18.6.3 The Think First Programme
18.6.4 The Thinking Skills Programme
18.7 EVALUATION OF GENERAL OFFENDING BEHAVIOUR PROGRAMMES
18.8 ISSUES RELATED TO OFFENDING BEHAVIOUR PROGRAMMES 18.8.1 Engagement and Programme Non-completion
18.8.2 Targeting and Selection
18.8.3 Manualised Groupwork
18.9 SUMMARY
LEARNING OUTCOMES
BY THE END OF THIS CHAPTER, YOU SHOULD BE ABLE TO:
Understand why offender risk assessment and offending behaviour treatment delivery are important in the management of offenders
Distinguish the methods of risk assessment for different types of offenders
Appreciate the principles of effective offending behaviour programmes and how issues relating to delivery may impact on effectiveness.
18.1 INTRODUCTION
During the latter quarter of the twentieth century, the What Works movement contributed a plethora of evidence regarding the effectiveness of different ways of working with offenders. Perhaps the most influential evidence came from a number of collective reviews of large numbers of studies, using the method for integrating research findings known as meta-analysis. These evaluations have advanced our knowledge concerning the components of interventions most effective in reducing recidivistic behaviour and have been particularly influential in shaping modern day correctional service provision. As a consequence, offending behaviour programmes, and the associated practice of risk assessment, are now commonplace within many correctional services of the world, including those within the UK, Europe, Australasia, and North America.
Andrews, Zinger, Hoge, Bonta, Gendreau, & Cullen’s (1990) meta-analysis of intervention evaluation studies is perhaps one of the most prominent of the What Works papers to date. Within this and subsequent publications, Andrews and colleagues founded the evidence-based principles of risk of recidivism, need and responsivity of offender programming, which have since influenced the design and implementation of offending behaviour programmes. These shall each be discussed in turn.
The risk principle dictates that to impact on offending behaviour the level of intervention received by an offender should depend on the level of risk that s/he poses. Offenders assessed as high-risk cases should, therefore, receive a greater level of intervention than offenders who are low risk. Clearly, the accurate assessment of risk is necessary to allocate offenders to individually appropriate levels of intervention.
While the risk principle is concerned with the dosage of programming, the need principle determines the targets of treatment. Andrews and Bonta (2010) make the distinction between general problems or needs present within the offenders’ lives, and those linked to their offending behaviour. The need principle proposes that to reduce recidivistic behaviour, interventions should target only those needs (or risk factors) that contribute to offending behaviour. Such factors have been termed criminogenic needs.
The third principle, responsivity, is concerned with how the intervention is delivered to the offender. The responsivity principle states that those programmes that successfully match the style and method of delivery to the learning styles of the attendees will be more effective. The issue of responsivity will be discussed in more detail later in this chapter. We will first look into the theory and practice of risk, and need, assessment before considering some examples of tools commonly used by correctional services in North America and the UK.
18.2 RISK ASSESSMENT WITHIN OFFENDER MANAGEMENT
Risk assessment within offender management operates on much the same principles as risk assessment within other fields. When you apply for insurance, for example, your insurance company generally requests particular information about you and your personal circumstances. This information is used by the company to judge the likelihood of the outcome they are insuring against occurring. In general, the higher the likelihood of the outcome or, to put it another way, the more risky it is for the company to insure you, the more expensive your premium will be. Consider the cost of a car insurance premium, for example. A 35-year-old driver with 15 years of driving experience, no motoring convictions and no previous car insurance claims would be able to obtain a lower insurance quote than a 21-year-old with two years’ driving experience, a conviction for driving over the speed limit and a previous insurance claim to cover repairs to the vehicle acquired in an accident involving the car and a lamp post! It is easy to see that the likelihood of a future claim is higher in the second scenario than the first: the driver has less experience of driving, has displayed risky driving behaviour and has a history of crashing the car. The insurance premium is thus adjusted to reflect the higher risk to the insurance company of having to finance a future claim.
In much the same manner, the risk assessment of offenders draws on information about the offender and their circumstances in order to reach judgements about their likely future behaviour. Such assessments may consider how likely an offender is to reoffend, be reconvicted or cause harm to others. The choice of which future behaviour is the subject of the prediction will depend on the purpose of the assessment.
The efficacy of risk assessment, then, rests on the ability to accurately predict future behaviour. Prediction, however, is not an exact science. If we recall our insurance illustration for a moment; a life insurance company will use information relating to our age, current health status, lifestyle, family medical history and, maybe at some point in the future, our DNA to determine the probability of future serious illness or death. The premium is then calculated on this perceived risk. It is perhaps obvious, however, that such information will not allow for the prediction of more idiosyncratic threats to health, such as being hit by a bus or struck by lightning. In much the same way, an assessment of offender risk of reconviction or harm is open to the same inaccuracies.
