Assignment: Risk Assessment and Prevention

Assessment and Treatment

Learning Objectives At the end of t his cha pter, you should be able to: • expla in the importance of coord inating risk assessment and risk management

activities for reducing the ri sk of offending;

• describe the three basic steps in the risk assessment p rocess;

• discuss d ifferent risk assessment methods, including their strengths and weaknesses;

• explain the risk-need-responsivity (RNR) model of offender treatment;

• explain the good lives model (GLM) as an alternative approach to offender treatment;

• identify and describe elements of effective offender treatment programs; and

• discuss st rategies for ensuring that the design and delivery of treatment pro­ grams reflect best practices.

EJ The Psychology of Criminal and Violent Behaviour

Karla Homolka

St Catharines, Ontario-Karla Homolka and her husband, Paul

Bernardo-dubbed the "Ken and

Barbie killers" by the media-were

responsible for the sexual homicides

of three young women in Ontario

during the early 1990s. The couple's

first victim was Homolka's young­

er sister, Tammy. As a gift to Ber­

nardo, Homolka drugged Tammy

unconscious so that he could rape

her (Williams, 1999). Tammy died

during the sexual assault, when she

aspirated her own vomit. In two sep­

arate incidents committed nearly a

year apart, Homolka and Bernardo

abducted teenagers Leslie Mahaffy

and Kristen French from the streets

of St Catharines. They took each girl

home, where they sexually assault- she was an active participant. Photo:

ed and tortured her for several days The Canadian Press/ Frank Gunn

Photo 13.1 Although Karla Homolka told investigators that she was an un­

willing accomplice to the rape and

murder of young women, video tapes

of the crimes-found after .her plea

bargain was arranged-revealed that

before killing, dismembering, and

disposing of her. When the couple was arrested in 1993,

Homolka struck a plea bargain with prosecutors, agree­

ing to testify against her husband in exchange for being

allowed to plead guilty to manslaughter and thus avoid­

ing a possible first-degree murder conviction . She be­

gan serving her 12-year sentence in July 1993 (Shephard,

1999). Meanwhile, Bernardo was sentenced to life in

prison with no chance of parole for 25 years. He was also

declared a dangerous offender, which mean s he can be

imprisoned indefinitely (Canadian Press, 2015).

After six years in prison, Homolka unsuccessfully

applied for temporary releases under the escort of a

correctional officia l, claiming that they would help her

"develop a positive support network along with reduc­

ing [her] social isolation" (Shephard, 1999, p. 1). At the

same time, her parole officer reported that Homolka

needed to "focus on assuming responsibility and the

role as an assailant that she played in committing these

offences, before any real social reintegration can be

considered" (Shephard, 1999, p . 1).

Normally, federal offenders are re­

leased on conditions after serving

two-thirds of their sentence, but the

Parole Board of Canada (PBC) con­

tinued to have Homolka detained

b ecause they considered her likely

to commit another offence ca using

death or serious harm (Mofina &

Audbry, 2001).

Despite the concern surround­

ing Homolka, finding suitable

treatment to reduce her risk of re­

offending proved complicated. At

the time, Canadian correctional fa­ cilities contained so few female sex

offenders-only 16 in the entire

federal system-that no treatment

program existed for them (Roman,

2001). Correctional officials first

tried to deal with the situation by

transferring Homol ka to an institu­

tion for offenders with mental disor­

ders, where they planned to adapt

a program intended for male sex offenders ("Homolka

moved," 2001). They soon concluded that none of the

in stitution's treatm ent programs would be beneficial

so moved her again. Eventually, she was treated at an­

other correctional facility, with a program designed for

juvenile sex offenders. She had completed most of this

program when her parole applicat ion was rejected for

a fourth time (" Twelve years," 2005). In fact, Homolka

neve r received ea rly release and remained in prison

until her sentence expired in July 2005.

The restrict ions on Homolka d id not end with her

release. In an unusual legal move, government officials

convinced a court to impose conditions on her free­

dom beyond the end of her sentence, on the basis that

she remained a danger to the public (Hamilton, 2005).

Howeve r, th e legal o rder was soon quashe d and all

the restrictions removed (Haines, 2005). Homolka is

now re married and, at last report, is living in Montrea l

with her husband and children (Wilton, 2016).

13 Assessment and Treatment

Introduction

After people are convicted of a criminal offence, they typically come under the authority of the correctional system, which is responsible for administering sentences and prepar­ ing offenders for eventual reintegration into society. Carrying out these responsibilities is complicated because, as we have seen throughout this book, offenders differ from one

another in many aspects, including upbringing, social skills, education, personality, and cognitions (Van Voorhis & Salisbury, 2014). This diversity means an approach that is ap­ propriate for one offender may be inappropriate or even counterproductive for someone else. Put succinctly, one size does not fit all when it comes to managing offenders. Correc­ tions officials plainly recognized this point when they struggled to find suitable treatment for Homolka. The process of assessing an offender's personal characteristics and circum­ stances and matching them to appropriate treatment and other management strategies is known as offender classification (Andrews, Bonta, & Hoge, 1990; Clements, 1996; Van Voorhis & Salisbury, 2014). This chapter focuses on how offenders are assessed, differenti­ ated, and managed in ways that promote public safety.

Risk of Offending A risk represents a possible adverse or undesirable outcome (Hart, 2001; Kraemer et al. , 1997). While we might not think in these terms, considerations of risk govern many of our daily choices. Decisions to invest in a mutual fund or undergo a medical procedure are both influenced by our perceptions about the likelihood of an unwanted outcome, wheth­ er it be declining financial markets, poor health, or something else. Risk is no less a con­ sideration in the criminal justice system, although the adverse outcome of concern in this context is offending behaviour. It is evident that Homolka's risk of offending explicitly or implicitly entered into the decisions to find appropriate treatment for her and to reject her applications for temporary escorted absences and early release. In fact, Homolka's risk was considered so great that efforts were made to have restrictions imposed on her even after her sentence was over.

Effectively reducing the risk of offending requires two complementary activities: risk assessment and risk management (Hart, 1998b). Risk assessment centres on developing an accurate picture of the risk posed by an offender. This activity has sometimes been referred to as risk prediction, but as Hart (2001) has pointed out, a good assessment goes beyond merely forecasting the likelihood that someone will commit a crime. It also in­ volves determining what it is about an individual offender and his or her circumstances that may contribute to future offending behaviour. A parallel can be drawn to the field of medicine. Knowing that a patient is in critical condition and likely to die is not especially helpful unless the physician understands what is putting the patient's life in peril so that a suitable course of treatment can be initiated. Similarly, once it is understood why an offender poses a risk, it is possible to undertake risk management, which involves imple­ menting strategies and delivering services designed to mitigate the risks identified during the assessment process (Hart, 1998b; Snowden, 1997). In short, the best way to prevent acts of criminal and violent behaviour is to assess and manage offender risk in a coordinated fashion (Guy, Douglas, & Hart, 2015).

offender classification The

p ro cess of assessing an of­ fender's personal character­ istics and circumstances and matching them to appropriate management strategies and services aimed at lowering his or her risk of reoffending.

risk of offending The likeli­

hood that a particular person will engage in an act of criminal or violent behaviour,

usually with in a specified time frame.

risk assessment The process of evaluating the likelihood of someone's risk of offending and the reasons that the

person poses a risk.

risk management The pro­ cess of implementing strate­ gies and services intended to mitigate identified risks.

The Psychology of Criminal and Violent Behaviour

risk factor A characteristic associated with an increased likelihood of future criminal behaviour.

protective factor A char­ acteristic asso ci ated with a decreased li kelihood of future criminal behaviour.

Risk

Mitigates identified risks

Informs risk management strategies

Management Implement strategies

to reduce the

Figure 13.1 The interconnected nature of risk assessment and risk management

Risk Assessments of Offenders

Steps in the Risk Assessment Process

As we shall see shortly, the risk of offending can be assessed in several ways; however, all methods have three basic steps in common. The first is information gathering. While the precise information collected is influenced somewhat by the method used and the risk of interest, assessments typically canvass the offender's past antisocial behaviour, social ad­ justment, psychological adjustment, and current or anticipated circumstances.

The value of any assessment depends on the accuracy and comprehensiveness of the data it is based on. To safeguard against incomplete or biased information, it is generally recommended that evaluators use multiple sources (Borum, Swartz, & Swanson, 1996). Information may be gathered by interviewing the offender, victims, and collateral informants (e.g. the offender's family and friends) as well as reviewing relevant documents such as the individual's criminal and mental health records and police reports . In some cases, evaluators may administer psychological tests or use other assessment tools.

The second step is to identify and evaluate the presence (or absence) of relevant risk and protective factors. A risk factor is something associated with an increased likelihood of future criminal behaviour (Kraemer et al., 1997). For example, we know from previous chapters that individuals who exhibit poor educational achievement or prior antisocial behaviour have a greater probability of committing crime, so these characteristics are risk factors. On the other hand, a protective factor is something associated with an increased likelihood of prosocial behaviour (Kraemer et al., 1997). Most protective factors represent the opposite condition characterized by a risk factor. Using the examples above, it is evi­

dent that high educational achievement and the absence of previous antisocial behaviour are protective factors. Because these types of factors are generally the reciprocal condition of one another, it is usually unnecessary to refer to both; therefore, we will frame our re­ maining discussion only in terms of risk factors .

13 I Assessment and Treatment

A distinction can be made between static and dynamic risk factors . A static risk factor is permanent or at least extremely resistant to change, whereas a dynamic risk factor is changeable. Returning to our examples once again, prior antisocial behaviour constitutes a static risk factor because it is impossible for an offender to undo these misdeeds. By com­ parison, offenders may choose to upgrade their education, making educational achieve­ ment a dynamic risk factor. This changeable quality has two important implications. Most

notably, it means that the risk currently posed by an offender may differ from past or

future risk. Risk is expected to fluctuate depending on the status of relevant dynamic risk factors, including whether the offender is currently abusing drugs or alcohol, experiencing psychosis, maintaining relationships with antisocial peers, or failing to comply with a pre­ scribed treatment and medication regime (Douglas & Skeem, 2005). In addition, dynamic risk factors make ideal targets for treatment precisely because they can change and reduce an offender's risk.

