orientations (Robinson and Shapland, 2008; Shapland et al., 2011) but also on the justice culture and legislative/organisational context in that country (see discussion in the Introduction). Some emphasise healing and resolving conflict, particularly where participants know each other; others reducing reoffending. All would see meeting victims’ needs as important, as well as various outcomes in relation to those responsible for harm. However, some would aim at answering victims’ questions, others at reparation and yet others principally at closure and emotional healing.
In relation to victims, it is widely accepted that offences may affect victims differently, depending on the context, victims’ previous experiences of victimisation and other aspects of their lives (Shapland and Hall, 2007). Hence victims’ needs will differ. We have already seen that the kind of risk assessment advocated by the experienced facilitators we interviewed majors on consultation with all participants and finding out their own individual perspectives, expectations and needs in relation to what restorative justice can bring. Preparation with victims and others harmed should explore their own expectations and seek to meet those (rather than any stereotypical idea of what, say, burglary victims normally want). The individualised risk assessment and mitigations in risk assessment for restorative justice should therefore allow those needs to be ascertained and the most suitable form of restorative justice procedure adopted to try to meet those needs. The likelihood of participants having altruistic components to their expectations for restorative justice should also aid in each party being willing to try to meet others’ needs as well (Shapland et al., 2011; Van Camp, 2016).
A key aim for many restorative justice providers (and their funders) is to reduce reoffending, or at least the extent of reconviction/rearrest. There is now significantly more evidence available about whether this is achieved, through meta-analyses and reviews, both for adults and young people (College of Policing, 2016a, 2016b; Sherman et al., 2015), though much of this evidence relates to face-to-face conferencing. The evidence suggests that there is a significant reduction in at least the frequency of reoffending after restorative justice. Looking at those who are most likely to reduce their criminality, however, suggests that the mechanisms involved are relational and agentic – both Shapland et al. (2011) and, earlier, Maxwell and Morris (2001) in Australia and Hayes and Daly (2003) in New Zealand found that it was experiences during face-to-face conferencing which related to the likelihood of lower recidivism. Relevant aspects were the conference making the person responsible for the harm realise the harm done; the extent to which he or she felt actively involved in the conference; how useful the process was found to be; consensus in coming to an outcome agreement; the person responsible feeling remorseful, he or she feeling involved, and shaming not being stigmatising.
Reducing reoffending in relation to those who offend on several occasions essentially involves promoting desistance – the process of giving up committing offences. Promoting desistance can involve affecting people’s agentic views as to how they wish to lead their lives in the future (early desistance) and/or maintaining the desire to desist and providing support in surmounting the obstacles to desistance that their social context throws at them. It has been argued that restorative justice, particularly face-to-face conferencing, provides an individualised process which both maintains motivation (through the victim and family members supporting the person responsible’s suggestions on turning their lives around and, as relevant, apologising for the harm done) and also provides solutions to problems relating to training and reskilling, accommodation and employment (through participants’ suggestions and social capital in constructing the outcome agreement) (Shapland, 2022b).
Again, the emphasis is on individualised measures and on working actively with people responsible for harm (and victims). A risk assessment for restorative justice which starts off the process by putting the same concentration on individual cases and on what all participants wish to achieve is likely to be one which can lead through to outcome agreements which support the journey towards desistance. The philosophy of risk assessment adopted by those facilitators we interviewed is one which is congruent with the theoretical underpinnings of how to meet victims’ needs and to try to promote desistance.
A restorative, people-centred view of risk for criminal justice
We have seen that the type of risk assessment being used by the experienced facilitators interviewed in the research can be characterised as being individualised, consultative, and exploring what the potential participant sees as the aims of the process for them. Decision-making is shared between facilitator and participant, and attempts are made to be transparent about aspects of the restorative justice process being constructed, though ultimate responsibility for the restorative justice going ahead lies with the facilitator, with the individual participant only being able to determine their own participation (because restorative justice is always voluntary).
As we will discuss later, this is different from the group-based and more actuarial risk assessment often used in criminal justice. We are not arguing that all criminal justice risk assessment is entirely actuarial or group-based. Some probation work, particularly, for example, on sentence planning, is ideally individualised to the particular needs of the offender and his or her social circumstances. It should be undertaken with the offender and focus not just on criminogenic needs on an RNR paradigm, but also on strengths and responsivity (the learning style and abilities of the offender). However, such probation work is accompanied and, it can be argued, structured by a risk assessment based on likelihood of reoffending, an ultimate outcome measure calculated on a group basis. Nellis (2022) had also warned that this individualisation may be compromised by new digital means of calculating risk and creating profiles.
