21.3 ID AS A RISK FACTOR FOR OFFENDING
One of the main pieces of work to influence views on intellectual disability as a risk factor for offending has been that of Hodgins (1992). She used the census cohort of 15,117 people born in Stockholm in 1953 and still living there in 1983. Only participants living in Sweden at the 30-year follow-up period were included in the analysis. She had available to her contemporaneous information on admission to a psychiatric ward and most serious diagnosis on discharge from the psychiatric service, including schizophrenia, affective disorders, paranoid states, alcohol and/or drug abuse and other mental disorders. Intellectual disability was defined “as those who are placed in special classes for intellectually deficient children in high school and were never admitted to a psychiatric ward” (Hodgins, 1992, p. 478). While around 29% of men with no disorder and no disability had been convicted of an offence, among men with ID around 50% had been convicted of an offence. Looking at convictions in two or more age groups, 10.8% of men with no disorder fell into this category while 32.7% of those with ID had offences at two or more age groups. Offenders with no disorder had an average number of 7.3 convictions and men with ID an average number of 10 convictions. Findings for women were similar.
Lindsay and Dernevik (2013) have questioned this classic study on a number of bases. Firstly, those of us who work in generic ID services have no concerns that half of the men we work with have committed an average of 10 crimes for which they have been convicted. When one looks at the types of offence recorded the data seem even stranger. Of the men with ID who have committed crime, 25% are recorded as committing a traffic offence. Most of the studies in the field note that road traffic offences are relatively uncommon in this group (Lindsay et al., 2010, Raina, Arenovich, Jones, & Lunsky, 2013). Furthermore, Hodgins’ (1992) methodology did not take account of the Swedish education system at the time, which included special schools and special hospitals for people with ID. Therefore, the sample was more likely to be made up of people who had attended special classes in mainstream schools for reasons of intellectual limitations or behaviour disorders.
21.4 ASSESSMENT OF OFFENDERS WITH ID
While it is so obvious that it hardly needs to be said, it is a basic requirement that the process of assessment must be understood by the clients completing it. Since people with ID are typified by significant deficits in literacy and comprehension skills, all assessments must be suitably adapted to simplify the language and concepts employed. Unfortunately, there are examples where this obvious tenet has been ignored. For example, it is fairly common practice in a research study on assessment in prisons to distribute a number of assessment schedules for prisoners to complete on their own in their cells. However, given the difficulties with reading and writing, this procedure excludes those offenders who have ID. Therefore assessments have to be altered to cater for the linguistic limitations of the client group (Lindsay & Skene, 2007).
PHOTO 21.1 It is a basic requirement, that the process of assessment must be understood by the clients completing it.
Source: © sinseeho/iStockphoto
A second important consideration in the context of lengthy forensic assessments is that because of literacy deficits, all material will have to be read and explained to respondents. Therefore both the item and the response categories require explanation. These issues have been dealt with extensively elsewhere (e.g. Lindsay 2009; Taylor & Novaco, 2005) and it is not our intention to review them in detail here. However, this has two consequences. Assessment will take much longer and it is impossible to give the respondent a series of questionnaires and receive them back the following week. Secondly, because the assessor has to read the item and explain their responses, all such assessments take the form of a structured interview. The client’s reactions to questions, and emotional responses will all be available to the assessor as part of the process. This is a significant amount of additional information and can add to the richness of any assessment process and report.
Because both the assessment instrument and the process require a different approach, the adaptations for assessment of offenders with ID are extensive. Following from this, it is important that the psychometric properties of the assessment remain intact and that the integrity of the process is not undermined or reduced by these adaptations. Several authors have now reviewed the psychometric properties of adapted assessments that can be used with offenders with ID.
Keeling, Rose, and Beech (2007a) investigated the psychometric properties of adapted versions of a number of assessments. One of the difficulties of their study was that it was a convenience sample drawn from the population of special needs offenders of the Australian Correctional System. This population was more diverse than offenders with ID and included participants with acquired brain injury, significant literacy deficits and communication deficits. Although the population was predominantly individuals with ID, it also included men with borderline intelligence and even with average IQ (with acquired brain injury). In order to assess the validity and integrity of the adaptations, they compared these individuals with 53 mainstream sexual offenders. They found that the Social Intimacy Scale, the Criminal Sentiment Scale, and the Victim Empathy Distortions Scale broadly retained their psychometric properties after adaptation and simplification. The least successful adaptation was in the Relationship Scale Questionnaire, which had a low internal consistency. Test-retest reliability was high and there were good correlations between the original and adapted versions, especially for the Social Intimacy Scale and the Victim Empathy Scale.
Williams, Wakeling, and Webster (2007) assessed the psychometric properties of six self-report measures with sex offenders with ID. These instruments were: the Sex Offender Self Appraisal Scale (Bray, 1996); the Sex Offender Opinion Test (Bray, 1997); the Victim Empathy Consequence Test (Offending Behaviour Programmes Unit, unpublished); the Self-Esteem Questionnaire (Webster, Mann, Thornton, & Wakeling, 2007); the Emotional Loneliness Scale (from Russell, Peplan, & Cutrona, 1980); and the adapted Relapse Prevention Interview (from Beckett, Fisher, Mann, & Thornton, 1997).
