Aggression and Violent Behavior 16 (2011) 20–35

Contents lists available at ScienceDirect

Aggression and Violent Behavior

Many sizes fit all: A preliminary framework for conceptualizing the development and
provision of cognitive–behavioral rehabilitation programs for offenders
Devon L.L. Polaschek ⁎
School of Psychology, Victoria University of Wellington, P O Box 600, Wellington 6140, New Zealand

a r t i c l e i n f o a b s t r a c t

Article history: Over the last twenty years, the growing influence of the Risk–Need–Responsivity model (RNR; Andrews &
Received 21 April 2010 Bonta, 2006), and meta-analyses of “what works” can be seen in the number of jurisdictions seeking to
Received in revised form 25 October 2010 implement high quality and consistently delivered rehabilitative interventions for offenders. However, results
Accepted 25 October 2010
have created concern that interventions are “one-size-fits-all,” and that more attention should be given to
Available online 2 November 2010
differential treatment response. And although it has revolutionized high-level policy on offender
Keywords:
management and rehabilitation provision, the RNR model does not provide clear guidance on many
Offender rehabilitation important details of program design and delivery that differentiate one treatment from another. In some areas
Program theory of offender rehabilitation, the conceptual resources to guide such decisions appear to be absent or
Differential treatment underdeveloped. This paper surveys cognitive–behavioral group-based interventions for offenders, and finds
What works considerable diversity in their design and delivery. Several relevant dimensions are used to organize this
diversity into a conceptual framework of three levels of program, based primarily around levels of offender
risk and program intensity. Advantages of such a framework are that it will stimulate theory development and
empirical investigation of alternate delivery models, and in so doing, support ongoing progress in
rehabilitation, despite a political environment that is more and more caught up in punitive containment.
© 2010 Elsevier Ltd. All rights reserved.

Contents

1. Setting the scene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

2. Rehabilitation theory and program development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3. Historical trends in cognitive–behavioral program development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
4. Proposed dimensions for differentiating programs in a framework. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.1. Client characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.1.1. Criminal risk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.1.2. Readiness and responsivity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2. Program characteristics. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2.1. Intensity or dosage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.2.2. Treatment targets and program components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.3. Program delivery methods, change processes and context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.4. Facilitator-related program characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.5. Treatment integrity approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.2.6. Offender assessment, program individualization and treatment change monitoring . . . . . . . . . . . . . . . . . . . . . . 25
5. Proposed program framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.1. Basic-level rehabilitation programs (low-to-medium risk, low intensity) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.2. Examples of basic-level programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.3. Mid-level multi-factorial treatment programs (medium to high risk and intensity). . . . . . . . . . . . . . . . . . . . . . . . . . . 27
5.4. Examples of mid-level multi-factorial treatment programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5.5. High-level comprehensive forensic therapy programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
5.5.1. The therapy environment as a therapeutic tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
5.5.2. Comprehensive assessment of change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

⁎ Tel.: +64 4 463 5768; fax: +64 4 463 5402.

E-mail address: devon.polaschek@vuw.ac.nz.

1359-1789/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.avb.2010.10.002
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35 21

