Victim Empathy Intervention With Sexual Offenders

Introduction

Do sex offenders need victim empathy? Despite the regular emergence of review papers on the topic of empathy (e.g., Day, Casey, & Gerace, 2010; Hanson, 2003; Hanson & Scott, 1995; Marshall, Hudson, Jones, & Fernandez, 1995; Polaschek, 2003), there is still no consensus on whether victim empathy is an essential, important, unnecessary, or potentially dangerous component of sexual offender treatment. Since the publication of Hanson and Morton-Bourgon’s (2004, 2005) meta-analysis, which concluded that “the clinical presentation variables (e.g., . . . low victim empathy . . .) had little or no relationship with sexual or nonsexual recidivism” (p. 17), there has been discussion among treatment providers about whether victim empathy work could (or should) be removed from treatment programs. The need for such action is contradicted by studies into offenders’ experiences of treatment programs, in which they invariably report that victim empathy work was one of the most powerful facets of treatment (e.g., Levenson, Macgowan, Morin, & Cotter, 2009; Levenson & Prescott, 2009; Wakeling, Webster, & Mann, 2005). This discrepancy has not been reconciled. For instance, Day et al., although promising to address the question of treatment impact, drew conclusions only for researchers and provided no guidance to practitioners.

To laypeople, the answer to the question, “Do sex offenders need victim empathy?,” would be a straightforward “Yes.” The general public would probably feel the question does not need to be asked. Surely if someone had empathy with another person, they would not be able to assault them harmfully? Given the widespread practice of victim empathy interventions, it appears that treatment providers hold the same beliefs. The most recent survey of sexual offender treatment providers (McGrath, Cumming, Burchard, Zeoli, & Ellerby, 2010) indicated that between 87% and 95% of North American programs involve victim empathy as a treatment target. For adult community programs, victim empathy was the most commonly cited treatment target, and for adult residential (including prison) programs, victim empathy was the second most commonly cited treatment target. This position is not limited to North America. In England and Wales, for example, the current prison sex offender program devotes about one fifth of its time to victim empathy work.

However, as Gendreau and colleagues argued, apparently ‘common-sense’ or intuitive approaches to treatment can constitute “correctional quackery,” where correctional staff prioritize personal experience, values, and anecdotal evidence above evidence resulting from rigorous, large-scale research efforts (Gendreau, Smith, & Theriault, 2009). ¹ Correctional quackery results in services being delivered to offenders that do not produce the outcomes they intend, usually meaning that the service is intended to reduce reoffending but does not do so. In contrast, rehabilitation has been defined as “administering interventions intended to heal or manage a person’s (physical or psychological) disorder or affliction” (Ward, 2010, p. 288). Usually in the forensic context, rehabilitation consists of practices that aim to reduce or eliminate reoffending and is judged on its ability to do so. Aiming to ascertain whether victim empathy work in its current form is legitimate rehabilitation or “correctional quackery,” this article will examine the nature of current practice and will evaluate the evidence base for victim empathy development as a treatment target for sex offender rehabilitation. The kinds of evidence that would justify the conclusion that victim empathy intervention is rehabilitative would include, for instance, an empirically-based theory for the existence and causal role of victim empathy deficit in offending or evidence that impaired victim empathy is associated with higher reconviction rates as well as data indicating that treatment procedures designed to enhance empathy for victims reliably lead to a reduction in reoffending.

Victim empathy work could be consluded to be correctional quackery if it does not reduce reoffending or if it actively works against this rehabilitative aim, yet practioners or policy-makers continue to believe in its efficacy despite the evidence otherwise. A third alternative is that victim empathy intervention work may be seen as desirable because it achieves another aim: that of punishment. Ward (2010) has defined punishment as “the intentional imposition of a burden on an individual” (p. 287), through actions that are authorized by the state, intentional, reprobative, retributive, and harmful. Technically, according to Ward’s definition, an activity can only be punitive if this is its intention; however, in this article we adopt the position that an activity can be inadvertently punitive by causing overall harm even if this is not its intention.

What Is Victim Empathy?

One of the difficulties in reviewing this topic is the fuzziness of the concept of victim empathy. In the literature, the term “empathy” is often used interchangeably with “awareness” and, in some cases (as will be discussed later), “remorse.” Hilton (1993) suggested that empathy involves both cognitive and emotional components, noting that “The cognitive aspect requires being able to state what another person might feel, while the emotional element involves automatically feeling what the person feels” (p. 290). Jolliffe and Farrington (2004) emphasized the same distinction between cognitive and emotional processes. Consequently, Barnett and Mann (2012) defined “empathy” as a cognitive and emotional understanding of another’s experience, resulting in an emotional response that is congruent with a view that others are worthy of compassion and respect and have intrinsic worth. “Victim empathy,” therefore, could be defined as a cognitive and emotional understanding by a sexual offender of the experience of the victim of his or her sexual offense, resulting in a compassionate and respectful emotional response to that person.