The third basic step is to arrive at a judgment about risk. Every risk assessment in­ volves some sort of decision-making process that considers, weighs, and synthesizes the various risk and protective factors (Hart, 2001). Decision-making in this context is either the product of professional judgment or actuarial methods (Grove & Meehl, 1996; Hart, 2001). Professional judgment is involved if the evaluator has input into the final deci­ sion; actuarial methods rely exclusively on a fixed and explicit set of predetermined rules (Grove & Meehl, 1996; Hart, 2001). The critical distinction between these processes is the presence or absence of human discretion. Professional judgment allows the evaluator to exercise discretion, whereas actuarial methods eliminate it.

Risk Assessment Methods

While every risk assessment involves the same basic steps, considerable variation exists in how these steps are performed . The literature published over the past two decades de­ lineates at least half a dozen methods for conducting risk assessments (e .g. Hanson, 1998; Hart, 2001; Melton, Petrila, Poythress, & Slobogin, 2007; Otto, 2000). Four major methods are described here: unstructured professional judgment, anamnestic, structured profes­ sional judgment, and actuarial. The main distinction among them is the degree of spec­ ificity and structure surrounding information gathering, risk factor identification and operationalization, and decision-making.

Unstructured Professional Judgment Method As its name suggests, the unstructured professional judgment method lacks structure at all three steps of the assessment process (Hart, 2001). Evaluators gather any information they believe is relevant , using the methods they feel are most suitable and/or convenient in the circumstances. They also decide which factors to emphasize and which ones to ignore and how this information will be integrated into an overall judgment of risk (Bonta, 1996; Melton et al., 2007; Monahan, 2008). In general, this approach is highly dependent on the evaluator's personal knowledge and experience (Hart, 2001; Melton et al., 2007).

The main advantage of this approach is flexibility; it can be readily adjusted to accom­ modate a wide range of contexts and unique, case-specific features (Hart, 2001; Melton et al., 2007). The downside is that the approach varies tremendously from evaluator to eval­ uator, which critics claim leads to low interrater reliability-that is, the level of agreement between different evaluators' assessments ofthe same person or thing (Grove & Meehl, 1996;

static risk factor A perma­ nent, or at least very diffi c ult to change, characteristic associated with an incre ased likelih oo d of future criminal behaviour.

dynamic risk factor A changeable characteristic as sociated with an increased likelihoo d of future criminal behaviour.

unstructured professional judgment method The risk asse ss ment method that gives the e valuat o r complete discretion over all aspects of th e process, including information gathe ring, risk fa ctor identifi c ation and op­ erationalization, information synthesis, and making the judgment about ri sk.

interrater reliability The extent to which different e valu ators' independent assessments of the same person or thing match.

anamnestic method The risk assessment metho d that involves inquiring into an offender's history to id entify the sequence of perso nal and situational fa ctors that the evaluator believes led to past offending.

offence cycle The particular sequence of p ersonal and situational fa cto rs that led an offender to commit past criminal acts.

The Psychology of Criminal and Violent Behaviour

Hart, 2001; Melton et al., 2007). In other words, if one evaluator concludes that someone presents a high risk of offending, we expect other evaluators examining this person to ar­ rive at the same conclusion. Accuracy appears to be another problem. Research generally shows that unstructured professional judgment is not particularly accurate relative to other types of decision-making processes (Grove & Meehl, 1996, Hart, 2001; Melton et al., 2007). The absence of structure also means a lack of transparency (Hart, 2001). Only the evaluator knows what considerations were used and how they were put together, which makes it very difficult for anyone else to scrutinize or test the conclusions (Bonta, 1996; Hart, 2001).

Anamnestic Method The anamnestic method involves reviewing an individual's offence history to identify the circumstances surrounding his or her criminal behaviour (Hart, 2001; Melton et al., 2007). It is founded on the premise that, if the same factors and sequence of events that led to previous criminal behaviour should reappear, the offender is likely to repeat the same be­ haviour (Miller & Morris, 1988). Imagine that Offender X has an extensive criminal record for motor vehicle theft. A review of these offences reveals that some of his associates have substance abuse problems and he uses drugs whenever he is in their company. Once he starts using drugs , it quickly becomes habit-forming, and he requires a constant source of money to purchase drugs. Unfortunately, his work performance slips so dramatically when he is using that he is usually fired from whatever job he has at the time. In the absence of any legitimate source of income to support his habit, he resorts to the only means he knows for acquiring money fast: stealing cars. The behavioural sequence, or offence cycle, in this case is straightforward (see Figure 13.2) and shows that associating with drug users, per­ sonal drug use, and employment problems are all important risk factors for this offender.

As this example illustrates, the anamnestic approach imposes some structure on the risk assessment process by requiring evaluators to gather information about the offence history (Hart, 2001). This method does not, however, prescribe the risk factors that must be considered or specify rules for judging risk. The approach's primary strength lies in its ability to guide risk management (Melton et al., 2007). Once the offence cycle is known, strategies can be developed to interrupt the behavioural sequence and avert future offenc­ es (Hart, 2001). For example, Offender X's case presents a number of opportunities to in­ tervene and break his offence cycle by developing relationships with prosocial individuals, avoiding drug-using associates, attending drug counselling, securing short-term financial

Offender X's Crime Cycle

Associates Financial problems ....lliri.. Pe rsonal Steals cars

with peers from buying drugs and .....,.-+ drug use for money

using d rugs loss of employment

Substan ce abuse A vo id peers invo lved Financ ial assistance; counse llin g in the drug culture ; employment ski ll s

strengthen prosocial tra ining relationship s

Opportunities to break the crime cycle

Figure 13.2 Identifying the offence cycle through an anamnestic risk assessment

13 I Assessment and Treatment EJ aid, and participating in employment skills training. Despite its practical appeal, the an­

amnestic approach is largely unproven in terms of reliability and accuracy for predicting and preventing reoffending (Hart, 2001; Melton et al., 2007). Moreover, anamnestic as ­ sessments ignore the fact that risk factors shift over time and that offenders may behave differently in the future (Hart, 2001).

Structured Professional Judgment Method

A third and more recent trend in the field of risk assessment is the structured professional judgment (SPJ) method. Central to this method are guidelines intended to reflect good risk assessment practices with respect to training, qualifications, and assessment consid­ erations (Hart, 1998b, 2001). SPJ adds structure to the process in two fundamental ways: it specifies the information that should be gathered to carry out a risk assessment, and it identifies and operationalizes key factors that research, theory, and/or practice indicate are important for risk assessment purposes (Douglas & Reeves, 2010).

The HCR-20v3, now in its third version, is a popular example of the SPJ method (Douglas, Hart, Webster, & Belfrage, 2013). This tool was designed for assessing violent, rather than general, offending among adults with mental health issues. Its guidelines direct evaluators to rate the relative presence or absence of 20 risk factors according to definitions contained in the HCR-20v3 manual and then use this information to form an opinion about the offender's risk (Douglas et al., 2013). Thus, the HCR-20v3 structures the initial steps of the assessment process, including the information to be gathered and the risk factors to be examined, but leaves evaluators free to determine the relative impor­ tance of each risk factor and how the factors should be integrated into a judgment of risk.

Researching Criminal and Violent Behaviour

structured professional judgment (SPJ) method The risk assessment method that prescribes the minimum infor­ mation that must be gathered and identifies and opera­ tionalizes the risk factors that must be conside red but gives the evaluator discretion over how to synthesize this information and make the judgment about risk .

The HCR-20v3

Consistent with the SPJ approach, the HCR-20v3 guide­

lines outline some general principles of good risk

assessment practices and articulate the evaluator's

expected training and qualifications (Douglas et al.,

2013). Its 20 risk factors were identified through re­

views of the literature, using a rational selection pro­

cess based on reason or logic (Douglas & Reeves,

2010). That is, a factor might be selected because it

is empirically related to violence, implicated by con­

temporary theories of violence, widely recognized by

professionals in the field as relevant, or some combi­

nation thereof (Douglas & Reeves, 2010; Hart, 2001).

Ultimately, the goal is to capture a broad collection of

key risk factors (Douglas & Reeves, 2010) .

The risk factors of the HCR-20v3 are organized into

three scales: historical, clinical, and risk management

(see Table 13.1). The historical scale contains 10 items

related to aspects of the person's past. Five items ad­

dressing the person's current or recent functioning

make up the clinical scale. The remaining five items,

in the risk management scale, concern the person's

anticipated future circumstances (Douglas & Reeves, 2010). All the risk factors are defined, and evaluators

assess and rate each one on a 3-point scale, where

0 indicates the factor is absent, 1 indicates it is par­

tially or possibly present, and 2 indicates the item is

definitely present (Douglas et al., 2013). Evaluators are

instructed to determine whether an offender poses a

low, moderate, or high risk after taking into account

the number of presenting risk factors, the relevance

of these risk factors, and other considerations, such

as the nature and intensity of management needed

in the case. Thus, decisions about risk remain the re­

sponsibility of the evaluator.

The Psychology of Criminal and Violent Behaviour

Table 13.1 HCR-20v3 Scales and Items

Historical (Past)

H 1 . Violence

H2. Other antisocial behaviour

H3. Relationships

H4. Employment

HS. Substance use

H6. Major mental disorder

H7. Personality disorder

HS. Traumatic experien ces

H9. Violent attitudes

H10. Treatment or supervision response

Clinical {Present)

C1. Insight

C2. Violent ideation or intent

C3. Symptoms of major mental disorder

C4 . Instability

cs. Treatment or supervision response

Risk Management (Future)

R1. Professional services and

plans

R2. Living situation

R3. Personal support

R4. Treatment or supervision response

RS . Stress or coping

actuarial method The risk assessment method that prescribes all aspects of the process, including informa­ tion gathering, risk factor identification and operation­ alization, and the rules for combining this information and making the judgment about risk.

construction sample The original group of par­ ticipants studied t o develop a new psychological test or assessment tool.