5.6. Examples of high-level comprehensive forensic therapy programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

6. Conclusions and questions for the future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

1. Setting the scene
Political “get tough on crime” agendas in nations such as the
United States, Canada, the United Kingdom, and New Zealand are
causing prison populations to rise rapidly, as an increasingly harsh
and punitive lens is focused on criminals, whether or not crime is
actually increasing (Prisoners of politics: Canada's criminal-justice
policy, 2009). Some governments recognize the need to respond to
scientific research alongside public anxiety about crime. They
acknowledge—in policy, if not in their press releases—that an
affordable and humane approach to community safety does not result
solely from imprisoning more people for longer, but must include
effective offender rehabilitation. However, burgeoning convicted
offender populations relative to the available financial and human
resources for working with offenders may be driving at least two
important changes in rehabilitation practice that challenge current
scientific knowledge about program effectiveness: (1) deployment of
more readily available, much less skilled and less expensive program
deliverers, and (2) redirection of expensive, scarce and highly skilled
therapists to intervention with increasingly high-risk high-need
offenders.
There has been a degree of confidence that—at a broad policy level,
informed by meta-analysis and the Risk, Need, and Responsivity
principles—there exists an empirically sound basis for designing
programs that can work, if implemented well (Wormith et al., 2007).
For some time it has been noted that the question of “what works”
needs to, and has begun to shift to “what works for whom, and when”
(Lösel & Schmucker, 2005; Serin, Kennedy, Mailloux, & Hanby, in
press; Wormith et al., 2007). Most commonly the answer is framed in
terms of the type of intervention method (e.g., cognitive–behavioral
vs. other) or the target offender group (e.g., general, violent, sexual
offenders). As program evaluations have begun to report other factors
such as risk levels of participants, quality of treatment integrity,
background of therapy staff and the like, other types of answers have
begun to emerge. Further progress will likely only be made by looking
at in-program issues, such as how the program addresses client
attrition (Wormith et al., 2007, p. 882), program length, spaced vs.
massed sessions, manualization, ordering of components, approaches
to readying offenders for change, staff skills and delivery style, staff
training and oversight, the range of content, and so on.
Notwithstanding the progress that has been made, one of the most
common criticisms of current empirically driven approaches is that
programs often devolve to “cookie-cutter” or “one-size-fits-all”
(Howells & Day, 2002; Ward, Melser, & Yates, 2007), despite initial
intentions. But is that really the case? Those involved in the
rehabilitation of offenders know that there also are many differences
between these programs. Increasing sophistication is evident in the
offering of different intensities of programs to offenders with distinct
characteristics. In some jurisdictions there are now quite elaborate
suites of programs meeting tight standards of accreditation (e.g., HM
Prison Service, United Kingdom). This paper focuses on integrating
some of these evident differences between programs into a tiered
conceptual framework of treatment program design and provision.
The first aim is to describe systematically trends already in evidence.
However, the paper goes beyond what already exists: aiming to
stimulate new conceptual and empirical investigations of “what to
provide to whom,” in order to effect further improvements in
community safety and offenders' lifestyles. The focus here is on
programs that use primarily a group format, and are cognitive–
behavioral—broadly defined—in the methods they use to effect
change. The paper concludes by identifying some of the implications
of this analysis for future research and policy, particularly with respect
to the development of interventions with the highest risk offenders.
2. Rehabilitation theory and program development
Research that evaluates effectiveness of treatment programs, and
identifies the factors that cause and maintain future criminal risk is
essential to effective program development. However, there are
limitations to bottom–up empiricism (e.g., under-determination of
theory; Ward, Polaschek, & Beech, 2006), making top–down theory
development equally essential. Two broad types of theory are
important in guiding rehabilitation programs: theory of problem
(i.e., etiological theory) and theory of change (Kirsch & Becker, 2006).
Ward and Maruna (2007) defined rehabilitation theory as a hybrid of
these two types, but also including practice guidelines, and guiding
values and aims.
The dominant rehabilitation theory guiding cognitive–behavioral
intervention programs for offenders is known as the “RNR model,”
based on the principles of Risk, Need, and Responsivity (Andrews &
Bonta, 2006). During the 1980s, a growing interest both in
“differential treatment” (Andrews, Bonta, & Hoge, 1990, p. 23)—
another term for “what works for whom”—and in the body of relevant
empirical research led them to propose the Risk, Need, Responsivity
(and Professional Discretion) Principles (Andrews et al., 1990).
Numerous meta-analyses, and individual program evaluations sub-
sequently have confirmed the empirical status of these principles.
These findings have been disseminated so that they provide
invaluable guidance for those developing policy about the form and
content of programs, and the target clientele. Consequently, jurisdic-
tions that were responsive to the importance of scientific research in
developing and delivering programs have been provided with a
relatively accessible source of information for guidance.
Over time, the RNR principles developed into the RNR model (e.g.,
Andrews & Bonta, 2006): a “comprehensive rehabilitation framework
theory” (Ward et al., 2007, p. 218). However, as Ward and Maruna
(2007) note, there is confusion about exactly what the RNR model is,
and the underlying theoretical base is complex and difficult to apply in
practice. Ward and Maruna describe RNR (and its underlying PIC-R) as
“fundamentally a psychometric model” (p. 22), referring in part to its
strong emphasis on empirical correlates both of criminal behavior,
and of effective approaches to rehabilitation.
Perhaps because of its complexities and its psychometric base, the
RNR model and its associated empirical research base are often
translated into practice as a series of lists, summarizing key findings.
Alongside the three RNR principles, other examples include the
authors' “big four” and “central eight,” and lists of principles for
program design and delivery (Goggin & Gendreau, 2006; McGuire,
2002). The importance of lists such as these in improving program
design cannot be overemphasized. The skill and effort required for
policy-makers and program designers to make sense of the growing
research literature is likely to be a major cause of low-quality
programming (Gendreau, Goggin, Cullen, & Paparozzi, 2002).
Nevertheless, programs adhering to RNR-derived guidelines still
vary widely (e.g., programs for sexual vs. violent offenders). Some
variations imply significant underlying theoretical differences; both in
the understood etiology of the targeted problems, and the methods by
which change is thought to be effected. However, beyond describing
22 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
themselves as cognitive–behavioral, and as adhering to the RNR
model, many programs' underlying theoretical bases remain implicit
or opaque (Polaschek & Collie, 2004; Ward et al., 2007).
In addition to drawing on an overarching rehabilitation theory
then, programs have their own specific and local theory: whether it is
implicit or explicit. In need of further development is a series of
intermediate-level theories1 that meaningfully link rehabilitation
theories to the local theoretical basis for a particular program. Theory
at this intermediate level is currently quite well developed for sex
offender intervention; there is a range of resources outlining the
general theoretical approach to designing sex offender programs (e.g.,
Beech, Craig, & Browne, 2009; Marshall, in press; Marshall, Anderson,
& Fernandez, 1999). When there exists better-developed intermedi-
ate-level theory, the task of documenting a local theory (i.e., the
theoretical rationale for a specific program) is much less arduous and
difficult, and requires less “guesswork;” the local theory can address
itself more narrowly to how it operationalizes the intermediate-level
theory. For example, imagine that an intermediate-level theory
specifies that for high-risk violent offenders (a) criminal attitudes
are important treatment targets, including both attitudes to violence
and attitudes to rule violation; (b) preferred methods of addressing
these targets in the process of treatment are collaborative and
questioning, rather than comprising rational RET-style disputation of
expressed attitudinal content; and (c) intervention is most effectively
aimed at the schematic level, rather than being directed at individual
attitudinal statements. The local theory then may limit itself to such
matters as addressing the types of criminal and violent schemas
considered most important, and the specific collaborative processes,
methods and techniques that will be used, along with the rationale for
recommended sequencing of the relevant content and processes
relative to other components.
However, there are few resources at this level for interventions
with generally violent offenders and high-risk offenders. One of the
aims of the proposed framework is to stimulate theoretical develop-
ment at this intermediate level, by imposing some structure on the
heterogeneity within RNR-based programs. Before discussing some of
the factors on which such a framework could be based, I briefly review
the history of cognitive–behavioral rehabilitation program develop-
ment in corrections. This review provides a context for the current
status of intermediate-level theoretical resources.
3. Historical trends in cognitive–behavioral program development
Endeavors intended to reform offenders must be as old as crime
itself, but current cognitive–behavioral approaches mainly owe their
origins to the development of social learning theory and the advent of
the cognitive revolution. The spotlight thrown onto the causal or
maintaining role of cognitive activity in human behavior, and the
recognition that intervening directly with cognition could result in
significant improvements in health and wellbeing, drove a rapid
expansion in psychological rehabilitation approaches over the last
30 years. Historically, for offenders, these developments were more
heavily influenced by emerging cognitive–behavioral therapies for
psychopathological problems; approaches that draw more directly on
theories specific to criminal behavior (e.g., Reasoning and Rehabili-
tation; Ross, Fabiano, & Ewles, 1988) became influential only more
recently.
Several distinct trends in the development of cognitive–behavioral
rehabilitation for offenders can be discerned. The majority of current
programs fall into one of three broad categories: interventions for
violent offenders, interventions for sex offenders, and programs for
general offenders (Hollin & Palmer, 2006). Specialist interventions for
sex offenders originated in the health sector as much as the criminal-
1
The term “intermediate level” was proposed by Tony Ward (personal commu-
nication, February 22, 2010).
justice system. Gene Abel, a psychiatrist, was an influential figure in
their early development (Marshall, 1996). The still widespread use of
the terms “pedophile” and “sexual deviance” reflect a view of sexual
offenders as mentally disordered, and of sexual offending as caused by
a deviation in sexuality. Sex offenders then are seen as different from
other criminals, and the more usual sources of criminal development
are de-emphasized in explaining their behavior. This history still may
help explain why even current cognitive–behavioral rehabilitation
theory and practice with sex offenders diverge somewhat from the
RNR model (Kirsch & Becker, 2006), despite a lack of empirical
rationale for doing so (Hanson, Bourgon, Helmus, & Hodgson, 2009).
Empirically, all but the lowest risk sex offenders are more similar than
different to offenders with other types of index offences (e.g., Serin &
Mailloux, 2003).
The history of cognitive–behavioral violent offender programs also
owes much to therapeutic developments in mainstream clinical
psychology; most particularly to Meichenbaum's stress inoculation
training, on which Novaco's influential anger management programs
were based (Novaco, 1977). Low intensity anger management
programs became the most popular treatment for violent offenders
during the late 1980s and 1990s, though compelling evidence that
they reduce recidivism risk remains elusive (Polaschek, 2006b).
Low intensity anger management programs are no longer the
treatment of choice for violent offenders, and Meichenbaum's theory—
originally intended for anxiety disorders—has serious limitations
when applied to higher-risk violent offenders. Almost inevitably,
these offenders have broad rehabilitative needs that contribute to a
range of types of offending, including violence; and programs based on
the centrality of anger pathology in explaining their behavior, or on a
stress-based approach are too narrow. In cases where forensic
psychiatric services are tasked with treating higher-risk violent
offenders—especially when the mental disorders that led them to be
contained in such facilities are predominantly personality disorders—
the value for interventions of more comprehensive criminal psycho-
logical models of violence has begun to achieve recognition (Howells,
Day, & Thomas-Peter, 2004).
Although programs for “general offenders” have less often been
seen as the business of mental health services, this third tradition also
has some important roots in psychiatry and clinical psychology. For
example, Yochelson and Samenow's (1977) phenomenological
investigations underpin one key approach, Cognitive Self-Change.
Another example is the work of pioneering psychiatrist Aaron T. Beck
in understanding clinical cognition, which shaped offender programs
based on the cognitive distortions model (McGuire, 2006). Equally
though, rehabilitation for general offenders was informed by devel-
opments in social problem-solving that predate the advent of purely
cognitive approaches (McGuire, 2005). For example Reasoning and
Rehabilitation (R & R) programs began in the mid-1960s and although
they are often described as “cognitive skills,” they target interpersonal
behavior as much as the inner world of the offender (Antonowicz,
2005), making them cognitive–behavioral rather than just cognitive
(Fernandez, Shingler, & Marshall, 2006).
In recent years these three separate traditions—programs for sex
offenders, violent offenders, and offenders in general—have remained
substantially intact. Recent developments have instead been more
generic: moving toward group-based treatment, increases in program
intensity, and in multi-modal therapies, and the development of
programs that recognize specific responsivity issues (e.g., intellectual
disability, gender differences, and indigeneity). Today, to the degree
that there are still separate traditions, their separation is as much a
quirk of history, as it is a rational matter. With occasional exceptions,
all but the lowest risk offenders tend not to specialize exclusively in
one type of offending. And their criminogenic needs correspondingly
overlap (Hollin & Palmer, 2003; Serin & Mailloux, 2003); there are few
risk-related treatment targets that exclusively or even mainly map on
to just one offence type, with the possible exception of deviant sexual
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
arousal in some child-sex offenders (Hanson & Morton-Bourgon,
2005). Therefore, a fruitful strategy for the development of interme-
diate program theories, and for program development in general will
categorize program heterogeneity using variables other than index
offence type and putatively uniquely-linked treatment needs (e.g.,
anger-violent offending; Howells et al., 2005; child-sex offending-
intimacy deficits).2 What are these variables?
4. Proposed dimensions for differentiating programs in a
framework
Many of the obvious differences between programs are dimen-
sional rather than absolute. Any system for categorizing programs will
rely primarily on these dimensional variables; making the drawing of
boundaries between categories appear somewhat arbitrary. Below, I
describe and evaluate a series of client, program and other
characteristics that vary across rehabilitation programs, and that
may indicate potentially meaningful theoretical differences between
programs.
4.1. Client characteristics
4.1.1. Criminal risk
RNR's risk principle stipulates that “intensive service is reserved
for higher-risk cases” (Andrews et al., 1990, p. 20). Understandably,
terms such as “higher risk” and “intensive service” are not precisely
defined, and in program evaluation studies, there is substantial
variation in the estimated risk level of clients labeled “high,”
“medium,” and “low” risk, just as there are in the methods used to
assess risk. This lack of precision in risk measurement is one likely
reason that the risk principle has yielded the most variable effect sizes
in meta-analyses based on the RNR principles (Aos, Miller, & Drake,
2006; Smith, Gendreau, & Swartz, 2009).
Criminal risk is a variable that also has important theoretical
limitations. Given that it is most reliably assessed using historic or static
factors, it is best thought of theoretically as a “placeholder” or proxy
variable, and it is used on this basis as one of the two main organizing
variables in the proposed framework. The current level of criminal risk
offenders pose, particularly for more serious (e.g., interpersonal
offences) provides an indication of the range and depth of difficulties
that underpin their criminal propensity and the quality of their
lifestyle.3 Consequently, relative risk level also suggests the likely
intensity and quality of human service that will need to be invested in
rehabilitation for significant changes to occur. Since dynamic risk
factors are treatment needs, higher levels of risk correspond to higher
levels of, and more, needs; but not to distinctive needs. Second, at
higher levels of risk, a number of responsivity difficulties become more
prevalent, even normative. Consider, for example, PCL-psychopathy
(Skeem, Polaschek, & Manchak, 2009), and low readiness for change
(Polaschek & Ross, 2010). Finally, higher-risk offenders often spend
more time in “the system” over their life course, giving them an
institutionalized outlook, and too often, an entrenched criminal
identity, and few prosocial connections to the community. All of
these factors suggest that the stable trait-like aspects of their risk levels
provide a basis for differentiated programming.
4.1.2. Readiness and responsivity
Specific responsivity (Andrews et al., 1990; Andrews & Bonta,
2006) encompasses all of the factors that can impede or enhance
2
To clarify, what is argued here is that anger regulation difficulties and intimacy-
interpersonal relationship deficits are treatable risk factors that make a contribution to
criminal propensity in general.
3
Although there is controversy about the relevance of non-criminogenic needs to
reducing criminal risk (e.g., Ward & Stewart, 2003; Wormith et al., 2007), there is little
doubt that conventional life success is inversely related to the extent and duration of
involvement in crime (Farrington et al., 2006).
23
offender learning (Ward, Day, Howells, & Birgden, 2004). Some
specific responsivity issues have been addressed through the
development of specialized programs, such as those designed for
offenders with intellectual disability, for women, for indigenous
people, and youth.
Whereas being intellectually disabled or a member of an
indigenous nation may not in itself be criminogenic, there are other
characteristics that both affect responsivity, and generate risk. More
recently, Bonta and Andrews (2007) described responsivity as the
socio-biological-personality factors that can hinder treatment. Of
these, two—psychopathy and motivation to engage in treatment—are
the most commonly discussed (Howells & Day, 2007; Ward et al.,
2004; Wong, Gordon, & Gu, 2007).
Recently, motivation to undertake treatment has been absorbed
into a more complete construct: treatment readiness. Howells and Day
(2003) defined treatment readiness as “the presence of characteristics
(states or dispositions) within either the client, or the therapeutic
situation, which are likely to promote engagement in therapy, and
which, thereby are likely to enhance therapeutic change” (p. 320).
This definition captures the essential interplay between the client's
current readiness stage and the capacity of the intervention to
respond to the client. It suggests that both clients and programs vary
in their capacity to fit to the other in ways that enhance client
progress.
Concern has been expressed that many CBT rehabilitation
programs are highly structured, content-focused, and manualized,
with protocols requiring close adherence to the session structure
(Hollin, 2006; Marshall, 2009). They have little capacity to respond to
differences in client readiness, and are often pitched for offenders who
are ready to engage with what the program offers, and begin to
change almost immediately (e.g., anger management; Howells & Day,
2003). Readying unready offenders for treatments that require a
degree of readiness, can be tackled through preparatory programs
(Marshall & Moulden, 2006; see also Laws, 2002), or brief motiva-
tional interviewing interventions (McMurran, 2009). However, high-
risk offenders often present more extensive and durable obstacles to
treatment engagement, to learning and to change, and these
difficulties are more closely related to, or are themselves, the risk-
related goals of treatment. More intensive programs for these
offenders can take individualized approaches, addressing low read-
iness within the main treatment program (Howells & Day, 2007), and
over much of the course of the treatment (Polaschek, 2010). Most
high-intensity programs now take this approach.
Although “learning styles” is a term often listed among responsiv-
ity factors, little specific attention has yet been given to the
relationship between how programs disseminate content and how
offenders learn (Annison, 2006). However, there are important
differences in the extent to which programs can facilitate offender
learning; some of these are noted in the next section.
4.2. Program characteristics
4.2.1. Intensity or dosage
The other side of the risk principle concerns the delivery of a range
of levels of service; there are substantial differences in how the
intensity, or treatment dosage aspect of the risk principle is achieved.
In some jurisdictions, self-contained programs of different intensities
have been developed and are directed at offenders in particular target
risk bands (Bourgon & Armstrong, 2005, New Zealand Department of
Corrections, n.d.). In other settings, higher doses are achieved by
combining an individualized sequence of program modules into a
treatment plan based on an individual offender's range of needs. A
third approach has people stay more or less time in an open group
program, with rolling admissions.
It is not yet clear whether such differences are associated with
variation in outcomes. On the face of it, there are advantages and
24 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35

