---------------------------

CHAPTER 24

Sexually Harmful Adults

Belinda Brooks-Gordon

Sexually Harmful Adults - References Sexually Harmful Adults - References

This chapter provides a critical reflection of sexually harmful adults. 
First the notion of sexual harm will be considered, then the theories that have been put forward to explain sexually harmful behaviour will be outlined along with the tools that have been developed to assess the risk of sexually harmful behaviour. The treatments that are used in interventions to amend the behaviour of, or rehabilitate, sexually harmful adults will be described, and then an exploration will be undertaken of how, in an era of evidence-based treatment, these interventions are tested for their efficacy.

Sexual harm is also an intensely political issue, so the second part of this chapter provides a critical analysis of some of the main controversies surrounding sexually harmful behaviour. These controversies include 

• (a) the increase over the past two decades in the variety and type of behaviours considered sexual harmful in legislation, 
• (b) the disparity in policy between sexual harm and sexual offending, and 
• (c) how public fear, and the political perception of public fear, can result in the overly mechanist application of tools and policies to manage risk. 

These three issues will be discussed to show how some contemporary measures to reduce risk may be counterproductive and may actually result in greater risk for the vulnerable and others for whose protection the measures were intended.

Who and What Is a Sexually Harmful Adult?

A sexually harmful adult is someone over the age of 18 years of age whose behaviour to another causes harm. Such behaviour may be sexual in and of itself or the behaviour may result in sexual behaviour of another. Sexually harmful behaviours can be divided into contact and non-contact behaviours (Craig et al., 2008). 

• Non-contact behaviours include exhibitionism and the viewing of child pornography.
•  Sexually harmful contact behaviours include rape of a male or female, and sexual assault on a minor under 16 years old (often called ‘child molestation’ and sometimes associated with paedophilia). 

A sexually harmful adult may be male or female, but research indicates that sexual harm is predominantly a male activity with 80–95 per cent of contact sex offences being committed by men. They may be highly intelligent, socially skilled, or have a low IQ, be learning disabled and not fully understand the consequences of their actions (Cantor et al., 2005). [*1]

Prevalence and Incidence [*2] of Sexually Harmful Behaviours

• [*1] These authors found that although adult males who commit sexual offences scored lower on IQ tests overall than adult males who commit non-sexual offences, IQ differences did not occur across all sexual offender subtypes – the younger the victim age, the lower the sample group’s mean IQ. Non-sexual offenders’ IQs equated to general population means.
   
• [*2] The terms prevalence and incidence are not synonymous. How widespread a practice is at a single point in time (its prevalence) and how often it occurs (its incidence) are different entities. 
  A sexually harmful behaviour may be highly prevalent (i.e. be widespread) but have low incidence (i.e. not occur frequently). An example of this is rape by an adult known to the adult victim or so-called ‘date-rape’. Other behaviours may not be so prevalent in the population(s) studied but may have high incidence. An example of this might be abuse by teachers of their pupils, whereby the frequency of the abuse happening may be high to one individual but the prevalence of such abuse in the teacher population (and indeed the pupil population) may be low.

While overall police-recorded sexual crimes increased during the past decade from 33,090 to 53,540 incidents, there was a reduction of 7 per cent in police-recorded sexual offences in 2008 on the previous year. So the process has obviously not been one of continuous increase; indeed, any rise in the most serious sexual crimes (including rapes, sexual assaults, and sexual activity with children) halted with a peak in the middle of the last decade and figures have fallen every year since. 

For example, the most serious sexual crime 
[PAGE 373]
numbered 31,334 crimes in 1997, peaking in 2003/04 at 48,732 crimes and falling every year since to the recent figure of 41,460 incidents in the 2007/08 survey (Kershaw et al., 2008), whereas less serious sexual offences (which include soliciting, prostitution offences, and unlawful sexual activity between consenting adults) increased from 1,756 in 1997 to a peak of 15,320 in 2004/05 following the creation of many controversial new offences in the Sexual Offences Act 2003 (see Bainham & Brooks-Gordon 2004). The number of incidents subsequently fell to 12,080 in 2007/08.

Police-recorded crime, however, is subject to the vagaries of reporting whereas victim surveys such as the British Crime Survey are not affected by changes in reporting, police recording or local police activity, and it has been measured in a consistent way since the survey began in 1981. 

Such surveys show that that a proportion of violent and sexual crimes are not reported. Nonreporting occurs most commonly because victims feel the police could not do anything, or the victim may consider the issue to be a private matter and wish to deal with it themselves. It can be inconvenient to report, or an incident may be reported to other authorities. There may be fear of reprisal or a dislike or fear of the police, especially if there has been previous bad experience with the police or the courts. All of these issues may affect the reporting of sexually harmful behaviour.

While self-report surveys provide a better estimate of hidden crimes such as intimate violence, victim willingness to disclose incidents may depend on the sensitivity of the information, and this can be difficult to disclose face to face. 

