The Internal, Self-Management Dimension assists sex offenders to

• identify high-risk situations leading to abuse;
• analyse seemingly unimportant decisions that allow them to be put into high-risk situations; and
• develop strategies to avoid, or cope more effectively with, these situations.

Since sex offences result not from psychological or medical disorders but from conscious choices, these skills provide offenders with resources that enable them to make better choices in the future.

Offenders are informed that if they learn to recognize high-risk situations and enact alternatives, they will be less likely to choose to abuse. However, since no known form of treatment removes the offender's ability to make choices, some offenders who have taken part in effective treatment programs will choose to reoffend.

The External, Supervisory Dimension of RP facilitates supervision of sex offenders by probation and parole officers. The three functions of this Dimension are

• to enhance the efficacy of supervision by monitoring specific precursors to offences;
• to increase the efficiency of supervision by creating a network of collateral contacts which assists the probation officer in monitoring the offender's behaviour; and
• to create a collaborative relationship with mental health professionals conducting therapy with the offender.

Relapse prevention is based on the assumption that a variety of factors influence whether or not a sex offender will commit another abusive act. The interaction of these factors affects the probability of relapse.

By entering treatment, a sexual aggressor essentially declares an intent not to reoffend. As treatment continues, the aggressor becomes more confident of the ability to handle life's future difficulties. Occasionally, this attitude becomes unrealistic and overly optimistic, leading to a lack of attention to critical risk factors.

As a result of a series of what seem to be unimportant decisions, the offender may encounter a high-risk situation. High-risk situations are defined as circumstances that threaten an offender's sense of self-control and thus increase the risk of relapse.

When an offender deals effectively with a high-risk situation, self-management is reinforced. To the extent that the expectation of successfully handling future high-risk situations remains realistic, the probability of relapse decreases. Offenders are primed for relapse if they believe that surviving one high-risk situation means that they "have passed the test" and can now handle all high-risk situations. Should the offender fail to cope with a high-risk situation, self-management decreases, and there is a tendency to give in passively to the next high-risk situation.

One study examining precursors to sexual aggression found that there was a common sequence of behaviours in the relapse process.⁽⁴⁾ 1. Emotion The first change in the offender's typical functioning was emotional. Typically, offenders found themselves unable to deal effectively with a change in their emotional state.

2. Fantasy The second change involved fantasies of performing sexual abuse. For example, angry offenders may attempt to deal with anger by visualizing themselves sexually degrading a person.

3. Cognitive Distortion Fantasies were converted into distorted thoughts in the third step of the relapse process. Offenders frequently made up rationalizations justifying their soon-to-be committed acts.

4. Plan As the relapse process evolved, offenders refined a plan in their minds that would enable them to carry out their fantasized behaviour. An essential element of the plan was to establish circumstances for the offence that might make the offender appear less culpable.

5. Act In the final step of the relapse process, the plan was acted out.
In this relapse process, the initial change that reliably differentiates sexual aggressors from others is the predominance of deviant sexual fantasies. Thus, for most sexual offenders, the initial return of deviant sexual fantasy is defined as the earliest identifiable lapse.

So far, the relapse process has been described from the point at which a person encounters a high-risk situation. However, RP also examines events preceding high-risk situations. Although some sex offenders lapse in situations that would have been difficult to anticipate, the majority set the stage for lapses by putting themselves into high-risk situations.

Offenders can set up a lapse by making a series of seemingly unimportant decisions, each representing another step toward a tempting high-risk situation. Any single decision may appear unrelated to reoffending but, in reality, each choice brings the aggressor closer to the high-risk situation where the final decision to offend again or not is made.

If sexual offenders have not been prepared to deal with lapses, they may attempt to hide their errors from therapists and parole officers. They may believe that admitting even a momentary deviant fantasy will be considered an indication that they are totally out of control. Any effort to bury a lapse typically leads to additional lapses that are closer still to offending again.

In the RP model, a relapse is the recurrence of deviant sexual behaviour. Several factors, encompassed by the term "Abstinence Violation Effect," influence whether or not a lapse becomes a relapse. A major component of the Abstinence Violation Effect is the conflict between an offender's self-image as an abstaining sex offender and the recent experience of a lapse. This dissonance may be resolved by offenders' deciding that their treatment has failed and that they remain sexual offenders.

Attributing lapses to personal weakness heightens the Abstinence Violation Effect. If lapses are considered personal failures, the expectation for continued failure develops, possibly ending in the ultimate defeat: relapse. The Abstinence Violation Effect is also amplified if the offender selectively recalls only the positive aspects of sexually abusing victims in the past and forgets about the delayed negative consequences.

If aggressors selectively remember positive outcomes of previous offences but neglect the delayed negative consequences (e.g., arrest, incarceration), the probability of relapse increases. Due to the strength of this phenomenon, it has been named the "Problem of Immediate Gratification" or the PIG phenomenon.