disadvantages to each of these approaches. A single, intensive, self- comes, and whether all are needed. But in practice, a compartmen-
contained program may be particularly effective with higher-risk talized view of the relationship between the goals of a component and
offenders because it can offer a sustained period of time in which to how participants actually change often appears unsupported; parti-
develop a therapeutic alliance to facilitate engagement and change cipants often demonstrate changes on treatment targets in sections of
(Hogue, Jones, Talkes, & Tennant, 2007). It may be easier in a single the program that are not intended to initiate those changes (e.g., a
integrated program for offenders to learn how their dynamic risk participant is adamant that he does not subscribe to criminal thinking
factors work together to create their own particular high-risk early in the program, but suddenly acknowledges it in the relationship
situations. The disadvantages of such programs include the slow skills module). However, more intensive, and more overtly therapy-
throughput of clients, and there may be significant parts of the like programs can be less concerned with a compartmental content-
program that are not relevant to some offenders. based view of change; they can pay equal or more attention to the
Advantages of “mixing and matching” from a suite of program therapy process for individual clients, rather than focus solely on
modules may include a superficially better fit to individual offenders' content to effect change (see next section).
treatment needs: an offender who has never had an alcohol or drug
problem simply does not take that module. In addition offenders can
4.2.3. Program delivery methods, change processes and context
repeat a module if they don't “get it” (Marques, Wiederanders, Day,
An assumption of CBT rehabilitation is that in order to change,
Nelson, & van Ommeren, 2005). The overall intensity of the individual
offenders must learn and then use new information and new skills. A
offender's required curriculum can be varied by risk level. And
CBT approach is distinct in the range of methods and processes
although some jurisdictions offer different levels of intensity for a
employed to facilitate participant learning: including “modeling,
particular module (e.g., cognitive skills programs in the UK; McGuire,
reinforcement, role-playing, skills building, modification of thoughts
2006), otherwise there is the increased flexibility of being able to
and emotions through cognitive restructuring, and practicing new
place offenders of different risk levels into the same programs. One
low-risk alternative behaviors over and over again in a variety of
disadvantage of this approach may be that completion of a series of
situations until one gets very good at it” (p. 283, Andrews & Bonta,
stand-alone interventions, each of which uses relatively superficial
2006).
change processes (e.g., limited opportunities to practice new skills)
CBT programs vary in how many of these techniques they employ,
may allow more criminally entrenched offenders to move through the
and in what proportions. Programs range from those that rely more on
whole sequence without effecting much change. A second concern is
classroom-style delivery of information with limited opportunities for
the risk-enhancing effect of mixing of higher and medium-risk cases.
demonstration, role-playing and practice, through to interventions in
Third, there is likely to be significant repetition of the processes of
which there are extensive practice opportunities that may even allow
program orientation and group formation as new combinations of
for generalization across settings.
clients come together in each new module, which may disengage
Shorter programs must often rely on a higher proportion of time
some offenders. There is also likely to be some repetition of concepts
spent in more didactic teaching of information, with limited
and skills across modules, but for some, this may actually be
opportunities for applying the skills in the program itself. Conse-
advantageous, in consolidating learning.
quently, the change process relies more centrally on the participants
Rolling open programs can offer the benefits of both approaches
to apply what they learn to their life outside the program; this
outlined above. Offender clients can come and go on a basis related to
circumstance explains in part why it is so important that the
their treatment needs and to how readily they learn, and make
participant be ready for such interventions.
changes. Content can be tailored to each person's formulation. The
More intensive programs can set aside more time for reinforcing
program is always full because places can be taken up as soon as they
what is taught with a variety of types of supervised practice, detailed
become vacant. However, such programs require a very high level of
reviewing of homework exercises, and adjustments to skill execution
skill from therapists, excellent theory manuals and a high level of
as competence increases. Such programs also can be less rigidly
oversight to ensure integrity.
structured: allowing more opportunities for using the treatment
process itself, including other group members or program community
4.2.2. Treatment targets and program components
processes, to facilitate offender change. In other words, a group can
Programs usually set out to target particular clusters of risk-related
function as a collection of individuals; each working separately, or
treatment targets (e.g., stable dynamic risk factors). Decisions about
group members can be involved in assisting and supporting each
which targets to include are influenced by the program duration,
other's learning. Some groups are constituted only for the hours they
program theories, or the type of offending targeted for change.
are “in session” with members returning to disparate residential
Program components are correspondingly variable. The local
environments. In other situations, the group lives and works together;
(i.e., program-specific) theory should provide a rationale for the
the treatment environment extends out of both the time and space of
relationship between each program component and particular
a session. In therapeutically supportive settings, many more processes
treatment needs. This relationship often is not transparently obvious.
can be employed to stimulate, support, and observe change.
For example, do we teach problem-solving as a conflict resolution
Consequently, more attention can be given to the creation of a
strategy, or to reduce impulsivity, or both? The role of a component
therapeutic climate within the larger social units in which the
may also alter according to its relationship to other components. For
offender is embedded during treatment (Marshall, 2009).
example, distress tolerance may be taught early in order to equip
participants to deal with later emotional challenges in the program, or
later as part of an affect regulation model targeted at avoiding 4.2.4. Facilitator-related program characteristics
offending. Identification of offence processes (i.e., offence mapping) The fourth program-related dimension is staffing. Facilitators are
may serve to motivate offenders to develop individualized plans early the vehicles for achieving the general responsivity principle, which
in treatment, or may be used later in the program to identify offence- has two components (Bonta & Andrews, 2007): the ability to develop
supportive thinking, or still later, to undertake relapse prevention a good working alliance, and skill in effecting change in the participant
activities for dealing with high-risk situations following release. through the use of cognitive–behavioral methods. Although both of
The program content is arguably the aspect that receives the most these elements are essential to effective treatment, and can be
attention from designers, facilitators and evaluators, particularly with targeted in training (W.L. Marshall, in press), the first also depends on
shorter, highly structured programs. In these programs, there may be facilitators' pre-existing personal characteristics (e.g., interpersonal
concern about establishing which components produce which out- warmth and sense of humor; Ross, Polaschek, & Ward, 2008).
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
The need to stretch budgets and human resources across growing
prison populations generates incentives to design programs that can
be delivered successfully by people with little training and experience
in offender rehabilitation. Thus, therapists and facilitators vary very
significantly in the skills and experience they bring to their roles.
At one end of the continuum, staff employed primarily in
surveillance roles may be trained to become program deliverers
(e.g., custodial staff). Occupying the middle ground are staff members
with a human service background who may have experience in group
facilitation. Often these deliverers are not trained in rehabilitation
theory, and their correctional rehabilitation experience may be
limited to the program they are delivering. The highest standard of
therapists undergo intensive training, and are expensive, and often in
short supply: for example forensic clinical psychologists, trained in
group therapy skills, knowledgeable about the psychology of criminal
conduct, experienced in cognitive–behavior therapy, and committed
to empirically-supported practice. In addition to variation in facil-
itators' skill and preparation for their treatment role, the levels of
support, supervision and training provided for their ongoing work
also vary.
4.2.5. Treatment integrity approaches
The methods used to achieve treatment integrity logically also
should vary according to the nature of the program, background and
skill of staff, and difficulties posed by the client group. Very close
levels of supervisor scrutiny of sessions may be employed when there
is high concern about therapeutic drift, perhaps because staff have
only brief training: or because the program lacks adequate theory and
facilitation manuals. In more intensive and complex programs, and
where staff are more skilled, monitoring of integrity may be both
more intermittent, and may even be based on adherence to program
theory (e.g., session goals) rather than detailed compliance with
specific program content.
4.2.6. Offender assessment, program individualization and treatment
change monitoring
Programs diverge in the type and extent of information collected on
participants before they enter the program, and on the use of that
information. Some programs screen referrals for basic demographic
criteria and collect no information directly from participants before the
program begins. Others assess in order to ensure that the correct
program placement has been made. Still others use assessment
information to develop individual formulations and treatment plans
for each offender, that may then be used to tailor or adapt the program—
as far as is possible—to the offender's needs.
The quality and volume of pre-program assessment information
also is related to the methods used by programs for monitoring client
progress and change. In some programs, change is inferred from
learning-related variables such as attendance, participation in group
sessions, and completion of any homework assignments. In others,
measures of treatment targets are taken pre- and post-treatment:
often, psychometric self-report instruments are used, or less often,
structured clinical judgments or behavioral ratings may be employed.
The range of dimensions on which programs vary is almost
endless; some have been outlined above. Because they were chosen
for theoretical relevance, they can be combined in predictable ways,
providing a basis on which to propose a tiered organizational
structure for offender rehabilitation interventions.
5. Proposed program framework
The previous section indicates the wide array of considerations
that go into designing and delivering rehabilitation programs. In this
next section, a framework with three levels is outlined and illustrated
using existing programs. The framework uses offenders' risk level and
program intensity, or “dosage” to anchor each level. Although this
25
strategy initially may appear reductionist, it was chosen because
offender risk and program dosage proxy for a number of factors that
both theoretically and empirically, are central to any rationale for
tiered program provision. The framework makes explicit some
assumptions about clients (Table 1) and the consequent program
responses that follow (Table 2), and the implications of both for
program policy (Table 3). A key tenet of the proposed framework is
that simply increasing treatment dosage as risk increases may not
yield equivalent effect sizes across the spectrum from medium to very
high-risk offenders. Rather, at the upper limits, it is not simply
treatment hours that increase, but the degree to which the program
fits around the formulated idiosyncrasies of that particular offender,
including dynamic (i.e., “in-flight-adjusted”) tailoring of content,
support for learning, and manipulation of the therapy environment to
reinforce and consolidate change.
5.1. Basic-level rehabilitation programs (low-to-medium risk, low
intensity)
Basic-level programs are quite brief, varying from 40 to 70 h of
intervention, and groups are usually closed, so that the cohort moves
together through the modules or sessions. Constraints on time limit
these programs to a relatively narrow range of dynamic risk factors
and intervention components (e.g., cognitive restructuring, relaxation
training, and problem-solving). Given these limitations, these pro-
grams will be more theoretically coherent if they specialize. In current
practice, one example of this specialization is seen in “cognitive skills”
or similarly labeled programs that teach a series of related thinking
and self-regulatory skills (e.g., problem-solving, empathy, self-
control, goal setting, and attitudes to crime).
Basic-level interventions are very structured; their manuals
specify session content in detail. Careful manualization helps to
guard against therapeutic drift, given that facilitators of these
programs have limited training in therapy, group skills, the theoretical
model, or psychology. Treatment integrity monitoring therefore is
often achieved by screening videotapes or sessions for adherence to
the manual.
Basic-level programs typically use delivery methods and styles
that are more psycho-educational, and rely more on classroom-style
teaching: hence the use here of the term “program.” The brevity of the
program, the structured curriculum, and the need to impart a
significant number of new ideas and skills in a short time lean such
programs more in the direction of “telling” than “showing” partici-
pants how to think and act. There is limited scope for practical
exercises, role-playing, discussion and practice. Homework may be
given but again, time often does not allow for personalized feedback.
The program schedule does not allow facilitators to respond in an
impromptu manner to “teachable moments” (Havighurst, 1952) that
arise, for example, due to an event outside of group sessions.
Facilitators may be given training in group process issues, but the
focus is more usually on demonstrating the interpersonal skills and
delivery style—such as interpersonal warmth, humor, and being firm-
but-fair—that help develop a positive group atmosphere, and are
associated with good outcomes (Dowden & Andrews, 2004; Marshall,
2005). There is little opportunity to attend to more specific process
issues that may arise in the group. Basic-level rehabilitation programs
often limit the involvement of the group itself in the change process.
The group format may be seen primarily as a way of increasing
program numbers, and is used mainly to elicit relevant examples and
role-play or practice interactive skills. The group process is not itself
the central vehicle for engendering change, as it can be in higher level
programs.
Basic-level programs usually are pitched at a level that assumes a
degree of readiness to change from clients. Although these programs
may include one or two motivational sessions at the beginning, they
are not suited to dealing with clients who need more assistance to
26 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35

Table 1
Summary of assumptions about clients, by program level.