For this reason British Crime Survey (BCS) interviews since 2004/05 have included self-completion modules on intimate violence (for those aged 16 to 59 years of age). Based on the 2006/07 BCS self-completion module on intimate violence, approximately 3 per cent of women and 1 per cent of men had experienced a sexual assault (including attempts) in the previous 12 months. The majority of these were accounted for by less serious sexual assaults. Less than 1 per cent of both men and women reported having experienced a serious sexual assault (Kershaw et al., 2008).

Despite increased public and legal awareness of sexually harmful behaviour, the measures above show that some behaviour goes unreported and therefore undetected and unconvicted. Yet research has focused on the subset of individuals who are reported, detected, arrested and convicted. [*3] The evidence base must therefore be interpreted with this bias to the fore (see Crighton & Towl, 2007).

• [*3] The bias is greater in research in England and Wales than in the USA where self-referral and diversion schemes operate for individuals who are at risk, or fear they are at risk, of offending (e.g. see Scheela, 1992).

Theories of Sexually Harmful Behaviour

There are five main theories of why adults sexually harm others. According to Palmer (2008) these are: 

• [1] the preconditions model; 
• [2] quadripartite model; 
• [3] the pathways model; 
• [4] interaction model; and the 
• [5] integrated model.

[1]‘Four Preconditions’ Model (Finkelhor, 1984) 
This is a model to explain sexually harmful behaviour towards a child. The model proposes that there are four steps or preconditions that must take place before an adult commits child sexual abuse: 

• 1) motivation – in which there is sexual arousal towards a child, emotional congruence with a child, and blockage whereby the adult’s sexual needs are not met by a suitable sexual partner; 
• 2) internal inhibition to cause sexual harm must be overcome – whereby self-regulation against the behaviour is overridden by internal factors such as distorted beliefs about the harm caused, or by the disinhibiting effect of alcohol, drugs or extreme stress; 
• 3) external inhibition must be overcome – such as gaining the trust of the child or their family; 
• 4) resistance of the child must be overcome – such through strategies such as force, fear, bribery or other grooming techniques.

[2] Quadripartite Model (Hall & Hirschman, 1992) 
This is a model to explain sexually harmful behaviour to a child or adult and it also accounts for the differences between adults who sexually harm children. The theory suggests that sexually harmful behaviour requires the following conditions: 

• 1) deviant physiological sexual arousal (or preference) to a child; 
• 2) distorted beliefs of children as competent sexual partners able to make decisions about sexual activity; 
• 3) emotional disturbance or lack of emotional management or control; and 
• 4) problematic personality traits and/or vulnerability from own adverse early experiences.

[5] Integrated Theory (Marshall & Barbaree, 1990)
This is a general model to explain all sexually harmful behaviour. It aims to explain the background of such behaviour through early attachment and experiences. It suggests that poor early experiences lead to 

• low self-worth, 
• poor emotional regulation, 
• poor problemsolving and 
• inadequate social coping. 

Such states can all be reinforced by difficult social interactions with peer groups and prospective sexual partners and reinforced 
[PAGE 374]
by other cultural influences such as the media or social norms. All of these developmental, socio-cultural and situational factors make people vulnerable to being psychologically inadequate as well as being susceptible to inappropriate sexual and antisocial behaviour.

[3] Pathways Model (Ward & Siegart, 2002)
This is a theory which combines elements of all the above models into a more complex and comprehensive understanding of sexually harmful behaviour towards children. It is maintained that early life experience, biological factors and cultural influences may lead to vulnerability, which can lead to 

• deviant sexual preferences, 
• intimacy deficits, 
• inappropriate emotions and/or 
• cognitive distortions. 

These four issues can be dismantled into smaller components which are organised into pathways that lead to the abusive behaviour of a child (see Ward et al., 2006 for further analysis of and a unified theory using these components). It is probably the most influential model of sexually harmful behaviour in research and practice today.

[6] Confluence Model of Sexual Aggression (Malamuth et al., 1993)
This peripheral model is a theory, and one of rape only, which suggests that rape occurs when two paths, sexual promiscuity and hostile masculinity, meet and provide the site for rape to take place. It draws on social learning and feminist theory of a certain type of risk-taking, dominant, competitive male who enjoys power to try to explain the role of sexual behaviour in maintaining selfesteem and peer status. However, this theory remains to be validated using a sample known to have committed sexually harmful behaviours, and is less widely adhered to than any of the above models.

[7] Evolutionary Theory of Sexual Offending (Thornhill & Palmer, 2000)
Even more controversially, this model dismisses the influence of culture and psychological strategies of power and control to suggest that rape is a result of evolutionary mating strategies. It puts forward a notion that males have evolved to have profound sexual desire in order to pass on genes, and this manifests itself in the motivation for sexual activity and the need for multiple partners. Its proponents suggest that rape would only be employed when the conditions are favourable; these would include lack of psychological or physical resources, social alienation, limited sexual access to females, and unsatisfying sexual relationships (Craig et al., 2008).