Table 18.1 presents the four potential outcomes in relation to prediction. When an event is predicted to occur, it then can either occur (Outcome 1) or not (Outcome 2). Likewise when an event is predicted to not occur, it then can either occur (Outcome 3) or not (Outcome 4). The consequences of incorrect prediction (Outcomes 2 and 3) can be rather unpalatable. For example, in the case of offender recidivism prediction, an offender deemed by the risk assessment as unlikely to reoffend could, on the basis of that prediction, be given a non-custodial sentence or released from incarceration. If the prediction is incorrect and the offender recidivates (i.e. Outcome 3) the previous decisions, predicated on the incorrect risk assessment, will inevitably impact on society creating real and unwelcome consequences. At the individual level, the subsequent offending behaviour creates additional victims and the associated distress of victimisation whilst society is also affected by the monetary costs of the administration of further investigation and justice.
Table 18.1 Possible outcomes in risk assessment
Prediction
Reoffend Not-reoffend
Outcome 1 Outcome 3
Reoffend Correct prediction (true positive) Incorrect prediction (false negative)
Actual event Outcome 2 Outcome 4
Not-reoffend Incorrect prediction (false positive) Correct prediction (true negative)
A false positive outcome is also undesirable: an offender’s application for parole may be rejected on the basis of a risk assessment that predicts future offending behaviour. If the risk assessment was incorrect, and the offender would not have reoffended on release (i.e. Outcome 2 in Table 18.1), it could be argued that the faulty assessment had contributed to an infringement of the detained individual’s human rights and the squandering of public money spent on continued imprisonment. In all risk assessments, then, the overarching aim is to maximise the proportion of correct predictions whilst minimising the false positives and false negatives.
PHOTO 18.1 Should a prisoner be released? Keeping him/her in prison unnecessarily could be seen as an infringement of human rights, but underestimating risk and releasing someone back into the community could lead to further
victims.
Source: © New Punisher/Shutterstock
18.3 METHODS OF ASSESSING RISK
While there are a number of ways of determining an individual’s level of risk, these tend to fall broadly into two categories: clinical and actuarial risk assessments. Clinical risk assessment, as the name implies, draws on the clinical skills of the assessor to evaluate the offender’s risk level. Hence, clinical assessment involves the collection of background information and observation of the offender by professionals. The clinician utilises the information gleaned from such observations, alongside their experience and training, to guide the formulation of the risk prediction. Originating from the medical field (Monahan, 1981), this practice can be construed as a diagnostic assessment of risk (Howells & Hollin, 1989).
Actuarial prediction is really quite different. Actuarial methods typically employ statistical algorithms, or equations, to generate risk scores from specific items of information. These scores can be construed as estimated likelihoods of the occurrence of the event of interest (i.e. on a scale of 0 to 10, for example), or can be categorised into risk bands such as low, medium and high risk. The algorithms utilised within actuarial prediction are generally determined through large cohort longitudinal research projects assessing the factors associated with the outcome in question. The resultant assessments are hence derived through the comparison of the individual’s risk predictors against the amassed knowledge of a similar sample’s outcomes.
Both clinical and actuarial assessments draw on a range of data points. These data can be static (or historic factors), such as the offender’s number of previous convictions, their age, and their gender. Such factors are termed static as they are not amenable to change through intervention or otherwise. Furthermore, both types of assessment can also utilise dynamic, or theoretically changeable, factors such as psychological (cognitive distortions, attitudes to crime), social (employment status, the nature of peer relationships) or behavioural (levels of impulsivity, displayed aggression) measures. Hence the choice of information used by the two kinds of assessment can overlap partially or indeed fully; it is the employment of this information that differs between the clinical and actuarial prediction fields.
A longstanding debate within the field of psychology has centred on the comparative merits of these two methods of risk assessment. Advocates of the clinical method argue that the generalisation of information from a group to an individual is in itself problematic (Dingwall, 1989). Others have claimed that actuarial methods oversimplify the complexity of factors involved in an offender’s decision to reoffend or cause harm (Grubin, 1997; Grubin & Wingate, 1996; Litwack, 2001) and fail to attend to idiosyncrasies that might be highly predictive. In contrast, Grove and Meehl (1996) have argued that actuarial risk assessment “involves a formal, algorithmic, objective procedure (e.g. equation) to reach the decision” (p. 293) whilst clinical prediction “relies on an informal, ‘in the head,’ impressionistic, subjective conclusion, reached (somehow) by a human clinical judge” (p. 294). This view was predicated on the findings of their meta-analysis, which compared the predictive accuracy of both methods and concluded in favour of actuarial assessments.
This view is now generally accepted (Howe, 1994; Milner & Campbell, 1995; Quinsey, Harris, Rice, & Lalumiere, 1993). Moreover some authors have gone so far as to propose the “complete replacement” of clinical assessments with actuarial measures (Quinsey, Harris, Rice, & Cormier, 1998). Others have proposed that such statements are too hasty. Douglas, Cox and Webster (1999) argue that:
…the actuarial method, while useful, is not a panacea, and that it has limitations in the risk assessment field. The function of actuarial prediction methods is simply that—prediction. Risk assessment as conceived here is broader than prediction. Prediction is a necessary first step. Yet risk management and prevention are equally necessary steps. Once a person is defined as high risk, it is in everyone’s best interest to suggest means by which such risk can be attenuated. (p. 155).