Basic-level rehabilitation programs Mid-level multi-factorial treatment programs High-level comprehensive forensic therapy
programs

Readiness to Client relatively ready to engage with program, Client more resistant to program engagement Client reluctant to acknowledge treatment needs,
engage in develop working alliance with therapist. but will respond with some additional assistance. to seek assistance, may avoid engagement with
treatment therapist(s).
Level and range Relatively fewer problem areas, difficulties less More targets, somewhat more entrenched, but Wide range of treatment targets, chronically
of treatment entrenched. some areas of relative strength, functionality. entrenched problems, complex interactions
needs between them.
Change process Conscious acquisition of skills and knowledge from Client needs more assistance with learning; more Client finds new learning difficult; requires
didactic teaching and limited practice. use of active methods. significant support to learn.
Client determines what is relevant, takes what is Client needs more direction but open to applying Client unlikely to change without substantial
learned and applies it outside program. learning, with limited additional support. direction and support for change.

engage in change, have difficulty learning, or who have personality
issues that make them chronically resistant to influence. Individual
responsivity issues cannot be attended to in basic-level programs.
Tailoring of program content to offenders' needs and an offender's
particular previous experiences (e.g., examining previous offences in
detail) also is not usually possible in a basic-level group program;
there is little time to focus on individual offenders.
Facilitators' knowledge of participants is more limited and will
increase only somewhat during the program. Although groups may
practice exercises together, a high level of trust and familiarity
between group members is not essential to the program process.
Similarly, while there needs to be rapport between clients and
facilitators, the client does not need to develop the level of trust and
rapport with the therapist that is thought to be required to undertake
more therapeutically challenging changes (e.g., personality style).
Although a working alliance may develop that is suited to the
demands of the program, it will not be based on the depth or breadth
of shared experiences of a more intensive program, where there are
more demands and opportunities to take psychological risks (Ross
et al., 2008).
Client assessments prior to basic-level programs often are limited
to checking referral criteria (e.g., risk level and consent to attend).
Less often pre- and post-program change measures will be used. Self-
report questionnaires are the most suitable since they require little
facilitator training to administer and score. Reporting on change, if
done, often centers on whether participants understood the content,
and showed they grasped the skills for each module or session, if there
has been opportunity to demonstrate these skills. Progress may also
be judged on participants' demeanor and attitude to the intervention
and the facilitators. Usually behavior outside group makes a limited
contribution to evaluation of change; it is assumed that learning
content is sufficient for change to follow.
The outline above (also see Tables 1 to 3) is notable for the number
and range of limitations ascribed to basic-level programs. However,
for policy-makers and administrators, basic-level programs therefore
represent a way of getting some level of programming to a larger

Table 2
Summary of program responses to assumed client characteristics, by program level.

Basic-level rehabilitation programs Mid-level multi-factorial treatment programs High-level comprehensive forensic therapy
programs

Focus of responsibility More client-determined
for fit between Assessment limited to referral criteria
client and program or list of needs.
Group closed. Common content delivered
in fixed order to all participants.
Application Higher proportion of total treatment hours
of learning spent in teaching knowledge and skills.
More limited range of delivery methods,
opportunities for client learning.
Sources of influence Mainly offender self-reinforcement.
and support Residential environment may be actively
for change unsupportive, interfere with change.
Assessment and Assessment information may be limited;
monitoring of self-report scales may be used, or facilitator
client change impressions of learning, based on
demonstrations in group. Looking for
enhanced understanding, and limited
demonstration of skill acquisition.
Approach to Program focus assumes client ready to learn, Pre-program preparatory intervention, or early in
(un)readiness and use learning to change.
Balance between program- and client-
determined. Program broadly identifies
client needs, but constrained
in ability to respond.
Some more time in program for application
of learning. More scope for feedback on
homework to consolidate learning.
More sources of information to clarify learning,
practice with (group members, more than one
facilitator, residential staff).
Encouragement, reinforcement for change from
group members, facilitators. Residential environment
may be globally supportive of change.
May collect more detailed assessment information.
Judgments of progress informed by opportunities for
more extensive demonstrations of knowledge and
skill acquisition in group. May be more use of pre-
program/post-program assessment information.
treatment itself to develop readiness, followed by
intervention that assumes client is ready.
More program-determined
Formulation-driven assessment
“individualized treatment in a group context”.
More time in program for application of
learning.
More sources of feedback.
Progressive, graded practice.
More opportunities for generalization.
Clients may work with multiple facilitators, but
relatively longer relationship, more stability.
Family may be involved. Residential
environment may be used to elicit specific
changes from individual clients, support
behavioral experiments, provide reinforcement
for graded practice; generalization of learning.
Detailed initial assessment and formulation.
Change information derived from requiring
client to demonstrate learning in group and
outside it, may be required to demonstrate
generalization to community, work, etc.
Information on progress sought from outside
immediate therapy team. May use structured
clinical judgment instruments to monitor
change (e.g., VRS; Wong & Gordon, 2000).
Intertwined with/overlaps treatment of risk,
client unreadiness dealt with throughout the
course of treatment.
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35 27

Table 3
Policy implications of methods of program-client fit.

Basic-level rehabilitation programs Mid-level multi-factorial treatment programs High-level comprehensive forensic
therapy programs

Therapist skills Facilitators usually have limited training
background, may have dual roles (Probation
Officer, Custodial Staff).
Trained specifically to run this program: knowledge
of CBT limited. Background not assumed in
rehabilitation theory, or group process.
Manualization Program content and delivery highly manualized:
little/no facilitator discretion.
Treatment integrity Treatment integrity = adhering to manual. Staff
monitoring delivery closely scrutinized (videotapes/sessions)
for compliance with program manual.
Facilitators have psychologist, generic social
service professional training; may have generic
CBT training.
Facilitators trained to administer program:
training often longer to accommodate
complexities of program, and group process
issues.
Manual may not be so prescriptive, detailed;
manualization balances content with process;
more attention to process within group sessions.
Treatment integrity = staff self-reflection, peer
review, supervisor scrutiny of videotapes/sessions
for compliance with manual.
Facilitators highly trained in therapeutic
skills, cognitive and behavior change
methods, working with personality
disorder, etiological and rehabilitative
theory, and
group process.
Program manual oriented to achievement
of client goals and session competencies,
with latitude for therapists to choose how to
achieve goals with client, adjust goals and
methods as progress unfolds. Manualization
guides choices but draws more on common
and deep understanding of therapy approach
between therapists and with supervisors.
Monitoring of integrity requires skilled
supervisor who knows program theory well.
Extensive supervision, peer, group, and
one-to-one needed. Integrity monitoring
folded into professional supervision
and development.