There are, of course, questions to be asked of all of these theories and in all cases explanatory power is limited when tested against the circumstances and victims of sexually harmful acts. For example, the Thornhill and Palmer (2000) theory does not explain why men would rape non-procreative beings such as men or children. The confluence model does not explain why female adults carry out sexually harmful behaviour.

Assessing the Risk of Sexually Harmful Adults

The utitility of risk prediction is that it is possible to prevent the circumstances under which a sexually harmful act may occur. These include the ability: 

• [1] to identify high-risk groups from early antecedents of later harmful behaviour, with a view to providing preventative services; 
• 2) to construct aetiological theories in the view that antecedent correlates of behaviour may equate to causes, and 
• 3) derive predictive information for use in criminal justice decision making, for example in placement or release decisions (Blackburn, 1995). 

A number of psychometric instruments have been devised to determine risk and these divide into actuarial and clinical risk prediction instruments. 

• Actuarial or statistical predictive indices objectively indicate an optimal decision, whereas 
• clinical prediction involves a subjective evaluation of risk based on the client as well as clinical experience of the client group.

Although there continues to be opposition to the use of actuarial risk scales, these scales are in general use in the field of risk assessment and prediction of violent and sexually harmful behaviours and recidivism (Harris et al., 2003).

There are four main actuarial instruments currently in use for sexually harmful behaviour. These include 

• [1] the Violence Risk Appraisal Guide (VRAG; Harris et al., 1998), 
• [2] the Sex Offender Risk Appraisal Guide (SORAG; Quinsey et al., 1998), 
• [3] the Rapid Risk Assessment for Sexual Offense Recidivism (RRASOR; Hanson, 1997), and 
• [4] the STATIC-99 (Hanson & Thornton, 2000).

[1] The VRAG (Harris et al., 1998) 
was developed for use with men known to have committed a violent offence (whether sexual or not) and to predict any new sexual or violent contact offences. It contains 12 items and the item weights are based on the empirical 
[PAGE 375] 
relationship between the predictor and violent behaviours. Individuals are assigned to one of nine risk categories on the basis of their scores. The Hare Psychopathy Checklist (PCL-R; Hare 1993, 2003) is a major part of the VRAG and is the biggest influence on overall score. The PCL-R scores are based on semistructured interviews and a review of the file information. [*4]

• [*4] Clinical files usually contain some or all the following information: 
  • (a) a summary of institutional files, including police records, court records, previous psychological reports, case management reports; 
  • (b) notes from semistructured interview with the miscreant, including information on family background, education, employment, substance use, relationships, mental health, criminal history and future plans; 
  • (c) psychological test results; and 
  • d) treatment reports written by the group therapist and the treatment manager.

[2] The SORAG 
It aims to measure characteristics such as impulsivity, irresponsibility and callousness. The SORAG is a modification of the VRAG, with 14 items (10 common to the VRAG), and is designed to predict violent recidivism in men who have already committed sexually harmful contact behaviour. Again the PCL-R score is the most influential.

[3] The RRASOR (Hanson, 1997) 
was developed for men who have been convicted of at least one sexual offence and is designed specifically to predict sexual recidivism. It has four items: 

• number of prior charges or convictions for sexual offences; 
• age upon release from prison or anticipated opportunity to reoffend on the community; 
• any male victims; or 
• any unrelated victims. 

Items on the scale are weighted to reflect the magnitude of its relationship with sexual recidivism. 

[4] The STATIC-99 (Hanson & Thornton, 2000) 
was developed for men who were known to have committed at least one sexual offence. It is designed to predict either violent recidivism or specifically sexually recidivism. It has ten items, of which four are the same as the RRASOR, and it was constructed by combining the RRASOR with a nonactuarial instrument. On the basis of their score, individuals are allocated to one of seven risk categories.

All the above are actuarial risk scales which are objectively scored and give probabilistic estimates of risk based on the established empirical relationship between their items and the outcome (i.e. sexual harm). Probabilistic estimates suggest the percentage of people with the same score who would be expected to harm sexually within a specified period of opportunity. These scales have good predictive validity and have been cross-validated in new samples of those who sexually harm. The scales contain similar items because they were all empirically derived and their developers drew upon the same sex offender recidivism literature for items (Seto, 2005).

Interventions for Sexually Harmful Adults

Psychological interventions, including behavioural, cognitive-behavioural and psychodynamic therapies, are all used to help change the behaviour of sexually harmful adults. In addition, drug treatment may be given alongside or instead of these therapies. 

• Cognitivebehavioural interventions are the basis of sex offender treatment in prison systems and community programmes in England, Canada, New Zealand and the USA. 
• In the UK National Health Service, however, psychodynamic approaches are common (Grubin, 2002).

Behavioural interventions are associated with traditional classical and operant learning theory and are generally referred to as behaviour modification or behaviour therapy. The hallmark of these interventions is an explicit focus on changes in behaviour by administering a stimulus and measuring its effect upon overt behaviour. 