What Does Victim Empathy Intervention Involve?

The notion of “developing victim empathy” as an important treatment goal for sexual offender treatment is seen in an early influential text (Salter, 1988), which urged therapists as follows:

Sex offenders must show progress in developing empathy for their victims. This imperative is explicitly addressed in treatment by bringing adult victim advocates and sometimes adult survivors to groups, by assigning readings, by requiring the offender to write the assaults from the child’s point of view, and by having the offender write an appropriate apology. . . . The offender must demonstrate through words and behaviour that he is making progress in learning that victims have a point of view separate from his own, that he is growing in his ability to determine what the impact of his behaviour is on others, and that he is moving toward developing empathy for his and others’ victims. (p. 177)

In another influential text outlining a comprehensive approach to sex offender treatment, Marshall, Anderson, and Fernandez (1999) reviewed the content of a number of established treatment programs and reported the following methods in use: (a) the offender role-playing the victim; (b) the offender writing an account of the offense from the victim’s perspective and reading it aloud to the group; (c) group discussion of the reactions of members of the group who have been abused themselves; (d) reconstructions of the offense where the offender takes on the role of the victim, apparently as a form of covert sensitization; (e) enabling group members to read the written accounts of victims of sexual abuse and then asking the offenders to write a report on the effects they have read about; (f) showing video tapes to the group that describe the harm caused to the victims of sexual assault; and (g) arranging meetings between the offender group and adult survivors of sexual assault.

In Marshall’s own program (as described in Marshall et al., 1999) the following treatment methods were initially employed: (a) increasing offenders’ abilities to recognize different emotional states in themselves and others; (b) a group exercise that identified the harmful effects on sexual abuse victims; (c) personalization of this group exercise to each offender’s own victim(s); (d) each offender describing the offense from the victim’s perspective; (e) the offender role-playing his offense from the victim’s perspective; (f) requiring each offender to write two hypothetical letters, one from the victim to him and the other to the victim; and (g) a group exercise where the group read actual victims’ accounts and discuss the emotions therein as well as their own emotional responses to the accounts. Marshall et al. reported how they later abbreviated this component to just two activities: generating a list of effects and then writing the two hypothetical letters. Marshall et al. concluded their account by noting that it is unclear whether activities such as those described above target emotional empathy as well as cognitive empathy and expressed concern that there is no obvious way to measure the “emotional depth” of an offender’s apparent empathy for his victims.

Reviewing the standard approaches to empathy enhancement as described by Salter (1998) and Marshall et al., 1999 it appears that two main methods are utilized. First, there are a range of treatment activities that could be called psychoeducational. These involve providing offenders with information through various media about the impact of sexual abuse on victims. The meaning of the information is subsequently expanded through group discussion and other tasks intended to enable the offenders to reflect on the experiences of sexual abuse victims in general. As Hilton (1993) noted, these activities are presumably intended to develop cognitive empathy. Second, there are methods which require offenders to reexamine their own offense in order to experience an emotional understanding of the victim(s)’ experience during and after during sexual assault. These methods are more experiential, often involving psychodramatic activities (e.g., Mann, Daniels, & Marshall, 2002; Webster, Bowers, Mann, & Marshall, 2005).

Victim Empathy “Models of Change”

Ideally all psychological intervention should be underpinned by an explicit, evidence-supported model of change. A model of change accounts for how an intervention is supposed to work: what it should change, how it will change it, and why this is important. Reviewing the victim empathy literature, three possible models of change emerge, though these are often not articulated formally or in detail.

First, there is a common-sense model, such as that advocated by Salter above. Salter did not formulate or utilize a theoretical model of empathy to underpin her focus on victim empathy. The source of her advice appears to be intuition and “common sense.” Second, some authors, such as Pithers (1999), subscribe to a model of change that predicts that victim empathy training will increase motivation to desist from offending. As Pithers put it, this model rests on the assumption that “the client must perceive a profoundly significant purpose for immersing himself in a challenging treatment process and working to preserve and reinforce a new approach to life. One potential source of enduring motivation for sexual abusers may be greater empathy for their past and potential victims” (p. 258). Third, many authors subscribe to the notion that empathy inhibits aggression and so interventions that increase empathy will decrease aggressive behavior. This model of change is often stated rather simply (e.g., Curwen, 2003), although a full account of the evidence base for this model can be found in Marshall et al. (1999).

A fourth model of change might be as follows. Role-playing past traumatic experiences to uncover contemporaneous irrational thoughts is a commonly used cognitive therapy technique (e.g., Beck, 1995). In victim empathy role-plays (see Mann et al., 2002) the purpose of the role-play is to uncover self-serving biases, rather than self-damaging biases, but the principle is the same: Revisiting memories of past events in the context of cognitive therapy enables the client to reassess the extent to which their beliefs at the time were mistaken, which in turn assists the client to choose different courses of behavior should similar events occur. It could be argued, therefore, that revisiting a past sexual offense with a focus on the victim’s perspective enables an offender to see the situation differently and that the more accurate perspective taking achieved through hindsight develops a skill that will generalize to other, future situations.