Source: Douglas, K. S., Hart, S. D., Webster, C. D., & Belfrage, H. (2013). HCR-20 (Version 3): A ssess ing risk of violence-User guide . Burnaby, BC: Mental Health, Law, and Policy Institute, Simon Fraser Unive rsity, p. 115.

One benefit of the SPJ method is greater interrater reliability because every evaluator examines the same risk factors using a common set of definitions (Hart, 2001; Melton et al., 2007). The validity of the judgments is also expected to be better because evaluators consider well-established risk factors. This arrangement minimizes the problem of eval­ uators who might overlook important risk factors or attach great weight to extraneous details (Douglas & Reeves, 2010). Finally, transparency is increased because the infor­ mation and risk factors going into the judgment are obvious (Hart, 2001). The approach's main drawbacks are the time and expense required to develop the guidelines (Hart, 2001). Suitable guidelines may not even exist for some cases, giving evaluators little choice but to rely on other approaches. Moreover, the evaluator's discretion over the judgment of risk is objectionable to those who prefer the purely mechanical process of actuarial methods (e.g. Quinsey, Harris, Rice, & Cormier, 2006).

Actuarial Method The actuarial method is the most structured risk assessment method. It specifies the information to be gathered, identifies and operationalizes the risk factors, and uses a predetermined formula or set of rules to make judgments about risk (Hart, 2001; Otto, 2000). Actuarial risk assessment instruments are usually derived through research with a construction sample, which is the original group of research participants used to de­ velop, or "construct," the instrument. Typically, the construction sample is composed of offenders studied for a specified time to see who engages in further criminal or violent acts. Statistical analyses identify the best combination of risk factors for predicting which offenders in the construction sample reoffended and which ones did not (Melton et al., 2007). These findings are then used to formulate the actuarial instrument.

13 Assessment and Treatment

This general approach was employed to develop the Violence Risk Appraisal Guide (VRAG), a 12-item actuarial instrument for assessing the risk of violence among offenders with mental disorders (Harris, Rice & Quinsey, 1993). All the items in the VRAG relate to fac­ tors that predicted subsequent violent reoffending in the construction sample. Like other ac­ tuarial instruments, the VRAG prescribes all steps of the assessment process. The information that must be collected is specified by the 12 items; each item or risk factor is operationalized so that evaluators know how to organize and score it; and judgments about risk are made using a fixed set of formal rules. Evaluators sum the 12 individual item scores to obtain a total VRAG score, which is converted into a precise estimate of the offender's probability of reoffending violently. Most important, the VRAG makes no provision for evaluators to modify this risk estimate, which is solely the product of a mathematical exercise (see the "Researching" box).

The high degree of structure associated with actuarial methods provides instruments such as the VRAG with many of the same advantages observed for SPJ, including good

Researching Criminal and Violent Behaviour

The Violence Risk Appraisal Guide

The Violence Risk Appraisal Guide (VRAG) was devel­

oped by a group of researchers working at a maximum

security psychiatric institution located at Penetan­

guishene, Ontario (Harris et al. 1993; Ouinsey et al.,

2006). The construction sample involved more than

600 mentally disordered, male offenders assessed or

treated at the institution. The researchers reviewed

each offender's file and recorded information concern­

ing approximately 50 potential "predictor" variables

covering aspects of the offender's sociodemograph­

ic characteristics (e.g. age, marital status), childhood

and adult adjustment (e.g. elementary school malad­

justment, criminal history, alcohol use), index offence

(e.g. relationship to the victim, victim injury), and psy­

chological assessments (e.g. 10, psychopathy).

After the participants were released back into soci­

ety, the researchers followed up, searching the national

database of arrests and convictions to see who had re­

offended . By this time, the participants had been free

for an average of nearly seven years. A participant was

deemed to have reoffended if he had been charged

with a new violent offence or returned to the institu­

tion in connection with an act that, in the research­

ers' opinion, could have resulted in such a charge.

Violence was defined broadly to encompass offences

ranging from simple assault to homicide and included

sexual assaults, armed robbery, forcible confinement,

threatening, and pointing a firearm (Harris et al., 1993;

Ouinsey et al., 2006). Based on these parameters,

31 per cent of the construction sample recidivated vi­

olently during the follow-up period (Harris et al., 1993).

The researchers performed a series of statistical

analyses to determine which study variables proved

to be the best predictors of violent reoffending in the

construction sample. The 12 most predictive variables

were selected for inclusion in the VRAG (see Figure

13.3). Harris and colleagues performed a further se­

ries of computations to develop a scoring scheme for

each item that was designed to optimize the instru­

ment's overall predictive ability. Scores for each item

were added to yield a total VRAG score, which could

range from -27 to +35. The researchers divided the

total possible score into nine brackets spanning sev­

en points each and calculated the recidivism rate for

each bracket based on the construction sample. For

example, 44 per cent of the offenders in the +7 to

+13 bracket recidivated within 7 years. The percent­

age of offenders who recidivated grew to 58 per cent

when the follow-up period was extended to 10 years

(see Table 13.2). It is assumed that other offenders

who were not in the construction sample but receive

similar scores will reoffend at the same rate.

The Psychology of Criminal and Violent Behaviour

1. Lived with both biological parents to age 16 (except for death of parent) 0 Yes • -2 0 No-+3

Score:

2. Elementary school maladjustment (up to and including grade 8) LI No problems = -1 D Slight or moderate discipline or attendance problems= +2

LI Severe (i.e. frequent or serious) behavior or attendance problems (e.g. truancy or

disruptive behaviour that persisted over several years or resulted in expulsion) = +5

Score:

3. History of alcohol problems LI 0 points ­ -1 D 1 or 2 points = 0

D 3 points""' +1 D 4 or 5 points= +2

Score:

Allot one point for each of the following : Alcohol abuse in biological parent Alcohol involved in a prior offence Teenage alcohol problem Alcohol involved in index offence Adult alcohol problem

4. Marital status (at time of index offence) D Ever married (or lived common law in the same home for at least 6 months) = -2 D Never married ­ +1

Score:

5. Criminal history score for convictions and charges for non-violent offences prior to the index offence (Comnier-lang Criminal History Score) D Score of 0 =-2 D Score of 3 or above == +3 LI Scoreof1 or2- 0

Score:

Cormier-Lang Criminal History Scores Robbery (bank, store) - 7 Break & enter (ind. intent to commit offence) - 1 Robbery {purse-snatching) - 3 Fraud (extortion, embezzlement) - 5 Arson and fire·setting (church, house, barn) - 5 Fraud (forged cheque, impersonation) - 1 Arson and fir'e-setting {garbage can) - 1 Possession of a prohibited/restricted weapon - 1 Threatenfn9 with a weapon ; pointing firearm - 3 Living on avails prostitution or procuring - 1 Threatening (uttering threats) ,.. 2 Trafficking in narcotics - 1 Theft over (ind. possession stolen property) - 5 Dangerous driving, impaired driving= 1 Mischief to public or private property ewer - 5 Obstruct a peace officer find . resisting arrest) "" 1 Break & enter and commit indictable offence - 2 Causing a disturbance - 1 Theft under (ind. possession stolen property) - 1 Wearing a disguise with intent - 1 Mischief to public or private property under ..; 1 Indecent exposure• 2

6. Failure on prior conditional release (includes parole violation or revocation, Score: breach of or failure to comply with recognizance or probation, bail violation, and

any new charges, including the index offence, while on a conditional release}

7. Age at index offence (at most recent birthday) Score: a ,,39 =-5 LI 27 = o LI 34-38 - -2 ·LI <26 - +2 LI 28-33 - -1

8. Victim injury (index offence only; most serious injury is scored) Score: 0 Death .. -2 l] Treated and released"" +1 LJ Hospitalized - 0 LJ None of slight (ind. no victim) = +2

9. Any female victim (for index offence) Score: i:J Yes• - 1 LJ No(ind.novictim)=+1

10. Meets DSM-Ill criteria for any personality disorder Score: 0 Yes - +3 LI No - -2

11. Meets DSM-Ill criteria for schizophrenia Score: a Yes - -3 LI No - +1

12. Hare Psychopathy Checklist-Revised score (PCL-R; Hare, 1991) Score: LI <4= -5 LI 15-24=0 LI s-9 - -3 LI 25-34 - +4 a 10-14 - -1 LI >26 = +2

TOTAL:

Figure 13.3 Violence Risk Appraisal Guide (VRAG) Source: Copyright © 2006 by the American Psychological Associat ion. Adapted wit h permission . V.L. Ouinsey, et al. (2006). Violent offenders: Appraising and managing risk (2nd ed .). Wash in gton, DC: American Psycholog ical Association. The use of APA info rmation does not imply endorsement by APA.

interrater reliability, validity, and transparency. In addition, the use of actuarial decision­ making means that human discretion and the frailties associated with it are eliminated from the judgment part of the process. In this respect, these methods address past concerns

13 Assessment and Treatment

Ta bl e 13.2 Recidivism rates associated with different VRAG scores

Rate of Recidivism

VRAG Total Score Over 7 Years Over 10 Years :;; -22 0.00 0.08

-21 to -15 0.08 0.10 -14 to -8 0.12 0.24 -7 to -1 0.17 0.31

0 to +6 0.35 0.48

+7 to +13 0.44 0.58

+14 to +20 0.55 0.64

+21 to +27 0.76 0.82

~ +28 1.00 1.00

Source: Copyright© 2006 by the American Psychological Association. Adapted with permission. V.L. Ouinsey, et al. (2006). Violent offenders: Appraising and managing risk (2nd ed.). Washington, DC: American Psychological Association, p. 286. The use of APA information does not imply endorsement by APA.

Table 13.3 Characteristics of the four major risk assessment methods

Risk Prescribed Prescribed and Prescribed rules assessment information operationalized for decisions method gathering risk factors about risk

Less structure Unstructured professional x x x judgment

Anamnestic ,/ x x Structured professional ,/ ,/ x judgment

More structure Actuarial ,/ ,/ ,/

Source: Copyright © 2008 by the American Psychological Association. Adapted with permission. J. Monahan "Structured risk assessment of violence." Violence assessment and management. R. I. Simon & K. Tardiff (Eds.), 2008. Washington, DC: American Psychiatric Publishing. The use of APA information does not imply endorsement by APA.

over the evaluator's ability to effectively process and integrate large amounts of information into clinical opinions and avoid many of the biases that exert subtle yet influential effects on human decision-making (Grove & Meehl, 1996; Quinsey et al., 2006). Actuarial methods are also grounded in scientific research and rely on risk factors with an established relation­ ship to criminal behaviour (or some other outcome of interest) in the construction sample.