Within sex offender treatment this is often used to address deviant sexual interest alongside penile plethysmography (PPG) for ‘objective’ measurement. Examples include 

• aversion therapy (exposure to deviant material followed by aversive stimulus), 
• covert sensitisation (imagine deviant sexual experience until arousal and then imagine a powerful negative experience), 
• olfactory conditioning (an unpleasant odour is paired with a high-risk sexual situation) and 
• masturbation satiation/ orgasmic reconditioning (masturbation to an appropriate sexual fantasy).

A range of interventions falls under the heading of cognitive-behavioural treatment. 

These interventions have been characterised on a continuum, in the middle of which are interventions seeking to change the individual’s internal (cognitive and emotional) functioning as well as their overt behaviour (McGuire, 2000). This best represents cognitive-behavioural treatment as that which has developed from social learning theory. 

Finally, there are cognitive therapies in which the focus is exclusively on changing some aspect of the individual’s cognition. 
This approach is arguably more likely to have a base in some variant of cognitive theory, such as information processing, than in learning theory. Cognitivebehavioural treatment attempts to change internal processes – thoughts, beliefs, emotions, physiological arousal – alongside changing overt behaviour, such as social skills or coping behaviours.

Cognitive behavioural therapy is where the intervention involves: 

• (a) recipients establishing links between their thoughts, feelings and actions with respect to target symptoms; 
• (b) correction of persons’ misperceptions, irrational beliefs and reasoning biases related to target symptoms; and 
• (c) either or both of the following: 
  • (i) recipients monitoring their own thoughts, feelings and behaviours with respect to target symptoms, and 
  • (ii) promotion of alternative ways of coping with target symptom/s.

[PAGE 376

Psychodynamic psychotherapy involves regular individual therapy sessions with a trained psychotherapist, or a therapist under supervision. Therapy sessions are based on a variety of psychodynamic or psychoanalytic models. Sessions rely on a variety of strategies, including explorative insight-oriented, supportive or directive activity, and applied flexibly with therapists working with transference (the unconscious transfer of feelings to a person which do not befit that person and which actually apply to another (Greenson, 1967) ). Psychoanalytic interventions are many and various but usually include regular individual sessions with a trained psychoanalyst three to five times a week working at the infantile sexual relations level of psychoanalytic theory.

Drug treatments administered specifically for sexually harmful behaviour or impulses are sometimes incorrectly referred to as ‘chemical castration’ or more correctly pharmacological diminution of an abnormal sex drive. These drugs include 

• leuprorelin (Prostap), which switches off the production of testosterone. High testosterone levels are linked with an abnormally high sex drive in paedophiles. 
• Other drugs, such as cyproterone (Androcur), work in a different way, by opposing the action of testosterone in the body instead of interfering with its production. 

The effect is the same and results in a lowered or absent sex drive and an inability to have sex. Some sexually harmful men are treated with flouxetine (Prozac), an antidepressant that is also prescribed for obsessive compulsive disorders (OCD) along with psychological therapy.

Measuring Interventions

The rise of evidence-based medicine, the philosophical origins of which extend back to mid-19th century Paris, has influenced ways of working with those who sexually harm others. Evidence-based practice is the ‘conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients’ (Sackett et al., 1996). The practice of evidence-based medicine means integrating individual clinical expertise with the best available external clinical evidence from systematic research. 

By clinical expertise, Sackett et al. (1996) mean the proficiency and judgement that individual clinicians acquire through clinical experience and clinical practice. They suggest that when asking questions about therapy ‘we should try to avoid the non-experimental approaches, since these routinely lead to false positive conclusions about efficacy. Because the randomised trial, and especially the systematic review of several randomised trials, is so much more likely to inform and so much less likely to mislead us’ (p.72). Evidence-based practice is a key feature of psychological interventions and rehabilitative therapies for this group of people within the prison service (Crighton & Towl, 2008).

Past Meta-analyses of Interventions with Sexually Harmful Adults

Hall (1995) carried out a meta-analysis of 12 studies that compared sexual offender treatment with a comparison condition (alternative treatment or no treatment) and provided recidivism data for sexual offences. 

• A small overall effect size was found for institutionalised treatment, and 
• a medium effect size for outpatient samples. 

Both cognitive-behavioural and hormonal treatments appeared to be superior to behavioural treatments, although a possible criticism of Hall (1995) is that it may have overestimated the effectiveness of treatment because of its use of official recidivism data, which may underestimate actual sexually aggressive behaviour. 

The results suggested that the effect of treatment with sexual offenders was robust, albeit small, and that treatment was most effective with outpatient participants and when it consisted of hormonal or cognitive-behavioural treatments. A subsequent Cochrane review of randomised controlled trials identified only two relevant studies (total n = 286) and no clear effects of relapse prevention/group therapy (White et al., 2000).

In a later meta-analysis Gallagher et al. (1999) quantitatively synthesised the results of 22 studies (25 treatment comparisons) evaluating the effectiveness of different types of treatment for sexual offenders. Like Hall (1995),

•  cognitive-behavioural therapies were considered promising while 
• less support was found for behavioural, chemical, and more general psychosocial therapies. 