number of offenders, given serious resource constraints. They do this
in two ways. First, the number of hours is relatively low, leading to
good throughput. Second, staff who run these programs are relatively
inexpensive compared to highly trained therapists. However, there
may be a temptation to “roll out too thinly” such programs: the
importance of devoting sufficient resources to staff recruitment,
training and supervision, and to monitoring program integrity has
been highlighted in a number of studies (see Andrews & Bonta, 2006;
Bourgon, Bonta, Rugge, Scott, & Yessine, 2009; Goggin & Gendreau,
2006).
5.2. Examples of basic-level programs
One of the lowest-intensity examples of a basic-level group
program was developed in New Zealand in the early 1990s: the
Video Anger Management Program (VAMP) comprised 10 sessions,
and covered understanding and recognizing anger, restructuring
angry thinking, coping with provocation, communication and asser-
tiveness, conflict resolution, relaxation and time out, and so on; with
just 2 h devoted to each major area. An adaptation of the VAMP in
South and Western Australia was delivered by trained tutors, and
evaluated by Howells et al. (2005), using a large sample of male
offenders (mainly prisoners, mainly index offences of violence).
More recently, in the UK, the Pathfinder suite offered to those on
Probation Orders includes seven programs, most of similar intensity,
but varying in their breadth of focus and referral criteria (e.g., violent
offenders, people who are not responsive to treatment in group
settings, offenders with alcohol- and drug-related offending, and
general offenders). Hollin et al. (2002) evaluated these programs and
reported that all were cognitive–behavioral, most used group formats,
content was well documented, and they were run by up to two
facilitators who were mainly specifically trained and supervised
probation officers. Some programs collected information from
offenders before and after to enable change monitoring. There was
an emphasis on training to deliver the manualized program, and skills
in group work were sometimes noted to be overlooked in training.
Pre-program assessment for some of these British programs
includes psychometric tests, but the information is used most often
to decide suitability for referral, and less often to document pre- to
post-treatment changes. Methods used in programs include practical
exercises and assignments, role-playing and discussions between
participants and tutors (McDougall et al., 2009).
5.3. Mid-level multi-factorial treatment programs (medium to high risk
and intensity)
Mid-level interventions are significantly longer; current programs
in this category offer treatment over 100 to 300 or more hours. The
interventions target multiple dynamic risk factors, include a broad
selection of intervention components, and are able to use a wider
range of learning processes to engender client change. Typically, there
is a single or main group program that is often closed so that
participants work through the sequence of interventions together
(see Tables 1 to 3 for a summary of the characteristics of this level).
Mid-level programs are most suited to higher-risk clients (Palmer
et al., 2009; Polaschek, 2006a). A higher-risk client group will have
more severe and diverse treatment needs, and more responsivity
issues: especially readiness-related difficulties. Manualization of mid-
level programs, particularly closed-group programs, may still be
detailed but manuals explicitly emphasize both process and content
(Mann, 2009), and the degree of specification is related to the
qualifications and experience of the facilitators, the quality of the
supervisory oversight, and the clarity of the program theory.
Although more hours in treatment may be used simply to increase
the complexity of the content delivered, there are other—arguably
more effective—ways to use the increase in treatment time. More time
gives more flexibility for facilitators or therapists to use group
processes to help participants to learn, and to facilitate changes in
client readiness, in relevant thinking, and in personal and interper-
sonal skill development. There is more opportunity to practice skills,
to have group discussions and for group members to assist others and
to reinforce others' change (Serran, Fernandez, Marshall, & Mann,
2003). There is time to review homework, and to focus on the
circumstances of a single client during a session. Therefore there also
is more opportunity to develop a therapeutic alliance, and to use that
alliance to undertake therapeutic tasks that are not possible in shorter
interventions. Further, constructive alliances may develop between
the group members that can be harnessed therapeutically.
Even if the realpolitik is taken into account (Hollin, 2009), still,
these are programs that should be run by skilled, more deeply trained
facilitators, and should allow room for supervised clinical innovation
and therapist judgment. However in practice, facilitators at this level
currently span the range from highly qualified psychologists to more
broadly trained human service providers such as social workers, and
nursing staff, to facilitation-trained correctional officers.
28 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
Depending on the nature and degree of the readiness difficulties of
the potential client pool for this level of program, the best responses
to readiness problems also will vary. Pre-program individual
motivational interventions (Anstiss, Polaschek, & Wilson, in press)
or preparatory groups (L. E. Marshall & Moulden, 2006) may be
effective solutions. However, if the referral group is a particularly
challenging and high risk one (e.g., men with high PCL-psychopathy
scores; Anstiss et al., in press), methods of working with low readiness
may be best incorporated into the treatment itself; readiness
difficulties may continue to arise despite preparation for entering
treatment.
These higher-risk clients are also hypothesized to need more
assistance in recognizing and understanding their treatment goals.
Time spent focusing on how a particular treatment target is related to
a particular client's pattern of offending can make such acceptance
more likely, and this time is more likely to be available in mid-level
programs. Although they still are structured, there can be scope to
adapt the program as it unfolds, to the nuances of an offender's
formulation.
At this higher level of intensity, programs are more likely to be
delivered in residential settings that are set aside for therapeutic or
rehabilitative use. Living together allows program participants to
develop stronger—and potentially more therapeutic—relationships
with residential staff (e.g., custodial officers, residential workers, and
nurses) and with each other, assuming adequate supervision. It
reduces possible antisocial effects from contact with offenders who
are not involved in rehabilitation. Carrying out homework exercises as
a group or with others becomes possible. However, in the proposed
framework, programs at this level stop short of fully harnessing the
therapeutic change potential of the residential community in which
the group sessions are embedded. The program setting is generally
positive and supportive, but not directly integrated into intervention.
Just as mid-level programs can give more scope for engagement
with individual offenders in order to achieve change, so assessment
information can play a much more central role in that process than
with the basic-level programs. Individual clinical formulations of each
participant provide information that can be used within the program
itself, which is not usually possible with shorter programs.
More elaborate monitoring of change, including access to
information from outside observers also becomes more feasible than
in basic-level programs, particularly when there is close co-ordination
between residential staff or family, and therapy staff. Mid-level
interventions commonly also use psychometrics and self-report scales
to document treatment change.
Several programs at this multi-factorial level are offered to
clientele with specific types of index offences (e.g., sexual or violent).
One important issue with such programs is whether the intention is to
reduce risk of a specific type of criminal behavior, or to reduce all
types of criminal risk in particular offenders. Since those at high risk of
serious offending are rarely specialists, programs that conceptualize
them in this way may overlook important treatment targets, or
methods (Kirsch & Becker, 2006). Perhaps the most obvious are
factors related to a general antisocial orientation, such as criminal
attitudes (Hanson & Morton-Bourgon, 2004; Mills & Kroner, 2006). So,
for example, sex offender programs can incorporate Reasoning and
Rehabilitation components to enhance effectiveness (Marshall, in
press).
5.4. Examples of mid-level multi-factorial treatment programs
Several mid-level programs provide excellent examples of rigor-
ous, carefully designed and well-implemented cognitive–behavioral
interventions for offenders, often meeting stringent criteria for
accreditation. As a consequence, while they are state-of-the-art in
current practice terms, they may have been required to specify session
content at a level of detail that does not allow for the constrained
flexibility proposed here for mid-level programs run by skilled,
psychologically-trained staff.
Canadian Corrections' Violence Prevention Program (VPP) was an
accredited program for repetitive high-risk violent offenders (Vio-
lence Prevention Program, n.d.). Treatment was delivered for
approximately 240 h over 16 weeks, in 8 two-hour sessions per
week. The program followed the same broad approach as the Anger
and Other Emotion Management Program (Dowden, Blanchette, &
Serin, 1999), but was more intensive, more personalized to the
offender, and accepted violent offenders regardless of whether they
had difficulties with anger regulation, and included motivational
enhancement components (Correctional Service of Canada, 2000).
The program began with orientation and introduced basic motiva-
tional enhancement approaches, then examined participants' own
violence history, before moving on to skills training in anger control,
social problem-solving, attitude change, relationship development
and conflict resolution, developing positive lifestyles, and self-control.
Lastly, a comprehensive violence prevention plan was developed.
The program was delivered in closed groups, with two dedicated
program facilitators including a postgraduate qualified psychologist
with knowledge in criminal psychology, and an experienced program
delivery officer, who was a graduate in a related field.
On a continuum anchored at one end by a high level of detail and
structure, and at the other by extensive clinical judgment and
therapist discretion, the program took a more explicitly “centrist”
position. It accepted participants who were not ready to change when
they entered the program, and staff were trained to work dynamically
with low readiness. Staff also were trained to customize material to
the individual treatment needs; although there was a scripted
manual, staff were told that it provided an example of how to deliver
the material (Correctional Service of Canada, 2000). Facilitators were
encouraged to develop supplemental material, and could make
recommendations through a carefully managed quality assurance
process, for changes in structure and content. The VPP was delivered
across a wide range of facilities, and did not require a dedicated
treatment environment.
In a second example, Bourgon and Armstrong (2005) reported on a
tiered prison-based cognitive–behavioral program, offered at the
Rideau Correctional and Treatment Centre in Ontario, Canada to
offenders in general. Three levels of programming—5, 10, and
15 weeks, corresponding to approximately 100, 200, and 300 h of
treatment—were offered, with each program provided in a separate
dormitory in the treatment center, in a closed group format (Bourgon,
personal communication, 17 February, 2010). All new arrivals were
interviewed, and undertook a structured assessment for risk, and
treatment targets. Amenable prisoners completed additional psycho-
metric scales, and were placed in one of the three levels of treatment
program on the basis of assessment information. The programs
differed in scope and intensity, and focused on replacing criminal
thinking and learning new skills. The longer programs devoted more
time to addressing each target, and incorporated more interventions,
including more role-playing and skills practice (Bourgon & Arm-
strong, 2005).
A psychologist and a social worker, each with graduate degree
training, facilitated each program. Correctional officers staffed the
center; they volunteered for placement in the unit, and acted as case
managers for the prisoners. Custody and therapy staff worked as a
team to run the program and create a therapeutic climate in the unit,
but the correctional officers had limited training in the content, and
goals of the program, and thus varied in their ability to support the
treatment itself (Bourgon, personal communication, 23 February,
2010).
Some programs currently offered at this mid-level of intensity use
a different format: open, rolling, and focused more directly on the
individual offender (Marshall, in press). There are definite advantages
to such a model. Most obviously, the program contents can be fitted to
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
the offender formulation much more closely. However, for these
programs to be successful, they require both a very high level of
therapist skill and experience, and a common understanding of
intermediate-level treatment theory.
Cognitive Self-Change (CSC; Bush, 2001) is an example of a mid-
level rolling-group program that differs in several important ways
from both of the previous two closed-group examples, and from some
of the characteristics of the framework proposed for mid-level
programs (Table 2). Program facilitators are often trained custodial
staff in the prison unit in which the program is housed; thus requiring
correctional officers both to maintain security and safe containment,
but also to act in a therapeutic role; identifying and supporting
prisoners in changing entrenched criminal thinking patterns. On the
face of it, this conflict of roles should engender mistrust and a lack of
engagement in participants. However, Bush (2001) noted that there
also were clear benefits; custody officers were strong advocates for
the program, demonstrated increased professionalism and invest-
ment, and developed an approach to working with offenders that
blended non-antagonistic authority with effective skills for calling
offenders' thinking to account and supporting change across the
entire residential environment. This approach is consistent with other
research suggesting that such a blended-role approach can be
successful, if staff use their authority therapeutically (Skeem, Eno
Louden, Polaschek, & Camp, 2007). Staff are certified through training
to run the CSC program, and supervised closely.
CSC has an explicitly cognitive focus; it teaches a series of cognitive
skills through extensive use of personalized exercises and practice
(Vermont Department of Corrections, 2002). Staff compile a compe-
tency development plan for each participant soon after program
entry. Thus, program completion is determined by completion of the
core and optional tasks in the plan; the final task is a “self-risk-
management plan” much like a relapse prevention plan for crimino-
genic thinking habits. Individual stays vary from 6 to 22 months:
making it a mid-level program at the higher dosage levels.4 The