On the other hand, creating an instrument designed to achieve maximum predictive ac­ curacy in the construction sample has a cost. The more finely tuned an instrument is for one sample, the more poorly it tends to perform in different samples (Hart, 2001). Another concern is the inadvertent omission of risk factors. Useful risk factors may be excluded from actuarial instruments because the researchers did not think of them, it was impractical to include them in the original pool of test variables, or they appear unrelated to criminal behaviour due to the idiosyncratic characteristics of the construction sample (Douglas & Reeves, 2010). Two other limitations of actuarial methods mirror those mentioned for SPJ, namely, the time required for development and the absence of suitable instruments for some cases. Table 13.3 provides a summary of the four major methods of risk assessment and their associated characteristics.

adjusted-actuarial approach The risk assessment method

that involves generating an initial judgment of risk

using the actuarial method and then subjecting it to the evaluator's professional discretion.

The Psychology of Criminal and Violent Behaviour

The Clinical versus Actuarial Debate

For many years, there has been a great deal of controversy over whether professional judg­

ment (also called clinical judgment) or actuarial methods produce better assessment de­

cisions (e.g. Holt, 1970; Meehl, 1954). This dispute-sometimes referred to as the clinical versus actuarial debate-originated in other fields but has spilled over into the area of risk assessment during the past two decades (Dvoskin & Heilbrun, 2001; Webster, Hucker, & Bloom, 2002). Proponents of actuarial methods usually base their stance on the supe­ rior accuracy of this method relative to human judgment. They often point to the large body of research, in a wide array of fields ranging from health to employment, that shows actuarial methods typically exceed the accuracy of unstructured professional judgment (e.g. 1Egisd6ttir et al., 2006; Grove, Zald, Lebow, Snitz, & Nelson, 2000).

Meta-analytic reviews of research specific to risk assessment echo these general re­ sults and confirm that actuarial methods are also more accurate than unstructured pro­ fessional judgment in this field (Bonta, Law, & Hanson, 1998; Hanson & Morton-Bourgon, 2009; Mossman, 1994). Individual studies conducting head-to-head comparisons using the same information, sample, and outcome measures also show that actuarial methods come out on top (e.g. Bengston & Langstrom, 2007). In view of the accumulated evidence, it is widely recognized that the disadvantages of unstructured professional judgment usu­ ally outweigh the advantages and, as a consequence, this method should be avoided when­ ever possible (Heilbrun, Yasuhara, & Shah, 2010).

The debate has now shifted to one between structured professional judgment and ac­ tuarial methods. But why would anyone continue to use decision-making processes that incorporate human judgment in the face of evidence supporting the greater relative accu­ racy of actuarial methods? Research reveals that, when risk assessments are appropriately structured, they generally fare as well as actuarial methods in terms of accuracy (see Guy et al., 2015). Bear in mind that actuarial instruments are often fine-tuned to the con­ struction sample for maximum predictive accuracy and, as a result, there can be a notice­ able decline in performance when they are applied to other samples (e.g. Blair, Marcus, & Boccaccini, 2008). By comparison, the SPJ method is expected to perform robustly across different samples because it selects risk factors on the basis of the literature as a whole and is not optimized for a construction sample (Douglas & Reeves, 2010; Hart, 2001).

Supporters see one further advantage of the SPJ method. With control over the judg­ ment of risk, evaluators are able to consider the relevance of each risk factor based on the case's unique characteristics. Douglas and colleagues (2013) explain that evaluators might reasonably decide that an offender poses a high risk on the basis of only a few or even a single risk factor if that factor is highly relevant and concerning (e.g. the offender expresses a clear intent to carry out a terrorist attack). Of course, exercising this type of discretion is exactly what critics claim reduces the reliability and accuracy of non-actuarial approaches (Quinsey et al., 2006; Rice, Harris, & Hilton, 2010). At this juncture, there is insufficient ev­ idence to conclude that either the SPJ or actuarial method clearly outperforms the other in terms of predictive accuracy. The only way to settle this debate, if it needs settling, is to con­ duct more studies employing head-to-head comparisons, as is often done in the medical field when there are questions about best practices involving different medical procedures.

Perhaps in an attempt to find an intermediary position, another approach that com­ bines professional judgment and actuarial methods, known as the adjusted-actuarial approach, has emerged (Hanson, 1998; Melton et al., 2007). It is often construed as separate and distinct from the others discussed so far. The challenge that evaluators face is how to

13 Assessment and Treatment

blend these two methods to form a single overall judgment of risk. There are two possibil­ ities. One approach is to use the actuarial instrument to form a risk estimate that the eval­

uator may increase or decrease to account for important case-specific factors that would otherwise be overlooked (Hanson, 1998). The other is to generate an actuarial estimate that

the evaluator considers along with all the other gathered information (Litwack, 2001). The

difference between the approaches is really a matter of emphasis (Melton et al., 2007). With

the former, the actuarial estimate is the foremost consideration and it is modified only to the extent necessary. The latter places more emphasis on the evaluator's professional judg­ ment, and the actuarial estimate merely represents one piece of information among many.

The prospect of combining professional judgment and actuarial approaches has evoked a number of criticisms. On some level, a combined approach is an oxymoron . Once professional discretion enters into the decision-making process, it is no longer ac­ tuarial. As others have noted, a decision is either rule-bound or without rules; it cannot be both (Douglas & Reeves, 2010). Of course, actuarial purists warn against the dangers of allowing professional judgment, arguing that it will inevitably result in diminished ac­ curacy (Rice et al., 2010; Quinsey et al., 2006). Dvoskin and Heilbrun (2001) suggest that evaluators should not tamper with an actuarial estimate of risk once it is derived. They recommend contextualizing the information for consumers (e.g. courts, parole boards) by explaining the estimate's strengths and limitations. In other words, evaluators should freely identify and explain other considerations that might affect the actuarial estimate, but they should stop short of adjusting it based on these considerations. Unfortunately, there is virtually no published empirical research on the practice of modifying actuarial assessments. Descriptions of the few unpublished studies completed to date suggest that actuarial assessments of risk overridden by evaluators tend to be less accurate than the original estimate (see Hanson & Morton-Bourgon, 2009; Heilbrun et al., 2010).

Putting It All Together: Good Risk Assessment Practices While the clinical versus actuarial debate is often cast in black and white terms, the actual division may not be so distinct. Actuarial methods frequently incorporate one or more risk factors that require clinical or professional judgment (Litwack, Zapf, Groscup, & Hart, 2006; Monahan, 2008). The VRAG is a useful case in point because it includes items based on the Hare Psychopathy Checklist-Revised (PCL-R) and DSM-III diagnoses of personali­ ty disorder and of schizophrenia. All these items require evaluators to make professional judgments about whether these conditions are present. Thus, the VRAG eliminates profes­ sional judgment only from the final step by dictating how the items are to be combined and interpreted. Furthermore, while the actuarial and SPJ approaches differ with respect to the involvement of professional judgment in decisions about risk, they are very similar in many other ways. Both use empirical evidence to inform the selection of risk factors, albeit through somewhat different methods (Heilbrun et al., 2010). Each approach prescribes what information needs to be gathered, what risk factors must be considered, and how those risk

factors are operationalized. Litwack and colleagues (2006) have gone so far as to suggest that professional judgment and actuarial assessments share so much in common that the two are "blurring to the point where there is often no meaningful distinction" (p. 504).

The evident overlap in these approaches indicates some degree of agreement over how risk assessments should be conducted. Rather than emphasizing areas of discord, it may

The Psychology of Criminal and Violent Behaviour

Table 13.4 Essential e lements for good risk assessment practices

1. Risk assessment practices need to be evidence-based (Douglas, Cox, & Webster, 1999; Melton et al., 2007; Webster et al., 2002) . That is, the procedure should reflect current research and consider factors with established relationships to the hazard of interest.

2. When available, structured methods are preferable to unstructured methods (e.g. Bonta, 2002; Melton et al., 2007; Monahan, 2008).

3. Risk assessments should be comprehensive and balanced so that all major known risk and protective factors are reviewed and considered (Bonta, 2002; Hart, 2001; Lavoie, Guy, & Douglas, 2009; Rogers, 2000; Webster et al., 2002).

4. Risk assessment practices should be sensitive to changes over time and therefore should take static and dynamic risk factors into account (Douglas & Skeem, 2005; Hart, 2001; Reid, 2003; Rogers, 2000).

5. Risk reassessments should be carried out at appropriate intervals so that any changes in dynamic risk factors are noted and evaluated.

6. Risk assessment practices should provide helpful information that can be used to guide risk management strategies (Hart, 2001; Lavoie et al., 2009; Snowden, 1997; Webster et al ., 2002).

be more constructive to identify those elements that are generally recognized as essential for good risk assessment practices. No definitive list currently exists, but the six elements identified in Table 13.4 are widely supported in the literature.

Treatment of Offenders Treatment is one of the major strategies used to manage offenders and their risk of com­ mitting a future offence. The remainder of this chapter discusses the main assessment considerations surrounding offender treatment and the program features that enhance treatment efficacy.

The Decline and Revival of Offender Treatment

The idea that offenders should be treated to "correct" their behaviour (hence the term corrections) is not new. Offender rehabilitation was the dominant force behind correction­ al thinking and practices in North America throughout most of the twentieth century. Although several factors contributed to its waning influence during the latter part of the century, Robert Martinson's article on the effectiveness of prison treatment proved to be a pivotal event. Martinson (1974) began the piece by asking, "What works?" (p. 22) and then discussed a report that he and colleagues had completed on behalf of the New York State government, which reviewed, in painstaking detail, 231 studies on prison rehabilita­ tion carried out between 1945 and 1967 (Lipton, Martinson, & Wilks, 1975). Martinson's (1974) conclusion was that "with few and isolated exceptions, the rehabilitative efforts that have been reported so far have had no appreciable effect on recidivism" (p. 25). Although Martinson outlined several caveats to this statement, it seems the message that most peo­ ple took away from it was simply "Nothing works."