Some of the studies had problematic threats to validity because they involved comparisons between treatment completers and non-completers. This is important, because it is acknowledged amongst practitioners that non-completers are more likely to reoffend.

In Hanson et al. (2002), most of the results were based on matching/incident assignment studies. In incidental assignment studies, comparison groups were selected from offenders in which there were no reasons to expect differences in the treatment group. 

That analysis showed for the first time a significant
[PAGE 377] 
difference between recidivism rates for sex offenders who were treated and those who were not, suggested Hanson’s team. The study revealed, among the research sample, sexual recidivism rates of 

• 17.3 percent for untreated offenders, compared with 
• 9.9 percent for treated offenders. 

Though that is not a large reduction, the large sample size and widely agreed-upon research methods make it statistically reliable and of practical significance.

Hanson et al. (2004) found that examination of individual treatment programmes did not yield any difference in recidivism rates. While the study did not allow conclusions about what was effective or ineffective in the CSOP (Community Sex Offender Program) interventions, the findings do suggest that some ‘highly plausible interventions may have little overall effect’ and the findings of the review certainly contrast with the positive effects of cognitive-behavioural treatment found in previous reviews (e.g. Gallagher et al., 1999, Hanson et al., 2002). 

Controversy remains as to the merits of treatment, and the relevant evaluative research on the effectiveness of psychological treatments for sexual offenders is not straightforward (Quinsey et al., 1998). Disappointing results from interventions in England and Wales have been discussed by proponents of treatment who suggest that greater effects might be present yet only detectable using more sensitive psychometric measurement instead of reconviction data (Friendship et al., 2003).

Losel and Schmucker (2005) reported a meta-analysis on controlled outcome evaluations of interventions for sexual offenders. Allowing a wide remit for acceptable studies, the 69 studies containing 80 independent comparisons between ‘treated’ and ‘untreated’ offenders. Effects for violent and general recidivism were in a similar range. Medical treatments such as surgical castration or pharmaceutical medication showed larger effect sizes than psychosocial interventions. Of the psychological interventions, cognitive-behavioural approaches had the most robust effect. Overall, Losel and Schmucker found a 6 percentage point reduction in sexual recidivism following treatment, compared to untreated controls.

The largest systematic study was carried out by Brooks-Gordon et al. (2006) and also published as a Cochrane review. 
This study reports a systematic review of randomised control trials reporting the effectiveness of sexual offender treatment programmes. Electronic and hand searches were carried out for randomised control trials published between 1998 and 2003. 

Searches resulted in nine randomised control trials (RCTs) and these contained data on interventions for 567 men, 231 of whom were followed up for a decade. Analysis of the nine trials showed that cognitive-behavioural therapy [CBT] in groups reduced reoffence at one year for child-molesters compared with standard care (n = 155). When CBT was compared with a trans-theoretical counselling group therapy, the former may have increased poor attitudes to treatment. 

The largest trial compared broadly psychodynamic group therapy with no treatment for 231 men guilty of paedophilia, exhibitionism or sexual assault. Rearrest over 10 years was greater for those allocated to group therapy. 

These findings and the subsequent use of randomised control trials in clinical policy and research are important, because while they may not be popular with those who want to believe that current interventions work for all sexually harmful men forever, they raise important conceptual questions about interventions for sexual harm. 

For example, do interventions need to be longer term, with community-based relapse prevention (such as Alcoholics Anonymous)? In a parallel study, the same authors (Bilby et al., 2006) also analysed quasi-experimental outcome evaluations and process evaluations in qualitative studies.

Harkins and Beech (2006) review the various research methods of examining treatment effectiveness, and random assignment, risk band analysis and matched control groups are discussed. 
They conclude that different designs confer different advantages and also have methodological shortcomings. While there are those who feel that only the most scientifically rigorous methodology must be employed if one hopes to draw meaningful conclusions, others feel that less stringent criteria in terms of comparison groups can yield meaningful inferential results. 

As a means of overcoming some of the shortcomings of recidivism outcome studies discussed, they suggest that the examination of more proximate outcomes, such as change within treatment, provide a useful addition to studies of treatment effectiveness (Harkins & Beech, 2006). 

The last two meta-analytic reviews reported here represent the two sides of this equation. The strictly defined Brooks-Gordon et al. (2006) review conforms to the rigorous Cochrane collaboration methodology to provide a stringent evidence-based measure which meets the so-called ‘gold standard’. And the other review, by Losel and Schmucker (2005), takes the wider, more pragmatic remit in which treatment usually takes place in England and Wales. [*5]

• [*5] This review may have been unduly weighted by its inclusion of physical/biological interventions such as physical castration and anti-libidinal pharmaceutical treatment.

There are enormous difficulties in carrying out RCTs to evaluate interventions with sexually harmful adults. Randomised controlled trials are often complex and 
[PAGE 378]
difficult to carry out. It is not ethical to carry out doubleblind randomised trials in which neither participant nor clinician knows which group an individual had been allocated to and then wait for the individual to produce sexually harmful behaviour. 