program is delivered in open, rolling groups of 8 participants. Two
staff facilitate sessions of 2 h and the group meets twice each week.
Although sessions typically focus on one offender's thinking at a time,
group members are active participants in the feedback to that
offender (Vermont Department of Corrections, 2002). In its original
state of Vermont, CSC is delivered mainly to violent offenders.
5.5. High-level comprehensive forensic therapy programs
The third and final level in the proposed framework is character-
ized by: (a) targeting very high-risk offenders, or offenders at high
risk of serious interpersonal crimes, (b) a similar level of dosage
delivered in group treatment sessions as for the mid-level programs,
and (c) embedding of the program in a fully therapeutic environment
or setting.
Stipulating a cognitive–behavioral treatment model for high-level
programs is an interesting task, in that few such programs have yet
been described in the published literature. The approach proposed
here (see also, Table 1) draws from a number of existing therapeutic
traditions, but in practice, programs that resemble this category are
mainly quite new; in several cases they represent an innovative
response to increasing pressures to demonstrate the ability to change
the criminal risk of offenders who might formerly have been
considered untreatable (e.g., PCL-psychopaths, offenders with other
personality disorders).
In high-level programs participants are admitted into a relatively
stable community. High-level programs are relatively expensive and
resource-intensive. Multiple, highly trained therapeutic staff may be
working with the same small number of offenders. There may be
4
Only Stage I of the CSC program is described here. Following completion of Stage I,
participants undertake Stage II—a maintenance program—in the community.
29
extensive general rehabilitative programming alongside the active
therapy session time (e.g., occupational training or therapy, education
and work, art classes, sports and physical training, cultural and
spiritual programs, and literacy classes), with a much more explicit
focus on developing a healthy lifestyle alongside working on specific
offending issues. The interventions are housed in purpose-built
facilities, or at least in facilities dedicated to therapy. Considerable
time is spent in communication between staff members of the
community—updating progress for each resident and planning the
next steps in their treatment, aligning other aspects of their
management—and with the residents themselves: interacting with
them and monitoring them in their various activities.
Group treatment still dominates the formal program delivery, but
participants may be placed in a single group or in multiple groups:
simultaneously or sequentially. These new and innovative interven-
tions are more accurately labeled as “therapy” rather than “programs”
(Marshall, in press). They are still cognitive–behavioral and there are
still group sessions that center on educating, changing attitudes and
teaching skills, but sessions are delivered in an encapsulating
therapeutic environment that also operates broadly according to
cognitive–behavioral principles. In other words, these programs are
not simply set in a therapeutic milieu, or generally sympathetic
environment. Rather the people and other characteristics of the
treatment setting are linked into the treatment process. More than
simply protecting participants from the wider anti-change culture of
the mainstream prison, the context can be engineered to promote and
embed change. If all of the residents of a unit are engaged in the same
general program of learning and change, and if participants are
engaged in structured, supervised activity for much of the time, staff
may be able to gain control of the culture of the setting. With the co-
operation of the clients, it can be built into a prosocial and therapeutic
oasis within an otherwise antisocial or unsupportive institution.
The setting can be manipulated so that clients experience
contingent delivery of consequences for prosocial and antisocial
behavior from a variety of social sources including various staff, and
clients who may be further along in the program. In addition to
experiencing immediate positive and negative feedback for their own
behavior, participants can also be reinforced for increasing durability
of change on factors such as impulse control and self-regulation
through tiered access to privileges (e.g., token economies, tiered
employment systems, and opportunities for reductions in security
giving access to new opportunities).
The unit can provide many of the benefits and challenges of a small
community: opportunities for prisoners to practice skills both related
to criminogenic (social problem-solving and non-aggressive resolu-
tion of grievances) and non-criminogenic needs (e.g., developing
friendships, planning and executing ideas for unit projects, and
developing employment skills). Completion of therapeutic tasks can
involve assistance and support from multiple community members;
tasks can have high ecological validity, but there are opportunities to
practice initially with supportive and encouraging coaches, and later
with parties who are not cued ahead of time, but instead respond
naturalistically.
Therapy tasks thus can be graduated in difficulty to promote skill
development, based on feedback about client progress to date. Time-
limited individual intervention may be provided when specific
responsivity issues or other problems arise. Because a number of
therapy staff know the offender well, the characteristics of particular
therapy alliances can be capitalized on to effect change, or to get
offenders past obstacles (e.g., working with a woman therapist, with
an individual therapist who takes a different approach to that of the
group therapists, or who intervenes to resolve conflict at a time when
the group therapists' relationships with that offender may be
strained).
Within the community, the group itself can be targeted for change.
Working with a whole cohort as a social group has innumerable
30 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
benefits, given the myriad ways in which group members overtly and
inadvertently influence each other's thinking and behavior: even to
determining what will and will not be talked about in group
(Goldstein & Huff, 1993).
Highly skilled staff, a well-defined rehabilitation theory, substan-
tial program-relevant training, well constructed theoretical and
facilitation manuals and skilled regular supervision all can enable
such programs to take a distinctly different approach to treatment to
that which has become the practice standard for cognitive–behavioral
rehabilitation programs for offenders.
For example, the program manual for high-level therapy programs
is not highly prescriptive of the content that must be delivered in a
particular session. Instead the manual—drawing on rehabilitation and
treatment theories—may be constructed around a series of common
treatment goals and objectives that follow from those goals. The client
formulation determines which of the goals are relevant to that client.
Therapists use the formulation to develop collaboratively a treatment
plan for each offender, but once a plan is constructed and goals are
identified, the objectives become those that are theoretically specified
for that goal. Therapists are given some discretion about the methods
of intervention and therapeutic processes that are used to achieve the
objectives. For example, therapists choose from a manual-based menu
of options for how to achieve the objectives, and how to have the
client demonstrate that they have made progress on the pertinent
goals. However, they may also make a case in supervision for novel
options that achieve the same objective for a particular client
circumstance.
Current programs that most closely fit this proposed level
sometimes are based around having the client complete a selected
series of tasks or “assignments.” Assignments are structured so that in
order to complete them, clients must undertake a guided learning
experience over a number of weeks, and the integrated therapeutic
environment enables feedback from multiple sources about progress:
which can then be used to refine and extend the tasks. In essence,
these programs are delivering individualized treatment in group
settings.
When the program manual specifies what is to be achieved but not
how to achieve it, treatment integrity monitoring becomes a different
process; it is no longer sufficient to monitor content for conformity.
The guardians of treatment quality require much more skill and there
is greater need for peer supervision of both therapists and their
supervisors and monitors in order to avoid unintended drift and anti-
therapeutic innovation. The key tasks of the integrity monitor include
evaluating the quality of the individual client formulation, the links
between formulation and the treatment plan, the execution of the
plan including adjustments for client responses to date, and the
quality of monitoring of client progress. The manual serves to anchor
the evaluator by specifying common treatment goals, objectives, and a
range of ways of achieving these objectives, but there is no doubt
clinical supervisors need a refined grasp of rehabilitation and
treatment theory to ensure therapeutic integrity is maintained.
The target clients for high-level programs have a variety of labels,
including “PCL-psychopath” and “personality disordered.” They are
life-course persistent, career criminals with repeated involvement in
serious crime. They may have more than their share of other forms of
psychopathology such as head injury, learning difficulties, anxiety
disorders, and so on. In current political environments, many are
unlikely to return to the community without substantial intervention
and enduring demonstrations of change. Treatment may be used
simply to achieve a move into less restrictive—and less expensive—
institutional environments.
If considered on a stand-alone basis, then, the concept of
“treatment readiness” has limited utility in programs at this level.
Low readiness is widespread, even without factoring in the impact of
coercive and even involuntary processes that may be used to get
clients into these treatment environments. Factors underpinning low
readiness are normative because they also underpin risk: for example,
even to develop a working relationship may require progress on
criminogenic needs such as entitlement, mistrust, impulsivity and
poor affect regulation.
Mechanisms of change in psychotherapy in general remain poorly
understood (Johansson & Høglend, 2007). In the basic and to an
extent, the mid-level programs, the design suggests that offenders are
expected to change by learning in the program, and then setting out to
try new ways of thinking and behaving. However, there is virtually no
evidence addressing the accuracy of these assumptions, or whether, if
correct, they represent the most effective approach for all levels of
client risk.
Therapeutic changes may not occur through conscious interven-
tion, and may not be evident to clients. Research on implicit processes,
and implicit learning all suggest that there may be a parallel level of
processing that is not available to conscious awareness (Wilson,
Lindsey, & Schooler, 2000), but that nevertheless influences behavior
(Toderov & Bargh, 2002). If processes themselves are not consciously
accessible then changes in them may not be either, and some
preliminary research has found changes on implicit and explicit
measures may not even occur in synchrony (Polaschek, Bell, Calvert, &
Takarangi, 2010). Further, there is evidence to suggest that employing
methods that cause behavior to change without the client's active co-
operation (e.g., manipulating the environmental contingencies, or
even just persuading clients to try behavior they believe will not
work) engender cognitive change, possibly with more effect than
cognitive approaches themselves (Craighead, Sheets, Brosse, & Ilardi,
2007; Dowd, 2006). Although a hallmark of institutional interventions
during the heyday of behaviorism, the harnessing of elements of the
social environment as therapeutic tools is rarely described in more
contemporary publications on adult correctional interventions,
though its potential with higher-risk offenders is being recognized
in some newer program models. Lastly, it is likely that since different
forms of intervention can create change in the same target, equally, a
target area may change during a program module intended to change
some other factor. Taken together these points have two implications.
5.5.1. The therapy environment as a therapeutic tool
The first implication is that high-level programs can, and should,
use their therapeutic environments to generate change. Instead of
relying on the client to initiate new experiences intentionally, high-
level programs have more capacity to harness the behavior of others
and the characteristics of the residential environment to generate
change in participants.
Since the advent of cognitive approaches to therapy, behavioral
interventions have a much reduced role in the “therapeutic toolkit”
(Fernandez et al., 2006) except in residentially-based interventions
for juvenile or intellectually disabled offenders (Brown, Borduin, &
Henggeler, 2001; Day, 1988; Guerra, Tolan, & Hammond, 1994;
Holmqvist, Hill, & Lang, 2009). Such techniques represent an
extension of the general responsivity principle. They include methods
as diverse as manipulation of reinforcement schedules, and requiring
demonstration of behavioral practice in the residential setting. Many
programs currently rely much more heavily on cognitive aspects of
CBT and de-emphasize behavioral elements (Fernandez et al., 2006),
but in a high-level program a wide range of techniques of influence is
used.
In addition to manipulating contingencies in the residential
environment to engender change, it may be easier in high-level
programs to make more use of the supportive social environment to
encourage clients who are undertaking behavioral experiments.
Having clients test out their old or new assumptions about the
world is often considered to be one of the most effective techniques in
cognitive therapy for engendering change (Bennett-Levy et al., 2004).
Cognition and behavior interact reciprocally; enacting behavior will
change cognition (Fernandez et al., 2006). A focus on behavior may be
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
a particularly powerful route to change in clients who are resistant to
introspection, and struggle to develop insight using purely cognitive
methods.
5.5.2. Comprehensive assessment of change
The second implication that follows from understanding that
change can occur without offender awareness, is that monitoring of
change in high-level programs should not and does not need to rely
on client self-report. Self-report is considered problematic anyway:
program evaluations seldom have demonstrated links between self-
reported progress and distal externally-measurable outcomes such as
reconviction (Serin, Lloyd, Helmus, Derkzen, & Luong, 2010).
Although predictive validity failures often are attributed to offenders
“faking good,” they are just as likely a consequence of more general
factors such as otherwise helpful positive illusions about self-
improvement (Conway & Ross, 1984), insensitivity and inaccuracy
of recidivism as an outcome variable, the elapsed time from self-
report measurement to outcome, and unreliable or irrelevant self-
report measures (Loza, Loza-Fanous, & Heseltine, 2007).
In lower level programs, often the other main source of
information about change comes from the therapists themselves.