Their population was 211 men who had undertaken an HM prison service adapted sex offender treatment programme. The average IQ was 71.9 and the accepted participants had an IQ up to 80, well outside the range of ID. However, the literacy skills seemed similar, in that all the assessments and questionnaires were read to participants who were then helped to complete the answers. The assessments had good internal consistency with alpha coefficients in excess of 0.75. These two examples, (i.e. Keeling et al., 2007a; Williams et al., 2007) demonstrate the way in which existing assessments can be adapted to suit offenders with ID. The following sections will describe specific assessments for specific problems.
21.4.1 Assessment of Anger Motivated Offenders
The clinical and research literature on offenders with ID has demonstrated that anger and aggression are the most common problems (Didden et al., 2003; Lindsay, Michie, Whitefield, Martin, Grieve, & Carson, 2006; Lindsay et al., 2010; O’Brien et al., 2010). Several authors have developed assessments for this frequent problem. Oliver et al. (2007) reported that the Modified Overt Aggression Scale, an informant-rated measure of the frequency and severity of aggression, had high levels of inter-rater reliability when administered to a small number of people with ID as part of treatment outcome research study. Willner, Brace, & Phillips (2005) developed the Profile of Anger Coping Skills to assess the extent to which people with ID use specific skills in managing anger situations. Informants were asked to rate clients’ use of eight anger management strategies in specific anger coping situations. The strategy included relaxation skills, counting to 10, walking away calmly, requesting help, using distraction activities, cognitive reframing and being assertive. Willner et al. found that the assessment had acceptable psychometric properties and was sensitive to change associated with anger intervention. One of the most important aspects of the study was that it directly assessed the use of anger management strategies,
illustrating that participants employed the skills taught. It may be thought that counting to 10, and relaxing, were the most common pieces of advice given during anger management; however, these two skills were the least frequently used in the assessed range.
Novaco and Taylor (2004) evaluated the reliability and validity of several modified anchor assessment measures with ID detained offenders: the Novaco Anger Scale (NAS: Novaco, 2003); the Spielberger State-Trait Anger Expression Inventory (STAXI: Spielberger, 1996); and the Provocation Inventory (Novaco, 2003), a self-report anger reactivity scale; along with the Ward Anger Rating Scale (WARS, Novaco & Taylor 2004). The modified anger self-report measures were found to have high internal consistency and reasonable test-retest reliability. There were substantial inter-correlations between the NAS and STAXI providing evidence for concurrent validity. The WARS staff ratings of patient anger correlated significantly with the self-report measures and these self-reports were significantly correlated with records of assault behaviour. The NAS significantly predicted whether the patient had physically assaulted others following admission to hospital and the total number of assaults carried out. Therefore there was evidence of significant concurrent validity for the anger measures.
In a further development, Taylor et al. (2004) developed the Imaginal Provocation Test (IPT) as an additional ideographic anger assessment procedure with people with ID. This samples key elements of the experience and expression of anger and was sensitive to change associated with anger treatment. The IPT produced four indices relevant to the individual clients’ experience of anger: anger reaction; behavioural reaction; a composite of anger and behavioural reaction; and anger regulation. The IPT had good internal reliability and correlated positively with the NAS and the STAXI. Alder and Lindsay (2007) also produced a provocation inventory – the Dundee Provocation Inventory (DPI) – which is easily accessible and easy to use. It is based on Novaco’s (1975, 1994) analysis and construction of anger as an emotional problem. One of the facets of Novaco’s analysis is that the individual may misconstrue internal and external stimuli and respond to a perception of theft rather than a more appropriate, less aggressive response. The DPI was administered to 114 participants with ID, and Alder and Lindsay (2007) found that it had good reliability and convergent validity. The DPI correlated significantly with the NAS (r = .57) and with the PI (r = .75) indicating that the DPI and PI have good convergence. They also found a five-factor structure consisting of threat to self-esteem, external locus of control, disappointment, frustration and resentment. The strongest factor was threat to self-esteem and this certainly accords with Novaco’s analysis of anger and its relationship with threat. Therefore, the DPI may provide a quick assessment of provocation in relation to a range of relevant factors in offenders with ID.
21.4.2 Assessment of Sexual Offenders
Keeling et al. (2007a), and Williams et al. (2007) have developed or adapted assessments that can be used with sex offenders with ID with some confidence (as outlined above). Griffiths and Lunsky (2003) have developed the Social Sexual Knowledge and Attitudes Tool (SSKAT) revised, that is designed to assess knowledge and attitudes in a range of areas related to social and sexual relationships. These include parts of the body, adolescent development, romantic relationships, sexual relationships, procreation, birth control, sexually transmitted diseases, use of alcohol and illegal substance use and maintenance of sexual health. The SSKAT was used by Michie et al. (2006) in an assessment of sexual knowledge in sex offenders with ID. They tested a hypothesis of counterfeit deviance first proposed by Hingsburger, Griffiths and Quinsey (1991) referring to behaviour that is undoubtedly deviant, but may be precipitated by factors such as lack of sexual knowledge, poor social and heterosexual skills, limited opportunities to establish sexual relationships and social naiveté, rather than a preference or sexual drive towards inappropriate individuals or behaviours. If this is the case, remediation should focus on educational issues and developmental maturation rather than inappropriate sexuality. Griffiths, Quincy and Hingsburger (1989) give a number of examples illustrating the concept of counterfeit deviance.