There is also controversy regarding denial of treatment to this population. Given the belief inherent in policy that treatment works, despite weak evidence, there are human rights and ethical implications should a potentially helpful treatment be withheld within a controlled trial, not to mention implications for future potential victims. It is feasible that such a study could deny a sexually harmful adult an effective treatment, and could also affect the security classification of such a prisoner and decisions regarding parole (Friendship et al., 2002). 

Prisoners allocated to non-treatment groups could be denied an intervention package that could affect their chances of early release or re-categorisation to a lower security level (Hood et al., 2002). 

And as Prentky and Schwartz (2006) have pointed out: ‘treatment is not likely to be effective for all offenders and that treatment is likely to be effective for some offenders. 

Essentially, such a conclusion is accurate and, for most of us, obvious. Given the extraordinary variation in sex offenders, it would be only logical that some, but not all, offenders would benefit from treatment. Stated otherwise, treatment undoubtedly will help to restore some offenders to a non-offending lifestyle and will fail to touch other offenders.’ It is therefore important to ascertain with as much accuracy as possible whose behaviour is made less harmful by interventions.

Improving the Quality of Treatment Outcome

A collaboration formed to ascertain the quality of sexual offender treatment outcome research recommended that interventions use strong research designs, including random assignment to treatment and comparison conditions. It was also recommended that offenders are matched on risk prior to being assigned to treatment. 

Random assignment studies are politically unpopular and difficult to implement, but the benefits of these studies are such that researchers should advocate for random assignment studies whenever possible. Researchers using random assignment studies, however, should be prepared for breakdown in the randomisation procedure. Consequently, it was recommended that all participants (treatment and control) should be assessed pre-treatment on risk-relevant variables, and that researchers are vigilant to problems of treatment integrity, attrition, and cross-over (comparison group receiving equivalent services) (Beech et al., 2007).

But many difficulties still hamper researchers wishing to undertake evaluative studies and, as a result, randomised control trials are seldom undertaken in criminal justice settings. In such a situation, the ‘next best’ evidence has to be considered. Not surprisingly, there are a number of pragmatic trials carried out in quasi-experimental designs. In addition, more sophisticated randomised research designs, which address some of the difficulties associated with traditional randomised control trials such as cluster randomisation, might profitably be explored for the evaluation of programmes.

Cluster Randomisation

Increasingly, ‘cluster randomisation’ is employed in medical or educational interventions where traditional randomisation trials are not possible. It is a method whereby clusters of individuals rather than independent individuals are randomly allocated to intervention groups. In the case of those who have sexually harmful behaviour, this could be in secure communities or secure accommodation unit clusters.

This approach has many advantages. The reasons for adopting this method might be: 

• administrative convenience, to obtain co-operation of investigations, 
• ethical considerations, to enhance participant compliance, to avoid treatment group compliance, and/or 
• to apply the intervention naturally at cluster level. 

An important property of cluster randomisation trials is that inferences are frequently intended to apply at the individual level while randomisation is at group or cluster level. Thus the unit of randomisation may be different from the unit of analysis.

Analysis and pooling of clustered data, however, does pose some problems. The lack of independence among individuals in the same cluster (i.e. between-cluster variation of different offender characteristics such as motivation could be due to differences in characteristics of the therapist) can create methodological challenges in both design and analysis. 
Authors must account for intra-class correlation in clustered studies, to avoid ‘unit of analysis’ error (Divine et al., 1992) which pushes down p values and could lead to overestimating statistical significance and to Type I errors (Bland & Kerry, 1997; Gulliford et al., 1999). 

In addition, loss of precision can 
[PAGE 379] 
also occur due to: 

• the intervention being applied on a group basis with little or no attention paid to individual study participants; 
• permitting the entry of new participants to a group (cluster) after baseline; 
• entire clusters, rather than individuals, being lost to follow-up; and 
• over-optimistic expectations regarding effect size.

Precision can be improved in cluster randomisation trials, argue Donner and Klar (2000), 
first by establishing a cluster-level eligibility criterion to reduce betweencluster variability, 
second by increasing the number of clusters randomised, even if only in the control group. Then match or stratify the design by a baseline variable having prognostic importance. 
Also, obtain baseline measures of other possible prognostic variables, and take repeated assessments over time from the same clusters or from different clusters of subjects. Finally, develop a detailed protocol for ensuring compliance and minimising loss to follow-up. 

Interpretational difficulties will be present, and the ratio of the total number of participants required using cluster randomisation to the number required using simple randomisation is called the ‘design effect’. Thus a cluster randomised trial which has a large design effect will require many more participants than a trial of the same intervention which randomises individuals (Kerry & Bland, 1998).

So far this chapter has explored the theories of sexually harmful behaviour by adults, the ways in which risk is assessed in these adults, and if risk is present, what interventions can be used to address and/or change behaviour. 