Others disagreed with Martinson's rather bleak assessment of the literature. In a rebuke published the following year, Palmer (1975) reported that nearly half the stud­ ies Martinson reviewed actually showed some sign of positive treatment effects. A few years later, Gendreau and Ross (1979) responded with an article pointedly entitled "Bib­ liotherapy for Cynics," which examined 95 methodologically stronger studies conducted

13 Assessment and Treatment

after Martinson's report and noted the presence of beneficial treatment effects in many of these investigations. When the studies were quantified, it was found that treatment had a positive impact in 86 per cent of the investigations (see Andrews, Zinger, Hoge, Bonta, Gendreau, & Cullen, 1990). These and other assessments of the available research revealed that treatment worked at least some of the time or at least under the right conditions. Five years after the publication of his article , Martinson (1979) publicly conceded that his orig­ inal conclusion was inaccurate. By this time, however, the nothing works doctrine was so firmly embedded into conventional thinking that not even his retraction could dislodge it (McGuire & Priestley, 1995).

Although a steady flow of studies continued to report encouraging results, two in­

vestigations published in the early 1990s were instrumental in restoring confidence in the potential value of correctional treatment (Andrews & Bonta, 2010). One was Mark Lipsey's (1992) quantitative review of results from studies examining the effectiveness of delinquency treatments using what was then the relatively new technique of meta­ analysis . This review was not the first meta-analysis of the offender treatment literature, but it stood apart both in terms of its quality and sheer magnitude. Lipsey analyzed over 400 studies, nearly three times more than any other meta-analysis of offender treatment at the time, which made its results extremely convincing. In contrast to the view that nothing works, he reported that treating juvenile offenders was associated with an average 10 per cent drop in recidivism compared to untreated juvenile offenders. Lipsey (1992) acknowledged that the size of the reduction was very modest, but it was nonetheless statis­ tically significant and non-trivial. Just as noteworthy, Lipsey (1992) found that treatment success varied with the type, integrity, and intensity of the intervention.

The second notable investigation was conducted by Andrews, Zinger, and colleagues (1990) , who believed that offender assessment was the key to effective correctional treat­ ment. They hypothesized that treatments appropriately matched to the offender would outperform mismatched or unmatched treatments. To explore this possibility, they re­ viewed 80 studies that examined the effects of treatment on recidivism. The type of treat­ ment was categorized in one of four ways: criminal sanctions (the type or level of criminal disposition varied) , inappropriate treatment (no or inappropriate matching of treatment type and offender), appropriate treatment (appropriate matching of treatment type and offender), or unspecified treatment. The study revealed that, compared to inappropriate treatments or unspecified treatments, appropriate treatments significantly lowered recid­ ivism. Together, the results of these investigations went a long way to restoring faith in correctional treatment, as well as sparking interest in the use of assessments for matching offenders to treatments.

The Rise of the RNR Model of Offender Treatment

Over the past two decades, our understanding of what makes a treatment appropriate, or inappropriate, has grown tremendously thanks in large part to the pioneering work of Donald Andrews, James Bonta, and their colleagues (Andrews, Zinger, et al., 1990; Andrews, Bonta, & Hoge, 1990). These researchers combined theory and empirical eval­ uation to identify three core principles of effective correctional treatment they termed risk, needs, and responsivity (RNR). These principles describe the who, what, and how of correctional treatment and together form the RNR model that is now a mainstay of contemporary correctional systems in many parts of the world (Andrews & Bonta, 2010).

risk principle A guideline

of the RNR model stipulat­ ing that, for treatment to

be effective, the intensity level must be matched to the offender's risk of reoffending such that low-risk offenders receive little or no treatment

and high -risk offenders re­

c eive the most treatment.

need principle A guideline

of the RNR model stipulating that, for treatment to be ef­ fective, it should be directed

at dynamic risk factors.

responsivity principle A

guideline of the RNR model stipulating that, for treatment to be effective, it must be delivered in a manner that matches the offender's abili ­ ties and learning styles.

general responsivity An

aspect of the RNR model's

responsivity principle stipu­ lating that, for treatment to be effective, it must be struc­ tured and del ive red within the context of a positi ve

therapeutic envi ronment.

specific responsivity An

aspect of the RNR model's responsivity principle stipu­ lating that, for treatment to

be effective, the therapeuti c approach must be tailored to the offender's particular characteristics and abilities.

The Psychology of Criminal and Violent Behaviour

The Risk Principle The risk principle identifies who should receive correctional treatment. It specifies that treatment should be matched to the offender's level of risk such that the most intense treatment efforts should be dedicated to the highest risk offenders. Conversely, very little

or no treatment should be given to the lowest risk offenders. Research confirms the sound­ ness of this principle, generally showing that allocating treatment in this manner has the greatest overall impact on lowering recidivism (Andrews & Dowden, 2006; Landenberger & Lipsey, 2005; Lipsey, 2009; Lowenkamp, Latessa, & Holsinger, 2006). The reason it is better to target high-risk offenders may be no different than the reason a C student will probably benefit more from tutoring than an A student: there is simply more room for improvement (Lipsey, 1995).

Other factors may also lie behind this principle. Andrews and Bonta (2006) note that subjecting low-risk offenders to high levels of treatment not only appears unhelpful, but it may also be counterproductive and increase recidivism. Although the reason for this finding is not entirely certain, two hypotheses have been put forward . The most frequently cited possibility is that placing low-risk offenders in intensive treatment programs typical­ ly means surrounding them with higher risk offenders who model and reinforce antisocial behaviour (Andrews & Bonta, 2010). The other explanation is that putting these offenders into treatment programs disrupts and weakens many of the protective factors that contrib ­ ute to their low-risk level, such as employment and relationships (Lowenkamp et al., 2006).

The Need Principle The need principle concerns what should be targeted for treatment, namely, an offender's dynamic, or changeable, risk factors . The logic behind this principle is straightforward . Treating and changing dynamic risk factors will lower the risk of reoffending, whereas treatment directed at static risk factors or at factors unrelated to criminal behaviour will not. Accordingly, Andrews and Bonta (2006) recommend targeting factors such as antiso­ cial thoughts and attitudes, procriminal peers, substance abuse, and lack of employment skills, all of which have a known relationship to criminal behaviour. The RNR model does not dismiss the possibility of treating an offender's other needs, but these are seen as much less important because they do not directly impact the risk of offending.

The Responsivity Principle The responsivity principle addresses how the treatment should be carried out. Generally speaking, treatment must be delivered in a manner that matches the offender's abilities and learning style. There are two aspects to responsivity. General responsivity refers to the broad approach to treatment, which should involve a structured program that is de­ livered in a positive therapeutic environment. Program structuring is discussed later in the chapter. To create a positive therapeutic environment, client-therapist relationhips must be established that are characterized by trust, respect, collaboration, support, empa­ thy, and understanding (Andrews & Bonta, 2006; Marshall, Marshall, Serran, & O'Brien, 2013). Specific responsivity involves adjusting the broad treatment approach to suit the unique characteristics and abilities of the individual offender, including his or her maturi­ ty level, cognitive abilities, learning styles, and cultural background. For example, journal writing exercises might be useful for high-functioning, well-educated offenders but are

13 I Assessment and Treatment

entirely unsuitable for illiterate offenders. Similarly, treatment sessions delivered to an of­ fender with ADHD should be relatively short and involve active participation to maximize the likelihood of success.

One important aspect of responsivity is treatment motivation. As the cases of Darnell

Pratt and Brenden Sarginson illustrate, offenders differ vastly in their motivation to seek out and participate in treatment programs. Pratt showed little interest in receiving treat­ ment for his substance abuse after correctly anticipating that he was going to be released from prison even without it. In contrast, Sarginson turned himself into authorities when he breached a term of his probation and asked the judge for a longer sentence to facilitate his treatment (see the case study). Although it is not well studied, treatment motivation is thought to have a major impact on the extent to which treatment gains are realized.

Darnell Pratt and release date was coming up and he wou ld be allowed to leave Brenden Sarginson prison whether he finished the

Maple Ridge, British Col um bia­ program or not. In March 2005, 16-year-old Dar­ The documents further in­ nell Pratt committed a gas-and­ dicate that Pratt was occasional ­ dash at a service station. When ly "uncooperative" with his case Grant De Patie, the station atten­ management team and was dant, tried to prevent the crime, suspected of using drugs and Pratt struck him with the vehicle alcoho.1 while in prison (Chan, and dragged him several kilo­ 2010). Two days following his re­ metres, ca using his death. Pratt lease, he failed to return to hi s was convicted of manslaughter halfway house, putting him in and ordered to serve seven years breach of his release conditions in prison. He admitted drinking (Campbell, 2010). He was ar­ daily during the months prior to rested a short time later. At his the offence and was given the revocatio n hearing, it came to opportunity to participate in a light that he had been found prison substance abuse progra m drinking in a pub twice in two

Photo 13.2 Since being released from prison, Darnell Pratt has reportedly committed several parole violations and pied guilty to car theft.

(Chan, 2010). Paro le documents describe Pratt as exhibiting a negative attitude to the program and making minimal effort to complete treatment-re lated assignments. Correctional authori­ ties suspended him from th e program after only six weeks for failing to attend or arriving late for over one-third of the sessions. Pratt indicated that he was not interested in participating because his conditional

days, contra ry to his relea se conditions that required him to abstain from all drugs and alcohol use (Campbell, 2010).

Regina, Saskatchewan- One day after finishing a two-year stint in prison for sexua ll y assaulting a six­ year old girl, Brenden Sarginson went to a park, in

continued

The Psychology of Criminal and Violent Behaviour

primary human goods Perso na l characteristics, states of mind, a nd experi­ ences that are intrinsica lly valued by people and p romote thei r psychological we ll -b e ing.