When therapists form strong alliances with offenders, it can be argued
that they should be somewhat “captured” by the client. For example,
they should develop genuine optimism and hope for the client that
reduces their ability to judge dispassionately the client's progress. In
forensic settings where they may be reporting on progress to bodies
such as parole boards, this possible loss of judgment may have serious
consequences, although it may be ameliorated by the use of
structured clinical judgment methods and rating scales that anchor
the therapist with evidence (Marshall, in press).
However, in a comprehensive rehabilitative environment, infor-
mation about change can be collated from multiple sources, including
other residents. Furthermore, there often are multiple therapists and
adjunct staff working with the client, giving more depth of
information across a wider range of working alliances. And the
requirement for clients to demonstrate new behavior in the group, the
community, and even in related environments (e.g., if there are
opportunities to go out of the unit for work, or other activities) gives
more confidence about the extent of actual progress. It becomes
possible to document and measure change in a more structured,
rigorous manner in which client self-report and the observations of a
primary therapist are but a part. Newer instruments have been
designed for this purpose, and are beginning to demonstrate links to
outcomes for violent and sexual offenders respectively, over and
above static risk estimates (Olver, Wong, Nicholaichuk, & Gordon,
2007; Wong, personal communication October 13, 2009), something
that offender self-report scales have rarely been able to do.
5.6. Examples of high-level comprehensive forensic therapy programs
The Dangerous and Severe Personality Disorder initiative began
with a commitment by the British government in 2001 to “develop,
pilot and deliver new services specifically for people who present a
high risk of committing serious sexual and/or violent offences as a
result of severe personality disorder” (DSPD programme: Dangerous
people with severe personality disorder, n.d.). The Peaks Unit at
Rampton Hospital is one of the two health-system sites for these pilot
programs; two others are located in prisons. Hogue et al. (2007)
recently outlined The Peaks program, which is based in a purpose-
built unit housing 70 male patients, and “aims to provide a
comprehensive service specifically focused on addressing the full
range of clinical and offending-related needs of those within the DSPD
group” (p. 57).
The DSPD pilot programs may represent the best-resourced
empirically-guided rehabilitative services ever provided for high-
risk offenders, with high staff-to-resident ratios and provision of a
31
wide range of therapeutic and other interventions and activities
within a supportive milieu. Individual formulations guide all aspects
of therapy and care. Individual patients enter into a sequence of
interventions designed to engage them, alter those features of their
personalities and interpersonal styles that have previously prevented
intervention and change, and address any unaddressed needs, before
developing a comprehensive relapse prevention plan (Hogue et al.,
2007). The milieu itself is conceptualized as providing continuous
support for engagement in change, and as reinforcing discrete
interventions, through therapeutic alliances, and the use of regular
community meetings (Hogue et al., 2007).
Interventions include orientation, coping skills, cognitive analytic
therapy, dialectical behavior therapy, schema-focused therapy,
trauma-based therapy, and specific programs for sexual and violent
offending (Hogue et al., 2007).
Staff from multidisciplinary therapeutic teams work collabora-
tively with each patient. Patient behavior is monitored 24 h a day;
nursing staff complete structured rating scales, and this information is
used to monitor progress, adjust treatment plans, and undertake more
acute patient management. Adjunct activities such as occupational
therapy are used as sources of additional information about current
levels of engagement and progress, as well as to provide patients with
activities they personally value. Close monitoring of each patient
enables detection of sometimes subtle patterns of behavior that
parallel parts of previous offence processes (i.e., offence-paralleling
behaviors; Daffern et al., 2007).
Contrasting in a number of ways with the DSPD initiative is a series
of programs developed by New Zealand's Department of Corrections
for rehabilitating male prisoners at high to very high risk of serious
criminal behavior. The New Zealand prison system currently contains
four purpose-built medium-security units dedicated to these high-
intensity treatment programs. Te Whare Manaakitanga—formerly the
Rimutaka Violence Prevention Unit—has been open since 1998 and
treats prisoners with a high risk of future violence, while the other
three units have been opened within the last two years, and take men
at equally high risk of future serious crimes but with no or fewer
violent convictions than men at Te Whare Manaakitanga.
Te Whare Manaakitanga is a “work in progress;” the program is in
transition between the mid-level described in this framework, and
this comprehensive therapy program level. Its treatment program
includes around 300 h of mainly group intervention, delivered in a
closed group of 10 men that meets 3–4 times each week for 2.5 to 3 h.
The unit has 30 beds; three cohorts begin the program each year, and
program completers stay on average around 9–11 months. The
program is designed for very high-risk violent prisoners nearing
parole, and is primarily cognitive–behavioral in orientation. Each
group is led by two therapists—one psychologist and one program-
trained facilitator with a social service background—and works
through a manualized treatment program, but with some emphasis
more on using multiple methods and processes for achieving
therapeutic goals rather than on specific content.
Following some introductory sessions that socialize men into
group therapy, the program modules include offence mapping
(identifying common cognitive, affective and behavioral precursors
to violent and non-violent offences), identifying and modifying
offence-supportive thinking, skills-based interventions to improve
affective regulation, impulse control, communication and relation-
ships, and finally, developing plans for reintegrating into the
community, and managing situations where risk of reoffending may
become acute. The program's integrity of delivery is monitored
through close professional supervision—group and individual—and
observation of sessions.
Outside of scheduled group sessions, program residents can take
part in a variety of other unit-based activities; recreational, employ-
ment and personal development (e.g., cultural identity and educa-
tion). In addition to the six therapy facilitators, the therapy unit has a
32 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
managing psychologist and a Maori cultural consultant who provides
supervision and training to the therapy team in working with the
mainly-Maori unit residents, and intervenes as part of the team in
individual and group sessions. The therapy unit is located in a
custodial environment managed by Correctional Officers who are
trained in how the unit works. Custodial and therapy staff meet
regularly to ensure continuity of intervention, gather information
about current behavior and change, and manage the many issues that
arise in relation to prisoner management. Each prisoner completes a
comprehensive pre-program assessment and a formulation and
individual treatment plan is developed collaboratively from that
information, and is reassessed regularly.
A key part of the transition to this comprehensive therapy level has
been the promotion of a fully therapeutic environment: known in
New Zealand Corrections as a “community of change.” The involve-
ment of custodial staff in extending each prisoner's treatment goals
out into the residential part of the unit, and in feeding back therapy-
related observations about men's behavior (e.g., offence-paralleling
behavior and instances of new prosocial skill demonstration) has been
a crucial part of this transition. Regular community meetings are used
to foster the therapy environment, and give residents another
opportunity to practice various interpersonal skills. Regular welcom-
ing ceremonies and celebrations—that bring into the unit outside staff,
agencies and families—help to create a sense of community that is a
point of difference from other units in the prison. The community is
viewed as a context for having men in treatment demonstrate and
generalize knowledge and skills taught in their formal group sessions.
Implementing a fully therapeutic environment at TWM is a
considerable challenge; these are very high-risk offenders. Without
change, about half of the men who are eligible to attend the program
will return to prison within 12 months (Nadesu, 2007). Men who
attend the program have mean scores on the Psychopathy Checklist:
Screening Version (PCL: SV; Hart, Cox, & Hare, 1995) indicating they
are at the cutoff for a diagnosis of psychopathy (Cooke, Michie, Hart, &
Hare, 1999). Treatment is delivered with relatively few resources, in
an environment in which initiatives to create trust and prosociality
must constantly push against highly entrenched institutional beliefs,
and a strong criminal gang culture. As an example of a comprehensive
level program, TWM still has some developmental steps to take, but
serves as one example of a high-level correctional program for very
high-risk men.
A second corrections-based illustration of high-level therapy
comes from the Regional Psychiatric Center (RPC) developed by
Wong and colleagues in Saskatoon, Canada. The RPC operates as a
maximum-security psychiatric hospital inside a prison, under the
aegis of the Correctional Services of Canada. With over 200 beds, it
runs accredited cognitive–behavioral programs for high-risk sexual
and violent offenders, along with various other rehabilitative services
based on the RNR model (Wong et al., 2005). The hospital is staffed
with multidisciplinary mental health teams and security staff, but
therapy and security roles remain clearly demarcated (Maden,
Williams, Wong, & Leis, 2004). Intervention is intensive in that
patients spend up to 30 h a week in rehabilitation; including
constructive non-therapy activities (e.g., work and homework assign-
ments; Maden et al., 2004); the overall pattern of intervention for RPC
attendees is more like that of the DSPD; participants attend multiple
group and individual interventions during their stay (Wong et al.,
2007). One program offered in the hospital for violent prisoners is
Aggressive Behavior Control (Wong et al., 2007).
6. Conclusions and questions for the future
From the outset, an examination limited to group cognitive–
behavioral RNR-based rehabilitation programs might be expected to
uncover a degree of uniformity. But the process of reviewing relevant
research and writing, and organizing it within the framework outlined
here reveals diversity on a number of fronts. But does this diversity
matter? Is it even interesting? It could be argued that organizing this
framework along such prosaic and practical dimensions as the
number of treatment hours, and whether therapists have been to
university, is too far removed from theory to inform it, but instead
captures differences that are not meaningfully related to improving
rehabilitation outcomes.
The counterargument put forward here is that program designers'
decisions about all of these factors reflect theory, whether explicit and
well-supported, or implicit, and unquestioned: perhaps even un-
founded. Adherence to the RNR model, or any model that functions as
a high-level theory for rehabilitation does not provide direction to
developers for many of the decisions they have to make. From where
are the resources needed for these decisions to come?
Ward et al. (2006) proposed a meta-theoretical framework for
understanding different levels of etiological theory for sexual
offending. In this framework, Level I theories are global and multi-
factorial, and leave unspecified significant details about the inner
workings of the phenomenon they seek to explain. It is Level II
theories that specify the mechanisms of each factor and how they
interact with other factors, and there is a third level; which, can be
thought of as a local theory of the offender's most recent offence.
Although rehabilitation-related theories are only partly informed by
etiology—they also cover change processes practice guidelines, and
guiding values (Ward et al., 2007)—they have in common with the
Level I theories in this meta-theoretical framework that they also are
global and multi-factorial. Ward et al. (2007), in outlining the
conceptual resources needed for fully theory-guided rehabilitation
programs, proposed a second level they labeled “treatment or therapy
theories,” which they described as “local theories of change.” However,
this level may be better viewed as two levels of theory: at the bottom
level is the local theory for a particular program; outlining why this
program takes the people it does, why the components are ordered as
they are to effect change, why some components are included and
others are not, how content and process interact to maximize offender
engagement and learning in this program, and so on. Missing, in this
view, is a middle-level theory that provides the conceptual resources
not just for working therapeutically with offenders, but for doing so in
that general kind of program and with this particular clientele. For
example, following from the framework proposed in this paper, a
middle or intermediate-level theory—as suggested earlier—could be
constructed for high-level multi-factorial forensic therapy programs.
Another example is readily available in the sex offender arena, where
quite a number of books now outline—in detail rich enough to inform
the design of a specific program—how to rehabilitate sexual offenders
regardless of specific setting, and across a range of risk levels.
Imagine that instead of rehabilitating sexual offenders, a program
sets out to work with high-risk violent offenders, or PCL-psychopaths,
or moderate-risk offenders with major drug and alcohol addictions,
for example. In doing so, it quickly becomes apparent that these
middle-level resources are impoverished or simply absent; the
conceptual space between the rehabilitation theory and the program's
manuals is incompletely (or completely un)specified. For example, it
is not unusual for facilitators' manuals to be limited to a series of
chapters on source-therapeutic approaches used in the program (e.g.,
Beck's cognitive restructuring, motivational interviewing, mindful-
ness-based therapy, and dialectical behavior therapy), without an
explanation of which aspects of these therapies are effective in which
situations or modules, how to resolve any conflicts in their assump-
tions about their processes of change, and how they are actually to be
integrated with etiological theory.
An explicit aim of articulating this framework was to make a
preliminary step in carving up the theoretical space into smaller, more
manageable pieces, with the hope of stimulating more intermediate-
level theory development and dissemination for each level of
program.
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
The bases for parts of the framework are drawn from clinical
experience and are therefore speculative more than empirically-
based. One example of this more speculative material has to do with