In addition, the first section of the chapter explored how the effectiveness of interventions is measured. While none of these issues are without controversy, there are major controversies around what actually constitutes sexual harm and what steps are legitimate to monitor risk. The next section of this chapter proceeds to discuss these two major issues.

When the ‘Sex Offender’ Is Not Sexually Harmful

There has been a marked increase over the past two decades in the variety and type of sexual behaviours criminalised in legislation as policy makers have responded to calls for action with unparalleled increases in legislation. This occurred most notably in the Sexual Offences Act 2003 and has continued ever since, most recently in the Criminal Justice and Immigration Act 2008. 

The review of sexual offences, Setting the Boundaries, took the view that what was wrong in sexual relationships depended upon the principle of harm and an assessment of harm done to the individual. However, that which is ‘sexual’ was defined in the Sexual Offences Act 2003 to be what ‘a reasonable person’ would consider to be ‘sexual’ (s.70), and this broad statutory definition is highly problematic.

There is a difference between being sexually harmful and being a sexual offender. 
For example, in a year when there were 25,000 offenders on the sexual offenders register, only 26 had committed a serious sexual or violent offence (Daily Telegraph, 2004). 

One of the reasons for this is that the number of new sexual offences on the statute book has increased, many of which have little to do with sexual harm and more to do with regulating behaviour or appeasing a perceived public understanding of safety. There were 62,081 recorded sexual offences in England and Wales from 2005 to 2006 (Walker et al., 2006) but not all sexually harmful behaviours will necessarily be recorded as offences, and many sexual offences will not be sexually harmful.

Women working in prostitution who are good parents and adults may, because of their working lives in the commercial sex industry, end up with a conviction for prostitution. This conviction now constitutes a sexual offence. Those working in the sex industry who, because they wish to keep their working lives separate from their home lives, or for reasons of safety or company wish to work with other sex workers, prefer to rent rooms or work in a sauna or ‘parlour’. Yet those renting out rooms or running parlours (often ex-sex workers or past receptionists, known as ‘maids’) now fall foul of the laws on brothel-keeping and on conviction are subject to a two-year sentence according to Sentencing Guidelines Council recommendations under laws of brothel-keeping. This too counts as a sexual offence. These two examples show how misleading the statutory categories are.

The ultimate sanction of the criminal law, however, should be used sparingly and only for those behaviours which are demonstrably harmful. There has a been a highly punitive legislative approach resulting in the criminalisation, as opposed to toleration of, sexual diversity (Bainham & Brooks-Gordon, 2004) and this has created a situation whereby what is a sexual offence is not that which is necessarily sexually harmful. And this situation even extends to so-called ‘child offences’ whereby two young people under the age of 16 years engaged in mutual sexual experimentation are breaking the law.

[PAGE 380]

The Politicisation of Sexual Harm

The electorate and the media expect policy makers to be accountable for the effects of their responses to sexual offending (Hansard, 2002). The politicisation of sexual harm, following a number of high-profile child murders and with politicians aware of global media coverage, added to the statute book new laws in England and Wales and new punitive measures to old laws, in an uptariffing and redefining of what is sexually harmful. 

In the USA, such concerns led to Megan’s Law in 1996, which allows private and personal information on those registered as sex offenders against children to be made available to the community (Office of the Attorney General, 2004). One of the many examples of political or populist punitiveness in England and Wales came in the Sexual Offences Act which also redefined ‘child’ as a person under 18 years of age, which would have the effect of criminalising a same-age couple sending pictures to each other over the internet.

In the UK, provisions in the Sexual Offences Act 2003 included substantial increases in sentence length for many sexual offences and increased state control, in terms of notification requirements and supervision, for up to 10 years after a sentence has been spent. Exaggerating the risk of sexual harm is problematic and may increase public fear, and stigmatise and hinder rehabilitation of offenders who have changed their lifestyles, while wasting valuable resources on unnecessary surveillance (Soothill et al., 2000).

Sexual Harm and the Culture of Fear

From 2002, anyone who worked with children in any capacity (even as a volunteer) had to be vetted by the Criminal Records Bureau (CRB). This followed the deaths of Holly Wells and Jessica Chapman who were murdered by the school caretaker Ian Huntley, and suspicion of grown-up behaviour towards children has fostered a climate in which it has become normal for some parents to trust only adults who possess official clearance. In some quarters it is argued that the culture of ‘vetting’ is damaging relationships with adults and children, and the moral panic over paedophilia has arguably become a panic bordering on hysteria.

The ‘War on Terror’ has polarised ideas of security vs. liberty; this, a political practice, a speech act, is one way of framing, naming and constructing problems. It seeks to mobilise forces behind the idea that ‘we’ face a threat that calls for immediate decisions and special measures, argues Loader (2006). Legislative hyperactivity articulates genuine public insecurities about crime, immigration and social disorder. It generates a climate that inhibits, even actively deters, critical scrutiny of the state’s claims and practices, and risks fostering a vicious circle of insecurity (atrocity > fear > tough response  > atrocity) that ratchets up state powers in ways that become difficult to temper, dismantle and reverse.