Michie et al. (2006) compared sex offenders with controls and found that ID sexual offenders had higher levels of sexual knowledge than ID controls. In a second comparison, 16 sex offenders were compared with 15 controls. These authors found a significant positive correlation between IQ and the SSKAT total score for the control group but no significant relationship between IQ and the SSKAT for the sex offender cohorts. Michie et al. suggested two possibilities for the first finding:
By definition, all of the sex offender cohorts have some experience of sexual interaction. It is unlikely that these experiences of sexual interaction are random and one might therefore conclude that these sex offenders have given some thought and attention to sexuality at least in the period prior to the perpetration of inappropriate sexual behaviour and sexual abuse. Therefore, we can be sure that they have at least some experience of sexual activity, which is not the case for the control participants.
The second possible explanation was that these individuals might have a developmental history of increased sexual arousal. This may have led to selective attention and interest in sexual information gained from informal sources. Such persistence of attention would lead to greater retention of information through rehearsal and perhaps to a higher level of associated appropriate sexual activity such as masturbation. These behavioural and informal educational experiences would lead to a higher level of sexual knowledge. In this latter hypothesis, sexual arousal and sexual preference might be considered to interact with knowledge acquisition and perhaps, attitudes and beliefs.
In a second test of counterfeit deviance, again using the SSKAT, Lunsky et al. (2007) compared two categories of sexual offenders and control participants. They divided the offenders into deviant persistent offenders (those who committed contact sexual offences and offences against children) and naive offenders (public masturbation, indecent exposure). They found that the naive offenders did indeed have lower levels of sexual knowledge than the deviant offenders, although the naive offenders did not have poorer knowledge than the control group as might be expected from the counterfeit deviance hypothesis. However, the fact that they found differences between the two groups of sexual offenders suggested the importance of the assessments of sexual knowledge.
Cognitive distortions (distorted attitudes that legitimise offending) have, for some time, been considered important in the pathways leading to a sexual offence and consequently, in the assessment of important factors for sexual offenders. A number of assessments have been developed for mainstream offenders (such as the Bumby’s (2006) RAPE and MOLEST scales), but the language is overly complex for people with ID (Kolton, Boer, & Boer, 2001).
The most widely used assessment of cognitive distortions for sex offenders with ID is the Questionnaire on Attitudes Consistent with Sexual Offending (QACSO), which has been designed for and standardised on offenders with ID (Lindsay, Whitefield, & Carson 2007). The QACSO contains a series of scales that evaluate attitudes across rape and attitudes to women, voyeurism, exhibitionism, dating abuse, homosexual assault, offences against children and stalking. They found that six of the seven scales in the QACSO were valid and reliable measures of cognitive distortions held by sex offenders with ID (the exception was homosexual assault). Lindsay et al. (2006) also found that the rape and offences against children scales in particular discriminated between offenders against adults and offenders against children. The former had higher scores on the rape scale and lower scores on the offenders against children scale while the latter had significantly higher scores on the offenders against children scale. Therefore, it would appear that cognitive distortions in sex offenders with ID can be assessed with some reliability and validity.
Lindsay and colleagues were cautious when considering the relationship of cognitive distortions to risk. They wrote that changes in attitudes may reflect a number of processes such as suppressio
n, the influence of social desirability and even lying. The recommended that the results from the QACSO should be considered in relation to a range of risk assessment variables, including actuarial risk, socio affective functioning and self-regulation abilities.
21.4.3 Assessment of Arsonists (Fire-setters)
Despite the importance of this issue for individuals, communities and the courts, there are relatively few published studies regarding the assessment and treatment of fire-setters with ID. Very few measures specific to fire-setting behaviour have been developed and evaluated in terms of reliability and validity.
As in any forensic clinical context, it is important to systematically collect and collate information from a range of sources (self- and informant reports, behavioural observations, file reviews, clinical assessments) in order to formulate risk, needs and intervention plans. Taylor, Thorne and Slavkin (2004b) reported on the use of a newly designed Pathological Fire-Setters Interview (PFSI). This is a structured interview, augmented with collateral information from patients’ records, staff observations and patient and informant completed clinical assessments, that gathers information in a functional analytic framework (cf. Jackson, 1994) concerning: (a) demographic, personal, family and offending history details; (b) personal setting conditions; (c) situational setting conditions; (d) antecedents to fire-setting; (e) motives for setting fires; and (f) consequences (thoughts, feelings and actual) of setting fires.
Forensic Psychology Page 96