A final factor affecting the Abstinence Violation Effect is the individual's expectation about the likelihood of lapsing. For offenders who believe that treatment should erase all vestiges of their deviant desires, a momentary loss of control may be interpreted as an irreversible trend. In contrast, an offender may view a lapse as an expected event representing an opportunity to refine self-management skills through analysis of reversible mistakes. In the latter case, lapses can yield productive outcomes. In such cases, an offender acquires enhanced coping skills and maintains greater vigilance for the earliest signs of relapse.

Relapse prevention begins by dispelling an offender's misconceptions about the outcome of treatment and by describing more realistic goals.

It continues with an assessment of the offender's high-risk situations, which are the conditions under which relapse has occurred or is likely to occur in the future. The initial assessment also examines the offender's coping skills, since situations are considered high risk only to the extent that the person has difficulty coping with them.

After high-risk situations have been identified, interventions are designed to train the offender to minimize lapses and keep them from evolving into a full-blown relapse.

When introducing RP to offenders, we emphasize the development of realistic expectations about therapy and encourage an active, problem-solving approach on their part. We inform them explicitly that there is no cure for their disorder. They are told that treatment may diminish their attraction to abusive sexual behaviour, but that fantasies about this behaviour are likely to recur in the future.

Sex offenders are told that the return of a deviant fantasy does not necessarily signify that they are going to offend again, and that a critical part of treatment involves learning what to do when they feel drawn to repeat abusive sexual activity.

We tell offenders that they will discover a variety of situations in which they will make seemingly unimportant decisions. These decisions will either lead them closer to offending again or take them away from that danger. They are told that developing the ability to recognize these situations and enact alternatives will reduce the likelihood of acting out their deviant fantasies.

Initially, we recommended introducing the offender to RP concepts during the first therapy session. However, we have since discovered that the highly cognitive strategies of RP can heighten offenders' defences against recognizing the harm inflicted upon victims.

Empathy for victims represents a critical source of motivation for the offender's treatment and maintenance.⁽⁵⁾ To avoid having RP viewed by offenders as an interesting intellectual exercise with little relevance to their lives, we introduce it only after victim empathy has developed.

Since RP is a highly specialized approach to therapy, thorough assessment is necessary to determine issues to focus upon in treatment.

Assessment in RP includes three major tasks:

• specification of the offender's high-risk situations (including seemingly unimportant decisions creating those situations);
• identification of existing skills for coping with identified high-risk situations; and
• analysis of early antecedents to the offender's abusive acts.

Analysis of case records, structured interviews, self-monitoring, direct observation and self-report measures are among the methods used to identify risk factors.

While this dimension of RP often works well, sexual aggressors may not use their newly acquired skills. Furthermore, although the importance of acknowledging lapses to therapists and probation and parole officers was repeatedly stressed, offenders sometimes neglected to do so.

Generally, the Internal, Self-Management Dimension of RP was beneficial in enhancing self-control. However, at critical moments that determined the difference between lapse and relapse, this dimension sometimes proved inadequate. A new dimension was therefore created.

Since offenders are sometimes unreliable informants, it was considered essential that other methods of gaining access to information about their functioning be created. To enhance community safety, an External, Supervisory Dimension of the RP model was developed.⁽⁶⁾

Traditionally, probation supervision of sex offenders has been a challenging enterprise, because probation violations noted frequently among many other offenders (e.g., intoxication, neglect of supervision appointments) were rarely noted among sex offenders.

Specification of the precursors to offences and of high-risk factors provides parole officers with identifiable indicators of impending danger of relapse for an offender. Since parole officers monitor specific risk factors related to the offender's sex offences (rather than attempting to monitor all the offender's behaviours, many of which are irrelevant), the efficiency of supervision is increased.

Whenever parole officers detect such a precursor, they determine that the sex offender is involved in a relapse process. Since offence precursors often appear in a distinct sequence (i.e., emotion-fantasy-cognitive distortion-plan-act), the type of precursor exhibited provides an indication of the imminence of potential relapse. The parole officer may then determine the type of intervention required by the offender's lapse (e.g., additional condition of parole, consultation with offender's therapist, parole revocation).

A second element of the External, Supervisory Dimension entails teaching the principles of RP to offenders' collateral contacts. They learn that helping the offender identify factors involved in the relapse process increases the likelihood that the offender will avoid a reoffence. In the offender's presence, network members are encouraged to report lapses to the parole officer or therapist.

Even in sparsely populated regions where meetings with collateral contacts are impractical, information from the network may still be obtained. A checklist of the offender's risk factors can be completed weekly by collateral contacts and mailed to the therapist or supervisor.

Care must be taken in evaluating the ability of collateral contacts to serve this function. For example, a spouse is unlikely to disclose information about her husband if she fears abuse. Employers who treasure the compulsive work habits of some sex offenders may be reluctant to mention information that could lead to the loss of a good employee.