In a report called Licensed to Hug for the think-tank Civitas, Furedi and Bristow (2008) show how adults have become less inclined to volunteer for mundane activities such as school trips, cricket umpiring or football coaching because they are fearful of being thought predatory or sexually harmful adults. The authors state that the whole notion of harm requires re-examination as ‘child protection policies are poisoning the relationship between the generations and damaging the voluntary sector’ (Furedi & Bristow, 2008). In addition, parents were sceptical about the efficacy of the vetting procedure and felt it was burdensome and confusing.

Institutionalisation of the vetting of adults with CRB* checks
[*Criminal Records Bureau]

From October 2009, adults will have to register with the new Independent Safeguarding Authority (ISA), so it is estimated that one in four (or 11.3 million) people will be affected by the scheme. The alleged protective effects of a system of vetting are considered to be largely illusory by Furedi and Bristow (2008), who conclude that the national vetting scheme represents an exercise in impression management rather than offering effective protection. 

Aside from the fallibility of record-keeping and technical systems, vetting only takes account of what someone has done in the past; it cannot anticipate what they may do in future. The situation de-skills adults, who then also have a diminished sense of responsibility towards children. 

Adults feel increasingly nervous around children, unwilling to exercise authority or play a positive role. Such intergenerational unease has not made children safer in the past but rather creates conditions for greater harm, as adults lose the nerve or will to look out for any child that is not their own (see also Furedi, 2005). Perversely, it encourages adults to avoid their responsibility to look out for the well-being of children in their community. 

Thus the policy of attempting to prevent paedophiles coming into contact with children will result in the estrangement of all 
[PAGE 381] 
children from all adults  –  the very people who would otherwise protect them from paedophiles and other dangers.

While fear and risk were explored in the 1990s through ‘ontological insecurity’ in a risk society (Beck, 1992; Bauman, 2000), Walklate and Mythen (2008) take this further to argue that the view has developed to the extent that if risk constitutes one side of ontological security, then trust and trust relationships comprise the other. They question the extent to which forensic research has fully accounted for the impacts of global structures and processes that shape human agency on the formation of individual anxieties, and state that not withstanding recent efforts of Gadd and Jefferson (2007), 

• ‘fear-of-crime’ research ‘has not adequately appreciated the full macro climate of this doubt and uncertainty or the way in which it articulates with localised individual experience’ (p.215). 

Governments become embroiled in the politics of fear, and widespread practices and processes result in the paralysis of the ‘culture of fear’ where not just individuals but whole communities are being scrutinised and surveyed as subject identities are made up through risk discourses, socially de termined according to whether they fit the profile of offenders or innocents. A situation then reigns whereby the risk of sexual harm is greater than before.

Conclusion

This chapter has explored sexual harm by outlining firstly the theories that have been developed to understand the behaviour. It proceeded to discuss the interventions currently used to rehabilitate sexually harmful behaviours, and how those treatments have been evaluated and measured, before going onto to discuss the difficulties in measurement and evaluation in meta-analytic studies, the ethics of such analysis, and ways of overcoming such pitfalls.

The second part of the chapter shows how sexual harm has become a major challenge for social policy, and the social and political panic around sexual offending and sexual harm is discussed, showing how more laws were introduced to appease perceived public disquiet. The law is now so encompassing that even consensual adult activity in private has become confused in statute with sexual offending. In this way the law fails to equate sexual harm with sexual offending, and the result is confusion, conflict and further strain on stretched resources. The consequence has been to foster a climate of concern around children that has resulted in fewer activities for them to do, less freedom, and therefore less ability to deal with risk. The resulting culture of fear is, it is argued, a threat to security and safety.

Further Reading

Ward, T., Polaschek, D.L.L. & Beech, A.R. (2006). Theories of sexual offending. Chichester: Wiley. This text provides a good overview of the theories of sexually harmful behaviours.

[PAGE 382

Brooks-Gordon, B.M., Bilby, C. & Wells, H. (2006). Sexual offenders: A systematic review on psychological interventions. Part I: Quantitative studies. Journal of Forensic Psychiatry and Psychology, 17(3), 442–466.

Bilby, C., Brooks-Gordon, B.M. & Wells, H. (2006). Sexual offenders: A systematic review of psychological interventions. Part II: Qualitative studies. Journal of Forensic Psychiatry and Psychology, 17(3), 467–484.

These two studies provide the most comprehensive systematic review of the effectiveness of interventions for sexually harmful adults. The first study is a meta-analysis, followed in the second study by an analysis of lesser data in quasi-experimental studies, and all of the qualitative research on psychological interventions. Between them, these studies provide a comprehensive overview of what is known about the efficacy of treatment at the current time to stringent Cochrane Collaboration standards.

Furedi, F. & Bristow, J. (2008). Licensed to hug. London: Civitas.

This study provides a critical look at the culture of fear around sexual harm.

References

Sexually Harmful Adults - :  Sexually Harmful Adults - References