In the absence of any formally stated evidence-informed models of change, there are three key issues to establish in order to justify victim empathy as an intervention for sexual offenders. These are: (a) Do sex offenders lack victim empathy? (b) Does victim empathy deficit predict recidivism? (c) Do interventions that improve victim empathy reduce recidivism? These questions must all be answered affirmatively for a model of change to be appropriately evidence based.

Do Sex Offenders Lack Victim Empathy?

The relationship between general empathy and offending was examined in depth by Jolliffe and Farrington’s (2004) meta-analysis, investigating differences between offenders and nonoffenders on questionnaire measures of empathy. All the measures investigated were measures of general empathy, both cognitive and emotional. When socioeconomic status and intelligence were controlled for, no differences were found between sexual offenders and nonsexual offenders, causing the authors to conclude that low (general) empathy was not causally related to offending for sexual offenders.

In terms of specific victim empathy, it is well documented that sexual offenders often minimized the harm they have caused the victims of their sexually abusive behavior. Studies examining specific victim empathy deficits in sexual offenders have been interpreted as indicating that such specific deficits do exist (e.g., Fernandez & Marshall, 2003; Marshall & Moulden, 2001). However, the identification of poor empathy for past victims does not necessarily explain previous or future offending. There are many reasons why an offender minimizes the harm that his victim experienced, for instance, to reduce his subjective feelings of shame and guilt. The process of minimizing or excusing the harm one has done to others by one’s transgressions is normative (Maruna & Mann, 2006) and as such is not unique to sexual offenders. So, although (specific) victim empathy deficits may be observable in sexual offenders, this alone does not constitute an argument to address them in treatment, unless these deficits also predict recidivism, or if there is evidence that addressing them reduces recidivism.

Does Victim Empathy Deficit Predict Recidivism?

As meta-analyses provide robust analyses of effect sizes and thus constitute a powerful form of evidence, these highly cited studies seem to be the logical place to start when looking for more conclusive answers to the question concerning the hypothesized relationship between victim empathy deficits and sexual offending. Hanson and Bussière (1996) meta-analyzed the results of 61 studies examining factors related to recidivism in sexual offenders. They found that lacking empathy for the victim of the offense was unrelated to sexual recidivism, r = .03, 95% CI [.00, .06]. The variability between the three studies (total N = 4,670) that directly examined the relationship between victim empathy deficit and recidivisim was not significant, suggesting that they all yielded similar results.

More recently, Hanson and Morton-Bourgon (2004, 2005) meta-analyzed 95 studies of recidivism risk factors for sexual offending, involving more than 31,000 offenders and almost 2,000 recidivism predictions. Their findings have provided the most extensive empirical basis for recommendations about what factors treatment programs ought to target in order to have maximal impact on recidivism rates (e.g., Mann, Hanson, & Thornton, 2010). Within their meta-analysis, Hanson and Morton-Bourgon considered five studies (total N = 1,745) that had included “lack of victim empathy” as a predictor variable. Overall, lack of victim empathy did not significantly predict sexual recidivism (d = −0.08, 95% C.I. [-.21, .06]). Furthermore, the Q statistic was not significant, suggesting that there was no more variability between studies than would be expected by chance. A small relationship was discovered between poor victim empathy and nonsexual recidivism (d = 0.12, 5 studies), and between poor victim empathy and nonsexual violent recidivism (d = 0.19, 3 studies). This meta-analysis in particular has led to speculation that victim empathy work within sexual offender treatment may not be warranted (e.g., Mann et al., 2010; Mann & Marshall, 2009).

However, if the definitions or measures of victim empathy in the studies that contributed to these meta-analyses were flawed or atypical, then the results may not generalize to all measures or definitions of victim empathy. As noted, Hanson and Bussière (1996, 1998) entered data from three individual studies into their meta-analysis. The first, unpublished data from Reddon, Studer & Estrada (1995), appears also to have been included in Hanson and Morton-Bourgon’s later meta-analysis (see below) and involved posttreatment clinical ratings of “remorse/empathy.” The second study, (Schram, Milloy, & Rowe, 1991), studied juvenile offenders and involved data collected either from treatment staff or from treatment files, rated against the variable “Expresses empathy for victim.” The third study, Maleztky (1993), did not contain any reference to an empathy variable when the original report was examined. Hanson (personal correspondence, April 29, 2010) confirmed that the variable coded for victim empathy from Maletzky’s study was “lack of remorse.”