The RNR model illustrates why good offender assessment is so vital. To apply RNR principles, it is necessary to differentiate lower and higher risk offenders from one another, identify individual dynamic risk factors, and determine personal attributes that are likely to influence treatment response. All these tasks involve assessing offenders, but assess­ ments alone are insufficient. The assessment information must be used to match offenders to suitable correctional programs. Think back to the Homolka case. The indecision over her treatment probably related to the difficulties correction officials encountered in find­ ing a program that fit her p ar ticular risk, needs, and responsivity.

The Good Lives Model

Despite its success, the RNR model is not without critics. While acknowledging the approach 's success and merits, Ward and colleagues criticize its single-minded focus on risk factors, which in many ways represent problems or deficits in the offender (Ward, 2002; Willis & Ward, 2013). They argue that this emphasis on deficits is unappealing to offenders and does not inspire or motivate them to become engaged in treatment (Willis & Ward, 2013). The good lives model (GLM) was developed as an alternative framework that provides a more pos­ itive, strengths-based approach to offender treatment (Ward, 2002; Ward & Maruna, 2007).

The GLM is grounded in the belief that offenders, like everyone else, desire personal meaning and fulfillment in their lives, which is achieved through primary human goods. Primary human goods represent personal characteristics, states of mind, and experiences that are intrinsically valued and promote psychological well-being (Ward & Maruna, 2007). So far, Ward and colleagues have identified 10 classes of primary goods, which are listed

violation of probation conditions that prohibited him

from, among other things, being anywhere child ren

congregate (Pruden, 2010). Sarginson recogn ized that

he had taken the first step in his offence cycle, which

he described as "disturbing" (Pruden, 2010, p. A1). Of

his own volition, he went to police and admitted to vi­

olating the terms of his probation. At court, the prose­

cutor requested that Sarginson receive 12-18 months

for breaching his probation. Sarginson argued for a

two-year sentence so that he would serve the time in

a federal pen itentia ry rather than a provincial institu­

tion . He stated, "I need help, and I can't get it on the

street ... The only place for me to get it in Saskatche­

wan is the federal pen itentiary" (Pruden, 2010, p. A1).

The judge noted that the term of impri sonment was

too harsh for a breach of probation but acquiesced to

Sarginson's request. On hearing the news, Sarginson

is reported to have smiled and said, "Now I can go

up to the pen and I can start the treatment I need " (Pruden, 2010, p. A1).

l

Photo 13.3 At court for violating the terms of his pro­

bation, Brenden Sarginson, requested the judge im­

pose a lon ger sentence to facilitate his treatment.

13 Assessment and Treatment

Table 13 .5 Primary human goods identified by t he good lives model

1. Life

2. Knowledge

3. Excellence in play and work

4. Autonomy

5 . Inner peace

6. Relatedness

7. Community

8. Spirituality

9. Happiness

10. Creativity

Physical needs required for healthy living and functioning .

Desire to understand ourselves, others, and our environment.

Desire to engage in recreational activities for pleasure and develop compe­ tence in work and non-wo rk acti vities.

Desire to freely choose one's goals and the means of achieving those goals without interference from others.

Desire to ach ieve and ma intain a stable emotional state free of turmoil and stress .

Desire for w arm and caring atta chments to others, including friends, family, and romantic partners.

Desire to belong to social groups of people who share the same beliefs and interests.

Desire for a sense of meaning and purpose in life .

Desire for the feeling of gratification and g eneral contentment with one's life.

Desire to come up with new ideas or ways of d o ing things or to produce artistic works.

Source: Day, A ., Casey, 5., Ward, T., Howells, K., & Ves s, J. (2010). Transiti o ns to better lives: Offender readiness an d re habilitatio n. Por tland, OR : Willan , pp . 51-52 .

in Table 13.5 (Day, Casey, Ward, Howells, & Vess, 2010). Although everyone pursues some measure of all 10, the significance and priority attached to each class varies from individu­ al to individual (Willis & Ward, 2013). Most important, GLM assumes that people commit crime to attain , directly or indirectly, one or more of these primary human goods. For example, someone who lacks romantic companionship, which reflects the primary human good of relatedness, might stalk someone as part of an inappropriate effort to establish a relationship with a desired partner.

The GLM suggests that the key to treatment is to develop an offender's capacity to attain primary human goods in a prosocial manner (Ward & Marshall, 2004). This goal can be ac­ complished by strengthening the internal capabilities of offenders, such as furthering their knowledge, skills, and beliefs, as well as addressing their external conditions, such as provid­ ing them with assistance and opportunities (Ward & Gagnon, 2006). Treatment on stalking might work on enhancing the offender's interpersonal skills and confidence in social situa­ tions (i.e. internal capabilities) and on identifying social activities and other opportunities where meaningful and consensual relationships might develop (i.e. external conditions). It is expected that the risk of future criminal behaviour will diminish as a natural consequence of improving the capabilities of offenders and enriching their lives (Ward & Marshall, 2004).

As our discussion makes clear, RNR and GLM differ in terms of their principal ob­ jectives and the means used to attain them. For the RNR model, the primary goal is to decrease reoffending. Treatment, then, targets an offender's risk needs. The main objective of GLM is to improve the offender's well-being and life satisfaction-reduced offending, though expected, is a corollary of this goal. According to proponents, one major advan­ tage of GLM is that it focuses on achieving the very things that matter most to offenders so that they are more likely to be motivated to engage in therapy (Ward, Mann, & Gagnon, 2007). Detractors worry that GLM may undo many of the gains made in correctional treat­ ment over the last several decades by re-directing scarce resources away from an offender's risk-related needs to his or her general psychological well-being, a shift that has not been shown to reduce recidivism (e .g. Bonta & Andrews, 2003).

treatment program A highly structured treatment incorporating specific treat­ ment goals and treatment content that is defined and sequenced.

multimodal program An intervention that addresses more than one need.

treatment dosage The amount of treatment, usually measured in hours, deliv­ ered to participants during the course of a treatment program.

attrition In terms of a treat­ ment program, the number of intended participants who do not complete the program.

The Psychology of Criminal and Violent Behaviour

Elements of Effective Treatment

Structured Treatment Since the three original principles of RNR were first articulated, research has continued to identify treatment features associated with improved outcomes for offenders. A key finding from this work is that structured treatments are typically more effective than unstructured ones (Lipsey, 1992). Treatment structure is a function of its goals and content. Treatment goals are usually derived from criminological theory and must be specific and clearly stated (McGuire, 2004). Treatment content, in turn, must be designed to achieve the identified ther­ apeutic goals (McGuire, 2004). Thus, all the activities that make up the content need to be carefully defined and explained and their sequence set out. An intervention that incorporates these structural elements is usually referred to as a treatment program. The most effective treatment programs tend to work on developing identifiable skills and/or behaviours (Lipsey, 1992). Cognitive, behavioural, or cognitive-behavioural methods have all generally proven to be better at bringing about desired changes in offenders than alternatives such as punish­ ment, psychodynamic approaches, or unspecified methods (Lipsey, 1992; Li:isel, 1995).

Multimodal Program Multimodal programs, which have two or more treatment components, are generally more effective than interventions that rely on a single therapeutic activity (e.g. Lipsey, 1992). Typically, each component has a specific object that may or may not be addressed by the other components using different methods or techniques. The obvious advantage of multimodal approaches is that numerous treatment needs can be targeted within the framework of a single program. One meta-analytic review found that multimodal treat­ ments featuring cognitive-behavioural therapy (discussed on next page) coupled with other components, such as mental health counselling or employment upgrading, were associated with significantly larger decreases in recidivism than interventions restricted to CBT alone (Landenberger & Lipsey, 2005). Likewise, Andrews and Bonta (2006) note that targeting a greater number of dynamic risk factors (relative to the number of static risk factors) produces stronger treatment effects.

Treatment Dosage Research shows that the treatment dosage, or the amount of treatment delivered in a program, is another factor that impacts outcome. Reviews of the literature indicate that higher treatment dosages are associated with lower rates of recidivism (Lipsey, 1992; Landenberger & Lipsey, 2005). The dose required to achieve maximum treatment effi­ cacy is not known with any degree of certainty, but it is recommended that offenders re­ ceive around a hundred hours of programming (Lipsey, 1992; Smith, Gendreau, & Swartz, 2009). The treatment dosage an offender receives is a function of several factors, including the number of sessions per week, the duration of each session, and the number of weeks in the program. In their quantitative review of CBT programs, Landenberger and Lipsey (2005) observed that the frequency and duration of weekly contact have a stronger bear­ ing on outcome than does the length of the program. Thus, high frequency sessions that concentrate treatment into a small number of weeks appear to be more effective than de­ livering the same amount of treatment over a longer period of time.

Treatment dosage is also affected by attrition, which refers to the number of intended program participants who do not complete treatment. Attrition may come about in a variety

13 Assessment and Treatment

of ways. For instance, offenders may not commence programs that the courts or correctional officials directed them to take; they may leave the program before reaching the end; or they may be expelled for violating program rules or expectations. The case of Darnell Pratt, dis­ cussed earlier, is an example of attrition. Attrition is problematic because it often reflects an offender's lack of motivation to change. Furthermore, non-completers receive less than the full treatment dosage and therefore any therapeutic gains they derive are likely attenuated.

Treatment Setting

The treatment setting appears to be another important factor. Meta-analytic reviews of the research report differential treatment effects for programs implemented in the com­ munity versus institutions (e.g. Andrews & Bonta 2006; Lipsey, 1999; Lipsey & Wilson, 1998). McGuire (2002a) estimates that community-based programs may be nearly twice as effective than institutionally based programs, but only if the programs are structured. Unstructured programs tend to perform poorly regardless of where they are situated. Setting-related treatment differences could be the result of one or more of the negative impacts of institution-based programs previously mentioned, such as removing offenders from their social supports and/or immersing them in the procriminal environment of custodial institutions.