There is a further model of risk assessment used in criminal justice, which is individualised, but again, we would argue, different from restorative justice risk assessment in some key ways.
Clinical risk assessment, as used by, for example, health professionals and social workers, prides itself on being individualised to the individual being considered (Cree and Wallace, 2009). In its traditional paradigm, however, it involves the professional using their training and professional judgement to assess each risk, with the final decision being taken by the professional and sometimes not even being conveyed to the individual. An example would be a medical or psychiatric report to the court prior to sentencing. The report may be designed to affect decisions on process or on outcome, with the result being conveyed to the individual usually by the main decision-maker for that stage of the criminal justice process (e.g., judge or prosecutor), not the professional. Clearly, there is a difference in power, and sometimes in information possessed, between this clinical risk assessment and that used in restorative justice. Clinical risk assessment falls under a paradigm of rehabilitation, rather than restoration, and does not involve co-production with the individual being assessed. This means it is limiting the individual’s agency in being able to shape the process (and decisions).
Another key difference between clinical risk assessment and assessment for restorative justice is whether it is assumed that the individual can change (and may, in the future, wish to change). Actuarial risk assessment, based on the past, does not envisage individuals changing in the future unless, previously, similar individuals have so changed. So, for example, as discussed earlier, OGRS as a measure of risk takes on board whether similar people have reoffended within certain time periods. But it cannot say whether a particular individual may decide to ‘buck the trend’ and do something different. Clinical risk assessment essentially also relates to the past, but is based on the professional’s experience and training (and the relevant research) as to whether the individual is giving responses and has previously behaved in ways which suggest they may change. Restorative justice, however, assumes, from its core philosophy, that
Restorative justice offers us a personalised approach that empowers the individuals and communities affected by specific criminal harms to participate in co-producing the response to those harms (Weaver, 2011). Through the values of dignity, respect and inclusion, and the practice of constructive dialogue aimed at mutual understanding, consensus-building and peace-making, it provides a way of seeing risks and their mitigation as arising jointly from people’s needs, wishes, concerns and ingenuity. This is not to say that professional judgement or risk assessment tools are irrelevant; indeed, whether in criminal justice or restorative justice, there will be instances where the professionals deem that a person’s desired course of action is not safe, feasible or legal. However, focussing on next steps, and how identified risks may be mitigated, provides a way that criminal justice practitioners may collaborate with people who have committed criminal harm, as well as the people harmed and relevant communities, to develop shared understandings of risks and ways of managing them. Such an individualised and collaborative approach offers real potential for meeting people’s needs and facilitating commitment to agreed actions.
Conclusion
The conference which was the inspiration for this volume was titled, ‘Shifting logics in criminal justice’. When we started the research described in this article, we did not expect to discover a new paradigm for risk assessment. 8 This new paradigm has significant differences from those actuarial, group-based approaches and clinical risk assessments currently used to aid decision-making on processes and outcomes in criminal justice. It became clear, however, as we looked at the responses of experienced facilitators as to how they did the risk assessment required to proceed with restorative justice in a particular case, that both the philosophy and the practice were not the same as these familiar criminal justice modes. Restorative justice can bring, we think, a new logic to criminal justice. This logic seems to be one which is used across different modes of restorative justice (conferencing, direct mediation, indirect mediation) and across different countries. Fundamentally, it sees people as potential collaborators with underlying positive motivations who can be engaged with constructively to develop shared understandings of risks and ways for mitigating these, instilling dignity and respect to move towards a better future.
Commonly, the transfer of knowledge is the other way round. Restorative justice is compared against the tenets of the criminal justice system and criminal law, often to its perceived detriment (e.g., that it is in tension with philosophies such as proportionality of sentencing in a retributive sentencing paradigm, as argued by Ashworth (1993)). In this instance, however, it is possible that restorative justice may provide some pointers to a more adequate framework for risk in criminal justice. Could criminal justice add the potential for change and for agency on the part of participants without negating its core goals and tasks? Could we see sentencing or the operation of probation embody co-creation and respecting the agency of those involved? This would not remove the need for risk assessment, or the primacy of the need for decision makers to reflect societal dimensions, not just those of individual participants. It would though change the type of risk assessment and its outlook, to a more individualised, person-centred, empowering view of risk, which includes the strengths and goals of participants, which is attentive to the particular experiences of individuals’ lives, and which considers the risks of not implementing positive changes. Such a view, which we think is compatible with the core goals of criminal justice, could also be trauma-informed, consonant with its individualised approach, and aiming to support the person responsible for the harm and other participants to positive, wished-for changes, not just an absence of harmful behaviours.
Footnotes
Declaration of Conflicting Interests
Funding
ORCID iDs