Hanson and Morton-Bourgon (2004, 2005) entered data from five studies into the meta-analysis examining relationship with sexual recidivism. Hanson (personal correspondence, April 16, 2009) provided the following additional information about how victim empathy was defined in each of the five data sets. Smith and Monastersky (1986) contributed data in the form of a clinical rating of “offender acknowledges and understands the negative impact of the offense on victim (empathy).” Kahn and Chambers (1991) contributed data in the form of a clinical rating of “Expresses empathy for victim.” Marques and Day (1996) contributed data from the Californian Sex Offender Treatment & Evaluation Project, in the form of a pretreatment clinical rating of “Offender was concerned with the welfare of his victim yes/no.” Hanson (2003) contributed data from an evaluation of Manitoba’s Secondary Risk Assessment in the form of ratings by probation officers of victim empathy, using a structured scoring guide (the Sex Offender Secondary Risk Assessment). Each offender was scored on a 3-point scale, as having Adequate, Partial, or No Knowledge of Victim Issues (Hanson, personal correspondence, April 29, 2010). Adequate knowledge of victim issues was defined as (a) adequate understanding and awareness of victim issues caused by sexual offense, and (b) offender demonstrates remorse and concern for victim. Rater reliability for the scoring in this study had not been established. Reddon, Studer, and Estrada’s (1996) data took the form of posttreatment clinical ratings of “Remorse/Empathy.”

In all the data sets involved in these meta-analyses, the measurement of victim empathy relied on clinical judgment, and most of the variables were dichotomous in nature. The assumptions that underlie this approach to measurement are that (a) victim empathy is observable, (b) victim empathy either exists or does not, and (c) clinical judgment is a valid, reliable, way of measuring victim empathy. These assumptions are either unsupported or are in fact contradicted by the available scientific evidence. For example, the evidence base for the reliability of unstructured clinical judgment suggests that it is generally no more reliable than chance. Furthermore, the phrasing of some of the “empathy” items indicates a fuzzy or restricted concept of empathy. In Reddon et al., for example, empathy and remorse are conflated, but it is arguably possible for someone to feel one without the other (e.g., to feel regretful about what you have done because of the impact it has had on you and your family, rather than as a result of an empathic understanding of the victim’s experiences). In Marques and Day, the variable measured (“Concerned with the welfare of his victim”) does not actually require the offender to accurately understand the effects of his behavior on the victim, merely to express some concern (which could relate to concern about the effect of the criminal justice processes, the family’s reaction, and so on, rather than necessarily implying they understand the effect their behavior had on the victim). The Kahn and Chambers (1993), Schram et al., and Smith and Monastersky (1986) variables referred to acknowledging victim harm, which may or may not result from a cognitive and/or emotional understanding of the victim’s experience. It seems likely that different raters may have defined empathy in different ways, and yet it seems only the Hanson (2003) study attended in any way to interrater reliability by using a structured scoring guide (there is, for example, no mention of interrater reliability in the Schram et al., 1991, or Maletzky, 1993, reports), but even in the Hanson study, interrater reliability was not actually examined. In addition, the ratings used were sometimes taken pretreatment (e.g., Marques & Day, 1996), sometimes posttreatment (e.g., Reddon et al., 1996), and sometimes were derived from treatment files where the treatment program had been very lengthy (e.g., Maletzky, 1993; Schram et al., 1991). There is little evidence that posttreatment measures of risk factors are reliable, and so this may also have been a limitation of some of the studies included in the Hanson meta-analyses.

In light of these numerous limitations to the definitions and measures of empathy used in the various studies, it may be overly hasty to conclude from Hanson and colleagues’ meta-analyses that deficits in victim empathy are unrelated to sexual offender recidivism.

Do Victim Empathy Programs Reduce Reoffending?

None of the various meta-analyses of sex offender treatment program effectiveness have examined the impact of different treatment components at the level of detail required to draw conclusions about victim empathy intervention. Indeed, even if they had, the fact that the vast majority of sex offender programs incorporate a victim empathy component would mean little likelihood of detecting any differential impact depending on the presence or absence of such a component. However, although not a study focusing specifically on sex offenders, it is relevant to consider Landenberger and Lipsey’s (2005) meta-analysis of the effects of cognitive-behavioral therapy on offenders, examining a range of programs aimed at reducing reoffending. Landenberger and Lipsey meta-analyzed 58 studies, of which 7 reported containing a specific victim empathy component, defined as “Activities aimed at getting offenders to consider the impact of their behavior on their victims.” The meta-analysis showed that the presence of a victim impact component was significantly associated with worse outcomes (smaller effects) of treatment on recidivism. At face value, this finding indicates that it is not necessary to explicitly address victim empathy in order to reduce recidivism and indeed suggests that to do so will reduce the effectiveness of an intervention.