Cognitive-Behavioural Therapy Cognitive-behavioural therapy (CBT) is widely hailed as one of the most promising forms of treatment for offenders. CBT is not unique to the correctional field; it is a prominent therapeutic approach used by clinical psychologists to treat a wide variety of problems, including mood, eating, and anxiety disorders. The generic term cognitive-behavioural de­ scribes a group of related clinical interventions that meld behaviourism and cognitive psy­ chology (McGuire, 2000). The therapy's central tenet is that our thoughts and behaviours are inextricably linked and therefore, to change the latter, we must alter our cognitive con­ tent and processes-that is, we must change what and how we think (Dobson & Dozois, 2001). CBT assumes that people who exhibit maladaptive behaviour also have underlying cognitive problems. Treatment typically concentrates on getting individuals to recognize and change errors in their thinking and to develop skills that will help them avoid prob­ lem behaviours in the future (Dobson & Dozois, 2001).

Offender treatment programs based on CBT operate on the same principles. As we have seen, especially in Chapter 8, offenders' thought content and processes tend to differ from non-offenders' in several important respects. Offenders are more likely to display poor moral reasoning, distorted cognitions, and antisocial attitudes and to misinterpret the actions and intentions of others. Presumably, these maladaptive thinking patterns, together with poor interpersonal and coping skills, lead offenders to frequently make poor decisions and engage in criminal acts. When delivered appropriately, CBT is consistent with the principle of general responsivity and is one of the treatment approaches recom­ mended by leading authorities in this area (Andrews & Bonta, 2006; Smith et al., 2009). Cognitive-behavioural methods may involve cognitive restructuring, cognitive skills de­ velopment, or both approaches at the same time (Van Voorhis & Salisbury, 2014).

Cognitive restructuring is mostly concerned with distortions and other flaws in the way offenders think (MacKenzie, 2006). It works on getting offenders to monitor and change the attitudes, beliefs, and assumptions they hold that support and promote their criminal behaviour (Van Voorhis & Lester, 2009). For instance, cognitive restructuring

cognitive-behavioural

therapy (CBT) An approach

to psychological treatment that focuses on individuals recognizing and changing errors in their thinking as well as developing cognitive skills

as a way to avo id problem

behaviours.

cognitive restructuring

A form of cognitive­ behavioural therapy that fo­ cuses on changing behaviour by altering a person's thought content. It typically involves challenging the person's dis­ torted beliefs and working to replace undesirable thoughts with more appropriate ones.

The Psychology of Criminal and Violent Behaviour

cognitive skills development A form of cognitive­

behavioural therapy that fo­ cuses on changing behaviour by altering how a person pro­ cesses information and de­ cides how to act. It typically involves the use of modelling and role-playing exercises to

develop the person's ability

to analyze, consider, and re­ spond effectively to different situations.

might seek to identify and challenge the distorted beliefs of a child molester who views his or her actions as harmless and educational for the victim. Specific cognitive restructuring techniques include challenging offenders with evidence that contradicts their distorted be­ liefs, getting them to actively monitor their thoughts in journals, and having them replace undesirable thoughts with more appropriate alternatives. In short, cognitive restructuring aims to alter cognitive content, or what the offender thinks (Van Voorhis & Lester, 2009).

Cognitive skills are concerned with how an offender processes information and makes

decisions to act in a particular way (Van Voorhis & Lester, 2009). The aim of cognitive

skills development is to help the offender adapt his or her behaviour appropriately to dif­ ferent social situations (MacKenzie, 2006). Thus, cognitive skills address faulty decision­ making processes as opposed to faulty thought content. Cognitive skills interventions work on improving the offender's ability to effectively analyze a situation, generate possible responses to it, and weigh the consequences of those responses, as a way to deal effectively to a wide variety of interpersonal problems and stressful events (Dobson & Dozois, 2001). Interventions of this sort typically utilize modelling, guided instruction, feedback and re­ inforcement, and role-playing to teach offenders such skills as effective communication, critical reasoning, perspective-taking, goal-setting and planning, negotiation, and conflict resolution. Once they master the individual skills, offenders practise putting them togeth­ er to deal with the problems they encounter in their day-to-day lives (McGuire, 2006).

How Effective Are Offender Treatment Programs? As our discussion shows, the "right" kinds of offender treatment can make a difference. The question that remains is, "How effective are treatment programs at reducing recidi­ vism?" The body of accumulated research has grown to the point where it is feasible to cal­ culate meta-analytic estimates of treatment effectiveness. In fact, so many meta-analyses exist on this issue that investigators have begun "reviewing the reviews" (e.g. McGuire, 2004). In one such review, Lipsey and Cullen (2007) reported that every meta-analysis they looked at found that correctional treatment had a positive impact on reoffending. The minimum average drop in recidivism associated with treatment across the more than three dozen meta-analyses they examined was 10 per cent.

This reduction is modest and may even seem trivial in the absence of some appropri­ ate context. To gain a better appreciation of what this figure means, some researchers have compared the effect sizes of offender treatment programs with those achieved by inter­ ventions in other fields (e.g. Andrews & Bonta, 2010; McGuire, 2002b). Effect sizes provide a common metric for comparing the outcomes of different studies, with larger figures reflecting stronger treatment effects. The effect size associated with offender treatments overall lies in the neighbourhood of 0.10 (e.g. Losel, 1995). As Table 13.6 shows, treating offenders is not out of line with other medical and psychological interventions.

The problem with looking at the average effectiveness across treatments is that it lumps everything together. Of course, treatments are not all equal; their effectiveness depends greatly on the appropriateness of the intervention. Programs that reflect good treatment practices may be expected to attain recidivism reductions beyond the average 10 per cent. Most illuminating in this regard are analyses carried out by Andrews and Bonta (2006), who examined how adherence (or nonadherence) to RNR principles impacted offender treatment outcome. They reviewed 374 studies of treatment and criminal justice sanctions and rated each one according to the number of core RNR principles followed, from 0 (none

13 Assessment and Treatment

Table 13.6 Effectiveness of selected interventions

Type of intervention Target Mean effect size {r)

Medical

Cyclosporine (Rosenthal, 1991)

Bypass surgery (Lynn & Donovan, 1980)

Psychological

Counselling (Erford et al., 2015)

Cognitive-behavioural treatment (Butler et al., 2006)

Correctional

Offender treatments in general (Lose I, 1995)

Offender treatments adhering to RNR (Andrews & Banta, 2010)

Organ rejection 0.19

Angina pain 0.37

Youth anxiety 0.29

Depression 0.44

Recidivism 0.10

Recidivism 0.26

Source: Adapted from McGuire, J. (2002b). Criminal sanctions versus psychologically-based interventions with offenders: A comparative empirical analysis. Psychology, Crime and Law, 8, 183-208. Reprinted by permission of the publisher (Taylor & Francis Ltd, http://www.tandfonline .com).

of the principles) to 3 (all three principles of risk, need, and responsivity). As the level of adherence rose, the likelihood of reoffending dropped lower and lower. Table 13.6 shows that the effect size associated with programs adhering to the principles of RNR is much larger than the overall average. Andrews and Bonta (2006) estimate that following all three principles could translate to a 35 per cent decrease in recidivism. By comparison, treat­ ments that failed to apply any RNR principles were associated with increased recidivism.

Research has also shown that treatment programs based on CBT are very effective. A meta-analysis of European studies found, on average, a 23 per cent decrease in recidi­ vism with CBT compared to a much more modest 12 per cent decline with other treatments (Redondo, Sanchez-Meca, & Garrido, 1999). Other reviews have produced similar results. Lipsey, Chapman, and Landenberger (2001) examined 14 high-quality studies of CBT with offenders and found that treatment was associated with an overall drop in recidivism of 30 per cent. A broader review by Landenberger and Lipsey (2005), encompassing a larg­ er number of studies exhibiting a greater range of methodological quality, arrived at a slightly lower average reduction of 25 per cent. These researchers identified the treatment features associated with the greatest declines in reoffending (e.g. focusing on high-risk offenders, delivering a reasonable treatment dose, and including components that address anger control and interpersonal problem-solving) and then calculated that a CBT program incorporating these "best practices" might reduce recidivism by as much as 50 per cent.

Treatment Integrity

Unfortunately, standardized treatment programs incorporating many best practices fre­ quently fall short of expectations when they are more widely implemented (e.g. Tong & Farrington, 2006; Wilson, Gallagher, & MacKenzie, 2000). For example, Lipsey's (1992)

The Psychology of Criminal and Violent Behaviour

treatment integrity The extent to which a treatment is delivered in the manner intended.

meta-analytic review of the juvenile delinquency treatment literature discussed earlier in this chapter reported that treatment effectiveness was enhanced by smaller study size and greater researcher involvement in the design and implementation of the program. The essence of this finding has been replicated by other researchers comparing so-called demonstration projects involving small-scale, tightly controlled treatments to real-world applications in which treatment is part of an established program broadly available within a correctional institution or system (e.g. Andrews & Banta, 2006; Landenberger & Lipsey, 2005). One estimate indicates that correctional treatments operated on a large scale may be only half as effective as those that are part of small demonstration projects (Lipsey, 1999).

The difference in treatment effectiveness between demonstration projects and rou­ tine practice has two plausible explanations. One possibility is that the prominent role of researchers in demonstration projects sometimes leads to findings that are biased toward treatment success. The other is that demonstration projects tend to have better quality control, or treatment integrity, meaning that they are delivered in the manner intended (Hollin & Palmer, 2006, Lipsey & Cullen, 2007). Lipsey (1992) surmised that, when re­ searchers are heavily involved in small-scale treatment studies, there is probably better quality control and therefore stronger treatment effects. By comparison, treatment pro­ grams implemented on a larger scale, in the real world, may not be delivered as carefully or consistently and, as a result, treatment outcomes suffer. Neither explanation has been definitively confirmed or ruled out, but general opinion seems to favour treatment integ­ rity as the most likely reason.

Elements contributing to treatment integrity are numerous and varied. Andrews and Dowden (2005) identified 10 indicators they used to gauge the integrity of correctional programs (see Table 13.7). In a review of almost 300 treatment studies, they found a signif­ icant, positive relationship between the number of integrity indicators and treatment ef­ fectiveness. That is, better integrity was associated with better treatment results. Although treatment integrity can enhance the effectiveness of "good" treatments, Andrews and Dowden (2005) note that it cannot improve the performance of poor treatments that do not adhere to RNR principles. This makes perfect sense, as delivering an ineffective treat­ ment exactly as it was intended does not change the fact that the treatment was ineffective in the first place. It is akin to building an airplane with serious design flaws that prevent it from flying. No matter how meticulously the design plans are followed or how much care is taken to assemble the plane, it will never get airborne.