The authors endorse a combination of clinical and actuarial methods into an “empirically validated, structured clinical assessment” (p. 157) of risk and need. Such tools draw on the expertise of professionals’ decisions and allow them to highlight idiosyncrasies that may impact on risk, whilst also providing an empirically derived and validated risk score or banding. Indeed the combination of clinical assessment with actuarial methods, it has been argued, can play an important role in identifying the factors that contribute to the outcome of interest (Limandri & Sheridan, 1995). As such, the development of structured clinical judgment, sometimes referred to as third generation, tools allows for the objective assessment of risk whilst also supplying an empirically derived and individualised profile of criminogenic needs. In the next section of this chapter, we will examine some of the more commonly used assessment tools.
18.4 RISK AND NEED INSTRUMENTS FOR OFFENDERS
Within England and Wales, the prison population has been growing over recent years and, despite a slowdown in growth since 2008, at the time of writing (May 2017) the population is hovering around the 85,000 figure, with almost 4,000 of these being female prisoners. In the last year (to December 2016), an additional 267,000 convicted offenders have been serving their sentences in the community under the supervision of either the Probation Service, which carries out risk assessments of all offenders and manages the most high risk offenders, or one of the 21 community rehabilitation companies, who manage low and medium risk offenders.
The crimes that these individuals have committed cover the full spectrum from benefit fraud and shoplifting to the more heinous crimes of sexual assault, manslaughter and murder, and hence, at the point of contact with prison and probation services, they will have varying levels of risk and need. Risk and need instruments can therefore be used to estimate the risks that offenders pose to themselves, others and society, and to decide how the risks and needs are best managed by the correctional services. Indeed, a key principle of offender management within Her Majesty’s Prison and Probation Service is that “resources should follow risk” (National Offender Management Service, 2006a, p. 22). This rule of offender management echoes the earlier message of Andrews and colleagues in their discussion of the risk principle, which, if you remember, stated that higher risk individuals should receive longer and more intensive interventions than those of lower risk. The operational consequence of these principles is that the levels of containment, supervision and intervention imposed on an offender are determined by their risk of reoffending, risk of harm and level of need, all of which can be derived though the utilisation of risk and need assessment tools.
In practice then, risk and need assessment tools are often administered prior to sentencing to inform decisions on the requirements placed on the offender as part of their sentence. For example, the court may decide that the potential risk of harm to others an offender poses, coupled with their high risk of reconviction, is too high to allow a community penalty and hence sentence them to a period of imprisonment, or it may feel that a medium risk of reoffending can be mediated by attendance at a rehabilitative offending behaviour programme within the community. Furthermore, community agencies may use the outcomes of risk and need assessments to inform the detail of offenders’ sentence plans, parole decisions and risk management plans within custody or on release into the community. Case Study 18.1 presents an example of how risk assessment information can be used to inform decisions within prison settings. It also highlights the need not only to ensure useful and accurate assessments but also to ensure that the outcomes of such assessments are taken seriously and used to inform decisions about the management of offenders.
CASE STUDY 18.1 THE DEATH OF ZAHID MUBAREK
Consider the potentially volatile setting of a prison. Risk assessment can be used to minimise the risk of harm posed by a prisoner to other prisoners or staff. For example, when an individual’s risk assessment indicates that they pose a risk to staff, their management plan can
be adapted to include provision to manage this risk – they might only be allowed out of their cell when two or more prison officers are there. Likewise such assessments can identify whether an offender poses a risk to other prisoners based, for example, on offence history or ethnicity. Such information can be used by prison staff to inform cell allocation, for example. The high profile and tragic case of Zahid Mubarek, murdered whilst sleeping in his own cell by his cellmate, Robert Stewart, is an example of the consequences of ignoring the outcome of, or, as in this case, failing to conduct, such assessments. The report of the inquiry into the death of Zahid Mubarek (House of Commons, 2006), states that Robert Stewart was known to have strong racist views, a history of violent behaviour, and possible mental health problems. Since this tragic incident, the Prison Service has introduced and evaluated the “cell-sharing risk assessment form”. This is now completed as part of the reception process into the prison in an attempt to reduce the possibility of such events in the future. At the time of the inquiry, however, the training provided for the completion of these assessments was lacking and the report questioned “whether this initial assessment results in an accurate predictor of risk” (House of Commons, 2006, p. 500). It seems then that despite steps forward within this field of risk assessment, there is still some way to go.
The following three examples describe assessment tools used commonly within correctional services. These tools are general in nature and hence aim to predict, or derive treatment targets in relation to, recidivism or reconviction per se as opposed to any particular type or classification of reoffending behaviour.
Forensic Psychology Page 82