Landenberger and Lipsey’s (2005) meta-analysis only tells us about the likely performance of victim empathy programs generally if the programs included in the meta-analysis were typical of such interventions. In order to properly interpret the findings of Landenberger and Lipsey’s meta-analysis, it would be necessary to locate the individual programs included in the meta-analysis that incorporated victim empathy and examine what actually went on in their delivery of treatment. Without this information, it may be premature to discount all victim empathy interventions. Were we to examine the treatment manuals for each of the seven programs, it is possible that we could take issue with some detail or other about how therapists attempted to enhance empathy. Even if the manuals were uncontroversial, perhaps the program therapists worked in a way that would have increased shame, which is a particular risk with victim empathy intervention and which could impair rather than enhance the ability to empathize with others (e.g., Bumby, Marshall, & Langton, 1999). Unfortunately, the records identifying the relevant studies proved inaccessible. At present, therefore, we cannot discount the finding that victim empathy work seems to decrease rather than enhance the effectiveness of offender rehabilitation, but we cannot explain it either, or be confident that it generalizes to all victim empathy intervention.

Because Landenberger and Lipsey (2005) focused on general offender programs, and meta-analyses of treatment effectiveness with sexual offenders have not examined or commented upon the presence of victim empathy components, we should also examine individual studies of victim empathy interventions with sexual offenders. We found two such studies. First, Pithers (1994) reported on the impact of a victim empathy intervention. In this study, offenders attended a structured empathy-enhancing group, with five components, as follows: (a) The group members described their offenses. (b) The group members read memoirs of abuse survivors and described to the group how their own victims’ reactions may have been similar. (c) The group members experienced audio and videotaped accounts of survivors’ experiences and discussed these in group. (d) Group members wrote narratives of their own offenses describing the abuse from the victims’ perspectives. They then role-played the victimization twice, first playing their own role in the offense and second playing the victim’s role, during which they articulated thoughts and feelings from each perspective. Pithers evaluated this procedure with 20 sexual offenders (10 child molesters and 10 rapists) using the Perspective Taking and Empathic Concern scales of the Interpersonal Reactivity Index (IRI; Davis, 1983), the Cognitive Distortions Scale (Abel et al., 1989), and the Rape Myth Acceptance Scale (Burt, 1980). Changes following treatment were reported for the IRI total score; however, this finding must be disregarded because the IRI subscales are negatively correlated and so cannot be totaled (D’Orazio, 2004). There were no significant changes on individual IRI subscales following the treatment for either rapists or child molesters, but the empathy-enhancing treatment significantly reduced endorsement of rape- and child molestation–supportive beliefs, suggesting some generalization from specific to general empathic responding. However, given the very small sample size, these effects would need to be replicated on a much larger scale to be convincing.

Beech, Fisher, and Beckett (1998) evaluated two versions of the national prison-based sex offender treatment program in England and Wales. The shorter version of the program consisted of around 80 hours of treatment, the longer 160 hours. A main difference between the two versions of the program was the inclusion of victim empathy work; the longer version involved narratives and role-plays from the victim’s perspective, whereas the shorter version involved only one brief exercise explaining why victims sometimes do not report abuse. (The other main difference was an extended relapse prevention component in the longer program.) Tests of pretreatment intergroup differences indicated no significant differences between those who underwent the shorter and longer versions of the program, and as such there is a reasonable possibility that any differences between the groups posttreatment were due to the victim empathy work and/or the relapse prevention work. In fact, both versions of the program brought about similar levels of psychometric change in men who had a low level of need (as determined using a psychometric “deviancy” profile). However, in men with high need, the longer program was more than twice as effective at reducing scores on a combined measure of psychometric measures of minimizations of their own victims’ experiences (the Victim Empathy Distortions scale; Beckett & Fisher, 1994) and of the experiences of child victims of sexual abuse more generally (Children and Sex Scale, Cognitive Distortions subscale; Beckett, unpublished).

Again, however, as with Pithers’ study, we do not know whether the changes observed on these measures had any relationship with reconviction outcome. These studies, therefore, do not provide evidence that victim empathy intervention reduces recidivism in sexual offenders.

The Views of Offenders

Several “service user” feedback studies have asked treatment participants (i.e., offenders) specifically about their experiences of victim empathy work. Unfortunately, these studies tend to be carried out with treatment participants who are still incarcerated or have only recently completed treatment and so do not enable any longer term reflection on benefits. In one such study, Wakeling et al. (2005) reported that the victim empathy component was most frequently quoted as being “the most helpful or enlightening” part of a national prison treatment program, cited by 28 out of 46 participants (61%). The participants in this qualitative study identified three ways in which victim empathy was beneficial. First, they proposed that empathy work enabled self-development: facing the truth and helping them think. There were hints of a cathartic effect, exemplified in the words of one offender: “It helped to talk about the offense and get it out into the open. I felt better sharing the problems with others” (p. 181). Second, participants identified it as a benefit that empathy work enabled them to “take responsibility” for the offense and overcome beliefs that their offending had not been harmful. This effect could be seen as addressing the risk factor of offense-supportive attitudes. Third, the participants identified that victim empathy work increased their determination not to reoffend. This effect is in line with Pithers’ (1999) theory of how victim empathy works. However, an unspecified number of participants in this study also described intensely negative experiences of victim empathy work, using words like “traumatizing,” “shocking,” “heart-breaking,” “upsetting,” and “stressful.” Wakeling et al. concluded that offenders viewed victim empathy work as a “helpful but a difficult process.” This study therefore indicated that victim empathy is both rehabilitative and punishing.