Table 13.7 Indicators of program integrity

1. Specific model of desired practice

2. Selection of appropriate staff

3. Staff are trained on program delivery

4. Supervision of staff by someone trained on program delivery

5. Training manuals are available on program delivery

6. Procedures for monitoring service and assessing treatment gains

7. Clients receive adequate treatment dosage

8. New or fresh program

9. Small participant treatment group

10. Program evaluator involved in the design, delivery or supervision of treatment

Source: Adapted from Andrews, D. A., & Dowden, C. (2005). Managing correctional treatment for reduced recidivism: A meta-analytic review of programme integrity. Legal and Criminological Psychology, 10, 173-187.

13 i Assessment and Treatment

In practice, delivering and maintaining high-quality treatments is a tremendous challenge. Staff turnover, training costs, waning staff enthusiasm, scarce resources, and

the time-consuming nature of continually monitoring and supervising treatment all threaten treatment integrity. While all correctional programs face these problems to some degree, larger ongoing programs are the most vulnerable, which may explain the relatively superior performance of small, one-off, demonstration projects. Despite the importance of treatment integrity to treatment success, it has yet to make its way firmly into main­ stream correctional practice. Andrews and Dowden (2005) report that only two of their indicators were mentioned in even half the treatment studies they reviewed. References to most integrity indicators were present in only 20 to 40 per cent of treatment programs. The extent to which these results reflect a true lack of treatment integrity as opposed to the researchers' failure to include it in their published reports is less certain, but the figures suggest that much more could be done to improve efforts around treatment integrity.

Putting It All Together: Effective Programs in Practice Great strides have been made in our understanding of effective correctional treatment programming over the last few decades. We now have a much clearer picture of the ele­ ments that make treatment effective. For instance, we know to target high-risk offenders, focus treatment on dynamic risk factors, accommodate the offender's learning style, and use cognitive-behavioural approaches that are multimodal and of sufficient dosage. We have also learned that treatment integrity is critical to maintaining maximum treatment effectiveness. Attention is now increasingly turning to how this knowledge can be effec­ tively integrated into contemporary correctional practices. Program accreditation and program audit are two new developments in this area.

Program accreditation, which has been undertaken primarily in the UK, is a pro­ cess of assessing and approving treatment programs that reflect best practices (Goggin & Gendreau, 2006). The first step in developing a program accreditation process is to identi­ fy the essential elements that make treatment effective and then integrate them into a set of minimum program standards. Correctional treatment programs undergoing accred­ itation are then evaluated against these standards. Programs that satisfy the criteria are accredited and receive what amounts to an "official" stamp of approval. Presumably, only accredited programs can be used to treat offenders within the correctional system admin­ istering the accreditation program. The main advantage is the assurance that accredited programs incorporate the basic features that the available evidence suggests make treat­ ments effective. Accreditation does not guarantee that a particular program works or is better than another program. These are questions that can be answered only by careful research of the program and its impact on offender recidivism.

The primary focus of program accreditation is confirming that correctional programs are well designed at the outset, but it provides little insight into their day-to-day deliv­ ery. A program audit verifies treatment integrity by assessing whether treatment is being delivered the way it is supposed to be delivered (Hollin & Palmer, 2006). Hollin (1995) recommends that program audits tap three information sources. First, independent ex­ ternal observers should watch the program as it is being administered. Second, program participants should be solicited for their experiences in the program. Finally, program

program accreditation A process for assessing and approving offender treatment programs that incorporate best practices.

program aud it An inspec­ tion of an offender treat­ ment program that assesses whether treatment is being delivered in the intended manner.

The Psychology of Criminal and Violent Behaviour

staff should provide a self-reflective review of how they deliver treatment. Once this in­

formation has been gathered and collated, program supervisors and administrators can assess how closely the treatment being delivered comes to matching the original design. More important, shortcomings can be identified and strategies to correct them can be implemented. Regular program audits (e.g. once per year) help to ensure that problems are rectified and that program delivery does not drift over time.

Summary l. The risk of offending represents the likelihood that a particular person will engage

in an act of criminal or violent behaviour. Effectively reducing this risk requires the complementary activities of risk assessment and risk management. The risk assess­ ment process focuses on evaluating the level of risk a person poses and the reasons that the person poses a risk. Information from the risk assessment is used to inform risk management, which involves implementing strategies to reduce the offender's identified risks as much as possible. Thus, both activities should be carried out in a coordinated fashion.

2. Every risk assessment has three basic steps: information must be gathered about the offender; the information gathered must be used to identify the presence of risk and protective factors; and some sort of decision-making process must be used to form a judgment about the offender's risk based on the relevant risk and protective fac­ tors. One of two types of decision-making processes may be used for this purpose. Decisions based on professional judgment allow the evaluator to exercise discretion, whereas actuarial methods rely on a fixed set of predetermined rules.

3. There are four main methods for conducting risk assessments, which vary in terms of the degree of structure imposed at each step of the process. The unstructured pro­ fessional judgment method has no structure. The anamnestic method requires the evaluator to gather information about the offender's criminal history (i.e. the person­ al and situational circumstances that led to the crime) to identify his or her offence cycle, but it does not prescribe specific factors or how factors should be combined into a judgment about risk. The structured professional judgment method directs the evaluator to gather certain information and evaluate the presence (or absence) of par­ ticular factors that are operationalized; however, the evaluator has discretion over the final judgment about the offender's risk. Actuarial m ethods structure all steps of the risk assessment and, most important, it is the only method that eliminates human discretion from the judgments about risk.

4. Treatment is an important strategy for managing the risk of offending. The popular view of offender treatment during the 1970s and 1980s was that "nothing works," but more current research shows that it can reduce reoffending when carried out appropri­ ately. Our present understanding of what constitutes appropriate treatment has been largely guided by the risk-need-responsivity (RNR) model. The risk principle speci­ fies that treatments should be dedicated to high-risk r ather than low-risk offenders. According to the needs principle, treatment must target dynamic, or changeable, risk factors, not static ones. The responsivity principle recommends using a structured treatment delivered in a positive therapeutic environment (general responsivity), which should be tailored to suit the offender's unique characteristics and abilities (spe­ cific responsivity).

13 Assessment and Treatment

5. In recent years, the good lives model (GLM) has emerged as an alternative to the RNR

model. The GLM is a positive, strength-based approach to treatment grounded in the belief that everyone, including offenders, seeks fulfilling and meaningful lives, which are achieved through primary human goods. These goods represent personal char­ acteristics, states of mind, and experiences that are intrinsically valued and promote psychological well-being. GLM assumes that people commit crime to attain primary human goods; therefore, treatment is geared toward developing an offender's capa­

bilities to realize these things using prosocial behaviour. It is expected that the risk of criminal behaviour will diminish as a natural consequence of improving the capabili­ ties of offenders and enriching their lives. Furthermore, the GLM's emphasis on achiev­ ing what matters to the offender should serve to motivate them to engage in therapy.

6. Research evidence indicates that a variety of other features, such as a clearly defined program structure, a multimodal intervention, sufficient treatment dose, and delivery in community settings, also enhance treatment effectiveness. One particularly prom­ ising therapeutic approach is cognitive-behavioural therapy (CBT), which typically involves changing offender behaviour by altering their thinking through the use ofcog­ nitive restructuring, cognitive skills development, or both at the same time. Cognitive restructuring emphasizes getting offenders to recognize and change the errors in their thinking that led to their past offending. Cognitive skills development focuses on teach­ ing offenders how to analyze, consider, and respond appropriately in social situations.

7. The impact of appropriate offender treatment programs in reducing recidivism is rel­ atively small but non-trivial. Moreover, the size of these treatment effects is equivalent to many other interventions in education and medicine. Current efforts are looking at ways to ensure that existing knowledge on best treatment practices are incorporated into the programs delivered to offenders. One strategy is to develop an accreditation process that sets out required criteria for treatment programs based on the available evidence and then approves treatment programs that meet these minimum standards. Another development is the program audit, which involves inspecting a treatment program to verify that it is being delivered consistently and in the manner intended.

Review Questions 1. Explain the difference between static and dynamic risk factors. Indicate whether each

of the following is static or dynamic: a. exposure to criminal role models growing up b. symptoms of psychosis c. low educational achievement d. a history of past criminal behaviour e. a diagnosis of narcissistic personality disorder f. a social network dominated by antisocial peers

2. Identify the four risk assessment methods and explain one advantage and one disad­ vantage of each.

3. Describe the clinical versus actuarial debate and explain how it has shifted recently. Discuss your position on it.

4. Reread the case study of Darnell Pratt. Discuss whether the RNR or GLM model is a better treatment option, given his obvious reluctance to participate in treatment. Give reasons to support your choice.

The Psychology of Criminal and Violent Behaviour

5. Explain the concept of primary human goods as it relates to the GLM. For each exam­ ple in the following list, indicate the primary human good(s) the offender was trying to obtain: a. a female offender who steals food to feed herself and her children b. a young offender who deals and uses drugs as a way to escape from his conflicted

relationship with his parents c. a male offender who feels he has little control of his life outside his home but gets

satisfaction from dominating and abusing his family

Additional Readings Andrews, D. A., & Bonta, J. (2010). Rehabilitating criminal justice policy and practice.

Psychology, Public Policy, and Law, 16, 39-55. Craig, L. A., Dixon, L., & Gagnon, T. A. (2013). What works in offender rehabilitation: An

evidence-based approach to assessment and treatment. New York, NY: Wiley. Douglas, K. S., Hart, S. D., Groscup, J. L., Litwack, T. R. (2014). Assessing violence risk. In

I. B. Weiner & R. K. Otto (Eds.), The handbook off orensic p sychology (pp. 385-441). Hoboken, NJ: Wiley.

Otto, R. K., & Douglas, K. S. (2010). H andbook ofviolence risk assessment. New York, NY: Routledge.

Van Voorhis, P., & Salisbury, E. J. (2014). Correctional counselling and rehabilitation (8th ed.). Waltham, MA: Anderson Publishing.