We require the offender to inform network members about offence precursors and risk factors. The probation or parole officer later asks network members to summarize what they were told.

This procedure accomplishes two goals. First, the accuracy and completeness of information presented by the offender can be evaluated, enabling the parole officer to estimate how well the offender understands offence precursors and the importance of others in behavioural maintenance.

Second, informing the extended network about offence precursors destroys the secrecy necessary for commission of sexual aggression. Behaviour that once may have seemed unimportant to others, but which was centrally involved in the relapse process, can then be recognized as signals for concern.

The final element of the External, Supervisory Dimension of RP is the liaison between the probation officer and the mental health professional. Regularly scheduled meetings between these two are essential.

During these meetings, the extent and consistency of the offender's disclosures can be compared. In addition to ensuring that each professional possesses all available information, these meetings also enable detection of the offender's efforts to create conflicts within the supervisory team. Since scheduled meetings allow exchange of routine information, any telephone calls and messages between meetings are considered indications of critical events to be dealt with immediately as opposed to needlessly annoying disruptions in an overburdened schedule.

The combined functions of specially trained probation and parole professionals, collateral contacts and the collaborative relationship between probation and mental health professionals are referred to as the External, Supervisory Dimension of the RP model. Since offenders are not consistently reliable informants, these additional resources are vital to adequate treatment and supervision and, therefore, to the safety of potential victims.

Taken together, the Internal and External Dimensions of RP offer improvements over traditional treatment approaches for sex offenders.

Of the 20 rapists in our follow-up sample at the Vermont Treatment Program for Sexual Aggressors, three (15%) committed an additional sexual assault during the six-year follow-up period. In comparison, of the 147 pedophiles, four (3%) have committed new offences.

A comparison of the proportion of relapses relative to sample size for each of the two offender subtypes (pedophile and rapist) revealed a statistically significant difference [X2 (1, N = 167) = 3.91,p<.05 (statistic corrected for continuity due to low expected frequency of relapses)]. Put simply, in comparison to their overall representation in the study sample, more rapists than expected have relapsed, while fewer pedophiles than expected have relapsed.

In 1978, Sturgeon and Taylor⁽⁷⁾ examined the recidivism status of every sex offender released from treatment in 1973 at California's Atascadero State Hospital, an institution which employed a standard peer-group milieu therapy at that time. Their data revealed that rapists' higher risk of relapse occurred within the first year after release from institutional treatment. In contrast, recidivism among pedophiles was most frequent two to three years after discharge.

Further analysis of Sturgeon and Taylor's data revealed that the proportion of rapists that offend again during the first year after release was significantly larger than the proportion of pedophiles relapsing during that first year [X2 (1, N = 200) = 4.71, p<.05 ]. Thus, rapists appeared to offend again sooner than pedophiles after discharge from treatment. One might question, though, whether these short-term differences in treatment outcome faded with prolonged exposure to risk factors and potential victims.

Analysis of Sturgeon and Taylor's data also revealed that, at the conclusion of the five-year follow-up, rapists and pedophiles had reoffended at nearly the same rate. A statistical comparison of the proportion of the two reoffending groups revealed no significant difference in relapse rates [X2 (1, N = 133) = 0.62,p<.80].

Thus, when RP was not used, rapists and pedophiles had similar reoffence rates by the end of the fifth year after their institutional release. Since differential reoffence rates were found for rapists and pedophiles in our sample at the Vermont Treatment Program for Sexual Aggressors after a six-year follow-up, with pedophiles reoffending at a significantly lower rate than rapists, the discrepancy in these rates appears attributable to relapse prevention.

Differences in the dynamics of rapists and pedophiles may explain both the greater frequency of relapses among rapists during the first year after release from treatment and the discrepancy in the impact of RP on the two offender subtypes.

The highest frequency of rape relapses during the first year after release may reveal the influence of anger and power as the predominant motivations for sexual violence. In individuals who have difficulty dealing with anger and feelings of disempowerment, loss of behavioural control can occur very quickly, and few precursors to an offence may be observed. Thus, rapists may move from adequate self-management to relapse in a relatively short time.

The finding that pedophiles relapse with the highest frequency several years after discharge may be a reflection of their efforts to develop intimacy and relationships. Development of any human relationship, even the profoundly disturbing and coercive interaction of a pedophile and victim, takes time. Pedophiles are more likely than rapists to display precursive risk factors over a relatively long time. These characteristics allow greater opportunity for identification of precursors, therapeutic intervention and restoration of self-control.

While no constellation of interventions prevents relapse, RP appears to hold considerable promise for reducing sex offender recidivism.

Some advantages that RP offers over traditional approaches to sex offender treatment follow:

• a more realistic therapeutic goal of control versus cure;
• reliance on multiple rather than single sources of information about offender behaviour;
• integration of mental health and probation or parole professionals; and
• definition of behavioural maintenance as a continuum rather than an abstinence-relapse dichotomy.