Levenson and Prescott (2009) surveyed 44 civilly committed offenders in Sand Ridge Civil Commitment Center in Wisconsin about their perceptions of treatment. On 5-point Likert-type scales, the value of victim empathy received a mean rating of 4.6 for importance and 3.9 for satisfaction. In relation to other parts of treatment, victim empathy was rated as the second most important treatment component. In a larger-scale study of outpatient treatment participants, Levenson et al. (2009) surveyed 338 offenders, of whom 92% rated “understanding the impact of sex abuse on victims and others” as a “very important” part of their recovery. Only “taking responsibility” received a higher approval rating. Levenson et al. concluded that, in the minds of their participants, “successfully addressing sexual deviance hinges, in part, on fully recognising and acknowledging their problem” (p. 50).

An earlier service user study into community treatment in the West Midlands of England (Garrett, Oliver, Wilcox, & Middleton, 2003) did not ask specifically about the experience of victim empathy, but the authors noted that many of the 32 respondents raised this topic spontaneously: “The majority of subjects would have liked to spend more time discussing . . . victim issues, which suggests that sexual offenders have more interest than commonly assumed in understanding their offending and its impact on victims” (pp. 334-5). This research was greatly limited, though, by a poor response rate from one of the two settings surveyed, and by the likelihood that respondents did not perceive their responses to be anonymous. In this study particularly, therefore, it was likely that the responses could reflect socially desirable responding rather than genuine appreciation of the empathy sessions.

Do the views of service users constitute persuasive evidence of the value of victim awareness work? It is of note that neither of the two treatment components that service users say they value—those designed to enhance victim empathy and enable the offender to take responsibility for his offending—have an impressive evidence base to support their inclusion in treatment (McGrath et al., 2010; see Maruna & Mann, 2006, and Ware & Mann, 2012, for a critique of the concept of “taking responsibility”). Either service users are articulating a mechanism of change that has not been identified in the literature (which is perfectly possible) or they are telling researchers what they think we want to hear, based on their intuition of what common sense would think to be important (also perfectly possible). It is also relevant that three of these four studies involved incarcerated participants, who completed treatment behind bars and who had not yet been released at the time of their participation in the study. They, therefore, are presumably commenting upon the aspects of treatment that they believe have had the most impact on them rather than the aspects of treatment that turned out to be most useful after release. Consequently, their comments may reflect the fact that the participants have the greatest subjective sense of impact from the sessions aimed at generating victim empathy as the techniques used tend to be the most active, emotive, and kinesthetic activities within programs that otherwise mainly involve cognitive therapy techniques. Furthermore, the participants in all three studies described here were treatment completers. We do not know the views of treatment dropouts on the value of victim empathy, nor do we have any accounts of their experiences of this aspect of treatment programs.

Rehabilitation Versus Punishment

Ward and Salmon (2011) distinguished between punishment and rehabilitation practices (Ward & Salmon, 2011). The public may well view sex offender treatment from a punishment perspective, desiring that the offender acknowledges the harm and suffering she or he has caused the victim(s), and in turn, experiences the appropriate distress generated by such knowledge. That is, the guilt, shame, and remorse experienced as a result of becoming aware of the harm inflicted on the victim(s) are seen as a fitting, retributive, punishment for a wrongful act (s)he committed. Treatment professionals, on the other hand, operate within a rehabilitation perspective. Although the practice of punishment involves creating states such as remorse, guilt, responsibility, and blame, the rehabilitation of offenders is concerned with well-being, support, caring, and belonging (Ward & Salmon, 2011). There is an obvious tension between the two perspectives of rehabilitation and punishment, especially as rehabilitative work with sex offenders almost invariably takes place in a context where they are also being punished for their crimes.

Is it possible for an activity to be both rehabilitative and punitive at the same time? It is possible that victim empathy work is less about enhancing the offender’s well-being and more concerned with wanting him to experience remorse over the harm that he has caused. Remorse is a moral emotion that is inevitably distressing, and the intention to create it is arguably a component of punishment, even if the intention is constructive. If a lack of remorse is not criminogenic, an objective for offenders to express remorse may interfere with a focus on evidence-based treatment goals and the objective therefore becomes retributive and punitive. This is clearly a case of conflating an ethical or moral objective with a purely therapeutic task (Ward & Salmon, 2011).

Ward (2010) argued that aspects of victim empathy work involves all five components of punishment: (a) It is authorized by the state; (b) it is intentionally trying to evoke emotions such as guilt; (c) it is reprobative, as it necessarily involves disapproval of the wrongful act based on the assumption the act caused harm; (d) it is retributive, as it is a direct consequence of the wrongful act; and (e) it is harmful, in that it is intended to produce emotions that will be uncomfortable and distressing, such as guilt, as a way of motivating the offender not to engage in the wrongful act again. Ward proposed that it is important that clinicians acknowledge this tension and should understand the justification for, and theories of, punishment. He pointed to (parts of) victim empathy work as an aspect of treatment that most obviously meets the definition of punishment in that the goal is for the offender to accept responsibility—and this process will most likely cause the offender to suffer.

However, Ward (2010) also conceded that punishment as part of treatment does not have to be demeaning or unethical but can be implemented in a respectful way. In his view, treatment techniques that meet the definition of punishment can still be ethically legitimate if the aim is for the offender to redeem himself and be reconciled with his community (the “communicative” theory of punishment; Duff, 2001). This approach to punishment respects human dignity and human rights and is founded on the assumption that the offender has equal moral status to all other citizens. Do victim empathy interventions meet this standard?

Articulating how therapists can lose their way when working with sex offenders, Pither (1997) stated, “To the extent that there is social support for treating sex offenders, it is not because society thinks incarcerated sex offenders have the right to decent mental health services to improve their outlook on life” (p. 35). Because society views sexual offending as “reprehensible and repetitive” (p. 35), it can be a struggle for a treatment provider to continue believing that “abusers are people capable of creating meaningful change in their beliefs and behaviours” (p. 35). Pithers provided some examples of “harsh and denigrating” treatment procedures that denied the human dignity of offenders. One of these examples had occurred in a session designed to enhance victim empathy. A treatment participant had testified that during a victim empathy role-play he was physically forced to sit on the therapist’s lap and that the therapist told the offender to pretend he was masturbating the therapist. When the offender refused, the therapist forced his hand in the motion of masturbating and then threw the offender onto the floor (see Pithers, 1997, p. 40, for a fuller account of this testimony and other examples).

Pithers’ (1997) account of unacceptable treatment practices described a treatment program where the rehabilitative imperative had been abandoned and a purely punitive model had been adopted. Pithers emphasized that the practices in this program did not respect human dignity and were not ethically legitimate. In our view, victim empathy work may be especially liable to such a loss of therapeutic and ethical direction because of the highly emotive subject matter and the frustration that many therapists feel when faced with a client who does not appear to acknowledge the harmful nature of his offending. Because of these dangers, we suggest that a treatment activity that displays features of punishment can only be ethically justified if we are confident that it can be practiced in a way that respects human dignity and actually helps to rehabilitate offenders and reduce reoffending rates. Despite Ward’s (2010) conclusion that punishment can still be ethically legitimate within a treatment framework, Pithers’ account would suggest that victim empathy intervention may be what Ward terms an “ethical flashpoint” (p. 294) where there is more danger than benefit.

Rehabilitation, Punishment, or Correctional Quackery?

Is victim empathy intervention punishment, rehabilitation, or “correctional quackery”? Should it be retained in treatment programs and, if so, on what basis? We contend that correctional professionals should resolve the tension between rehabilitation and punishment by reference to the evidence base for victim empathy work, as outlined at the start of this article. The punitive elements of victim empathy intervention may be acceptable if there is evidence that undertaking this work is ultimately associated with reduced reoffending—benefiting both the offender and society. If there is no such evidence, then the ethics of delivering a potentially punitive experience becomes questionable, and it is legitimate to inquire into the possibility of correctional quackery. This is emphatically not to suggest that correctional treatment providers are engaged in fraudulent practice—this is not implied by the definition of the admittedly provocative term “quackery.” However, both treatment professionals and researchers are also humans and subject to the same biases associated with being human, such as the tendency to prioritize intuition and personal experience over scientific data—or, as may be a possibility in this instance, omitting to ask questions because the answers seem intuitively obvious.

We conclude from our review that the theoretical basis for victim empathy work with sexual offenders is inconsistently articulated, poorly understood, and largely untested empirically. It is still unclear whether, or how, current victim empathy treatment interventions, which usually focus on developing empathy for past victims, generalize to future situations. It may be that developing empathy for the victims of the offenses helps the offender to gain a greater ethical awareness and improves motivation to change and make amends (Ward & Salmon, 2011). Alternatively, there is some precedent in the cognitive therapy literature for work that encourages the client to reappraise past situations and identify different perspectives concerning what happened, as a means to encouraging greater cognitive flexibility in future situations. However, in sex offender treatment, unlike most mainstream cognitive therapy, the client is being asked to reappraise the situation in a way that makes him less rather than more psychologically comfortable. As noted earlier, in Ward and Salmon’s terms this makes aspects of victim empathy work punishing.

Some offenders who have completed treatment programs repeatedly tell us that this work is important and helpful, suggesting that there is a rehabilitative element to this work. It seems rather grand and paternalistic to dismiss their feedback as “telling us what they think we want to hear.” However, other than some offenders’ reflections, there is no other convincing evidence that victim empathy is a necessary component of the rehabilitation of sexual offenders. That is, there is no evidence that poor victim empathy predicts recidivism or that programs that address victim empathy reduce recidivism. The models of change that have been advanced to support victim empathy intervention are insufficiently articulated and lack empirical testing. Therefore, the evidence for empathy as rehabilitation is weak, but at this point, it is not possible to conclude that victim empathy intervention has been found to be irrelevant or harmful. In our view, these questions are still open. Until further research has investigated these questions more rigorously, the possibility of correctional quackery must be acknowledged, and consequently, the current enthusiasm for victim empathy intervention should be tempered.

Implications for Treatment

Because to date, service users have consistently reported that victim empathy is valuable, we believe that some victim empathy procedures should be retained in treatment programs. However, there are three provisos to this recommendation. First, those responsible for treatment programs should respond to the urgent need to conduct more research into the reasons why offenders report that these treatment procedures are helpful. Second, treatment providers ought to recognize that aspects of victim empathy procedures meet the definition of punishment and that some offenders themselves describe this part of a treatment as something that causes them to suffer. A key question, is whether the fact that offenders have caused much greater suffering to others justifies this component of a treatment program. Is there a legitimate argument that causing offenders to suffer is an appropriate rehabilitative activity? Treatment providers should therefore be able to articulate an awareness of the ethical complications of this work. It is our view that treatment should be delivered in a way that fully respects the human rights and dignity of those they are treating and there is some evidence that this imperative can be compromised when it comes to victim empathy intervention. This issue, being a moral and ethical dilemma, needs more open and extensive discussion among professionals in the field.

Third, it is our view that the evidence base for victim empathy procedures does not justify the prominence that this work currently holds within treatment programs. The possibility remains that victim empathy intervention is not rehabilitative. We therefore recommend that programs aiming to reduce sexual offending should devote less time to enhancing victim empathy and more time to addressing the underlying blocks to empathic concern. (See Barnett & Mann, 2012).

Implications for Research

It is obvious that we need more sophisticated research into victim empathy. The current deficiencies in both definition and assessment of empathy will, if they persist, prevent the attainment of stronger empirical evidence. Five directions for future research emerge in particular.

First, there is a need to develop better operational definitions and measurement tools for cognitive and, particularly, emotional empathy. Previous studies have almost exclusively relied on examining clinical ratings of empathy as a dichotomous construct; more sophisticated measures need to be developed. Self-report measures may not be the best approach; a better alternative may be comprehensive structured rating scales. Second, once more imaginative measures have been devised, there is a need to examine the predictive validity of such measures in terms of reconviction outcomes.

Third, relationships between improvements in specific victim empathy and improved general empathic responding should be studied. This research strategy is essential because interventions that improve general empathic responding are less likely to be perceived as punitive by participants. If improved general empathic responding generalizes to improved victim empathy, or if it lowers reconviction rates without the need for specific victim empathy work, then this intervention would be preferred to a victim empathy intervention.

Fourth (and this should be easy given the extensiveness of victim empathy interventions) we recommend more in-depth studies of the experiences of sexual offenders during victim empathy interventions, with the aim of creating improved models of change. And last, we suggest that desistance studies, following up sexual offenders who appear to have successfully desisted from offending, could profitably explore the long-term impact of victim empathy intervention from the perspective of the service user. It is possible that treatment participants praise victim empathy intervention only because they are “telling us what they think we want to hear,” and thus the findings from research to date reflect socially desirable responding. It is important that future research designs specifically address the extent to which socially desirable responding accounts for service users’ accounts of victim empathy intervention. For example, researchers should be unconnected with the treatment program and should consider experimental methodologies that could identify the extent to which sexual offenders believe that it is the “right answer” to say that victim empathy intervention was a particularly useful part of treatment.

Conclusion

Victim empathy has long been seen as an important, even essential, component in the treatment of sexual offending. This review indicates that such enthusiasm is not founded on empirical evidence. On the other hand, studies that imply that victim empathy is not of value are too flawed to be seen as definitive. Whereas the positive views expressed by service users about victim empathy intervention may be affected by demand characteristics (i.e., participants feeling that this response shows them in a better light or that it is what the researchers want to hear), there has been no investigation of this concern either. Overall, we cannot conclude on the current evidence whether victim empathy intervention is rehabilitative, punitive, or an example of correctional quackery. Until this uncertainty is resolved, the current enthusiasm for victim empathy intervention may need to be tempered.