what offenders need at different levels in order to learn and use what
is learned. Moderate and high-risk offenders often fail (i.e., recidivate)
after program involvement. How offenders fail—and succeed—has
been the subject of some research (Maruna, 2001; Zamble & Quinsey,
1997) but little is yet known about the connection between what
offenders learned in programs, how they used the experience after the
program, and what effect it had on the next part of their lives. Even
basic research identifying what participants gained from programs is
scarce (e.g., Marques et al., 2005); more would be very useful in
fleshing out this aspect of the framework.
Conducting an exercise like the one that formed the basis for the
proposed framework reveals that there is still much about “what
works for whom” that simply hasn't yet been investigated; it exposes
a number of gaps worthy of future research, gaps that have not just
been theoretical, but policy implications. Several examples are given
below.
One important question is whether the total treatment hours
serves as an adequate indicator of the dosage of a program, or whether
the way in which those hours are delivered is also important. In fact,
in the high-level programs total hours in group sessions becomes
somewhat meaningless, since therapy is regarded as an around-the-
clock endeavor. However, comparing outcomes for high-level and
mid-level programs that have the same number of in-group hours is
an obvious goal for future research. Currently, there are almost no
evaluations establishing the outcomes of the high-level comprehen-
sive programs. Are they more effective than mid-level programs?
Future research could also include a more systematic examination of
current differences in mid-level programs. It has been assumed here—
based on clinical experience with mid-level violence prevention
programs—that interventions around 300 h duration, for offenders at
moderate to high risk of serious offending, require skilled staff with
university-level training in psychology, and experience and training in
providing cognitive–behavioral therapy. The optimal approach outlined
in mid-level programs in the framework is predicated on manualization
of both therapeutic process and content (Mann, 2009; Marshall &
Burton, 2009) allowing therapists some room to use clinical skills to
tailor treatment to clients' interests and needs, and to respond to issues
as they arise in group, but providing sufficient supervision and
oversight to maintain the quality of the intervention.
However, currently some programs at this level use facilitators
who are untrained or minimally trained in psychology, and may have
little or no therapy background. Instead they have on-the-job training
in how to run a specific program, from a highly detailed manual, with
close invigilation of adherence with manual content to ensure high
quality treatment delivery. This “paint-by-numbers” approach to
rehabilitation has been justified as necessary when resources are
scarce, and in order to deliver rehabilitation to larger populations.
Does it work? Does it work as well as or better than programs led by
more thoroughly trained staff who have more leeway to balance
content and process? Is it more cost-effective? Head to head tests of
this issue are essential. Currently, skepticism about the paint-by-
numbers approach is fed partly by meta-analyses showing that
demonstration programs are more effective than “real-world” roll-
outs (Andrews & Bonta, 2006). But real-world dissemination often
confounds quality control problems with changes in: the quality of
staffing, approach to manualization vs. the role of clinical judgment,
and how “treatment integrity” is consequently achieved. There is
tentative evidence that correctional officers can be effective change
agents with more serious offenders (Henning & Frueh, 1996), and
elsewhere it has sometimes been found that clinical psychologists—
admittedly not adhering to research evidence on the most effective
treatment approaches—can be less effective than much less expensive
and qualified change facilitators trained to achieve a specific task
33
(Baker, McFall, & Shoham, 2008). Nevertheless, the “demonstration
program effect” implies that there is a trade-off between program
quality and quantity; replicating a demonstration program across
many sites is associated with a decrease in program integrity and
dilutes the program's effect size (Andrews and Bonta). Whether such
dilution is caused by reductions in the quality of the therapeutic
alliance, lower therapist skill or other factors is not yet clear.
What if less expensive, relatively less-qualified staff can run
effective basic and even mid-level programs? We can then turn our
focus to how best to achieve good effect sizes with the very difficult-
to-treat highest risk clients. One implication of the framework is that
if these clients require a very high level of therapeutic skill and
resources, the most expensive and highly trained therapists should be
working with these most destructive clients, rather than solely
overseeing the delivery of lower level programs; and similarly highly
skilled supervisors and trainers should be supporting their practice.
This discussion brings us back to the issue of offender risk.
Although challenging, it would be fruitful to explore ways of defining
levels of risk more absolutely in order to advance our understanding
of the relationship between criminal risk and treatability, at different
program levels. This is not an easy task. One way to define risk is in
terms of the survival time or proportion of the sample reconvicted—or
reconvicted for serious offences—within a fixed period of time. But
recidivism rates are tied to jurisdiction-specific legislation and
policies relating to the use of imprisonment, lengths of imprisonment,
and eligibility for parole. Even with these differences and difficulties
with comparing across jurisdictions, it appears that much of our
research on program effectiveness is based on moderate-risk
offenders (Gendreau, personal communication, June, 2007). More
research on high to very high-risk offenders is needed to promote the
development of intermediate-level theory for the mid-level and high-
level programs. This research is also needed to test a key assumption
of this framework: that treatment effects will attenuate at the highest
levels of risk unless not only dosage, but the power of the intervention
to engender change in each specific offender also increases.
The proposed framework outlined above mixes a survey of recent
and current practice with observations about their implications, and
ideas for future research. It is intended to stimulate discussion and
future theory and research, but the underlying enterprise has
important limitations. Summarizing the vast array of cognitive–
behavioral interventions for offenders into just 3 levels defined by
common principles requires, no doubt, that a variety of exceptions go
unnoted. Furthermore, rehabilitative programs are constantly moving
targets; while taking a snapshot for this review, it was evident that
several correctional systems were planning to implement significant
innovations in their rehabilitation design and delivery in the not-too-
distant future. And no significant attempt was made here to review
the scientific validity of the specific program examples discussed, nor
of most of the many program characteristics outlined in the levels of
the framework. Future research should examine these issues with the
purpose of being able to draw conclusions about how each level of
program could best be designed and delivered, within its constraints.
To conclude, in many western jurisdictions that formerly sup-
ported a more humane and effective approach to improving
community safety, harsh and lengthy imprisonment regimes are
gaining a stronger hold, even if crime rates are actually reducing
(Prisoners of politics: Canada's criminal-justice policy, 2009). These
developments threaten rehabilitation. What resources remain for
interventions may result in programs that are so poorly implemented
that they are doomed to confirm existing prejudices. However, a
review like this one demonstrates just how much progress we have
made, since the days in the 1970s when “it is fair to say that … other
scholars concurred with the assessment that treatment programs
were generally ineffective” (p. 420, Gendreau, Smith, & French, 2006).
More than at any time in the past, there are rehabilitation programs
for offenders that are broadly guided by scientific research and theory.
34 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
And there is diversity and depth to these offerings; they are far from
being one-size-fits-all.
Acknowledgements
I would like to acknowledge the contributions of Ralph Serin, and
two anonymous reviewers for the final version of this manuscript.
References
Andrews, D. A., & Bonta, J. (2006). The psychology of criminal conduct (4th edition ed.).
Cincinnati OH: Anderson.
Andrews, D. A., Bonta, J., & Hoge, R. D. (1990). Classification for effective rehabilitation:
Rediscovering psychology. Criminal Justice and Behavior, 17, 19−52.
Annison, J. (2006). Style over substance? A review of the evidence base for the use of
learning styles in probation. Criminology and Criminal Justice, 6, 239−257.
Anstiss, B., Polaschek, D. L. L., & Wilson, M. J. (in press). A brief motivational
interviewing intervention with prisoners: When you lead a horse to water, can it
drink for itself? Psychology, Crime & Law.
Antonowicz, D. H. (2005). The Reasoning and Rehabilitation program: Outcome
evaluations with offenders. In M., & J. (Eds.), Social problem-solving and offenders
(pp. 163−181). Chichester, UK: Wiley.
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based adult corrections programs: What
works and what does not. Olympia: Washington State Institute for Public Policy.
Baker, T. B., McFall, R. M., & Shoham, V. (2008). Current status and future prospects of
clinical psychology: Toward a scientifically principled approach to mental and
behavioral health care. Psychological Science in the Public Interest, 9, 67−103.
Beech, A. R., Craig, L. A., & Browne, K. D. (Eds.). (2009). Assessment and treatment of sex
offenders: A handbook. Chichester: Wiley.
Bennett-Levy, J., Westbrook, D., Fennell, M., Cooper, M., Rouf, K., & Hackmann, A.
(2004). Behavioural experiments: historical and conceptual underpinnings. In J., G.,
M., A., M., & D. (Eds.), Oxford guide to behavioural experiments in cognitive therapy
(pp. 1−20). Oxford, UK: Oxford University Press.
Bonta, J., & Andrews, D. A. (2007). Risk–Need–Responsivity model for offender assessment
and rehabilitation (No. 2007-04). Ottawa, Canada: Department of Public Safety and
Emergency Preparedness Canada.
Bourgon, G., & Armstrong, B. (2005). Transferring the principles of effective treatment
into a “real world” setting. Criminal Justice and Behavior, 32, 3−25.
Bourgon, G., Bonta, J., Rugge, T., Scott, T., & Yessine, A. K. (2009). Translating “what
works” into sustainable everyday practice. Program design, implementation and
evaluation. : Public Safety Canada.
Brown, T. L., Borduin, C. M., & Henggeler, S. W. (2001). Treating juvenile offenders in
community settings. In J. B., B. D., & W. H. (Eds.), Treating adult and juvenile offenders
with special needs (pp. 445−464). Washington, DC: American Psychological
Association.
Bush, J. (2001). Vermont's cognitive self-change program: A 15-year review. Corrections
Today Sunday July 1. Retrieved Jan 19 2010 from http://www.allbusiness.com/
public-administration/justice-public-order/1010235-1.html.
Conway, M., & Ross, M. (1984). Getting what you want by revising what you had.
Journal of Personality and Social Psychology, 47, 738−748.
Cooke, D. J., Michie, C., Hart, S. D., & Hare, R. D. (1999). Evaluating the screening version
of the Hare Psychopathy Checklist-Revised (PCL:SV): An item response theory
analysis. Psychological Assessment, 11, 3−13.
Correctional Service of Canada (2000, Junee). Violence Prevention Program Accredi-
tation case file. Unpublished report.
Craighead, W. E., Sheets, E. S., Brosse, A. L., & Ilardi, S. S. (2007). Psychosocial treatments
for major depressive disorder. In P. E., & J. M. (Eds.), A guide to treatments that work
(pp. 289−307). Oxford: OUP.
Daffern, M., Jones, L., Howells, K., Shine, J., Mikton, C., & Tunbridge, V. (2007). Editorial:
Refining the definition of offence paralleling behaviour. Criminal Behaviour and
Mental Health, 17, 265−273.
Day, K. (1988). A hospital-based treatment programme for male mentally handicapped
offenders. British Journal of Psychiatry, 153, 635−664.
Dowd, E. T. (2006). What changes in cognitive therapy? The role of tacit knowledge
structures. Journal of Cognitive and Behavioral Psychotherapies, 6, 141−148.
Dowden, C., & Andrews, D. A. (2004). The importance of staff practice in delivering
effective correctional treatment: A meta-analytic review of core correctional
practice. International Journal of Offender Therapy and Comparative Criminology, 48,
203−214.
Dowden, C., Blanchette, K., & Serin, R. (1999). Anger management programming for
federal male inmates: An effective intervention (Research report R-82). Ottawa, ON:
Correctional Service of Canada.
DSPD programme: Dangerous People with Severe Personality Disorder. Retrieved from
www.dspdprogram.gov.uk/ on March 9, 2010.
Farrington, D. P., Coid, J. W., Harnett, L., Jolliffe, D., Soteriou, N., Turner, R., et al. (2006).
Criminal careers and life success: New findings from the Cambridge Study in
Delinquent Development. London: Home Office.
Fernandez, Y. M., Shingler, J., & Marshall, W. L. (2006). Putting “behavior” back into
cognitive–behavioral treatment of sexual offenders. In W. L., Y. M., L. E., & G. A.
(Eds.), Sexual offender treatment: Controversial issues (pp. 211−224). New York:
Wiley.
Gendreau, P., Goggin, C., Cullen, F. T., & Paparozzi, M. (2002). The common-sense
revolution and correctional policy. In J. (Ed.), Offender rehabilitation and treatment:
Effective programs and policies to reduce re-offending (pp. 359−386). Chichester:
Wiley.
Gendreau, P., Smith, P., & French, S. (2006). The theory of effective correctional
intervention: Empirical status and future directions. In F., J., & M. (Eds.), Taking
stock: The status of criminology theory (pp. 419−446). New Jersey: Transaction
Press.
Goggin, C., & Gendreau, P. (2006). The implementation and maintenance of quality
services in offender rehabilitation programs. In C. R., & E. J. (Eds.), Offending
behaviour programmes: Development, application, and controversies (pp. 209−246).
Chichester: Wiley.
Goldstein, A. P., & Huff, C. R. (1993). The gang intervention handbook. Champaign, IL:
Research Press.
Guerra, N. G., Tolan, P. H., & Hammond, W. R. (1994). Prevention and treatment of
adolescent violence. In L. D., J. H., & P. (Eds.), Reason to hope: A psychosocial
perspective on violence and youth (pp. 383−403). Washington, DC: American
Psychological Association.
Hanson, R. K., Bourgon, G., Helmus, L., & Hodgson, S. (2009). The principles of effective
correctional treatment also apply to sexual offenders: A meta-analysis. Criminal
Justice and Behavior, 36, 865−891.
Hanson, R. K., & Morton-Bourgon, K. (2004). Predictors of sexual recidivism: An updated
meta-analysis (User Report 2004-02). Ottawa: Public Safety & Emergency
Preparedness Canada.
Hanson, R. K., & Morton-Bourgon, K. (2005). The characteristics of persistent sexual
offenders: A meta-analysis of recidivism studies. Journal of Consulting and Clinical
Psychology, 73, 1154−1163.
Hart, S., Cox, D., & Hare, R. (1995). Manual for the Psychopathy Checklist: Screening
Version(PCL:SV). Toronto, Canada: Multi-Health Systems.
Havighurst, R. J. (1952). Human development and education. New York: Longmans, Green.
Henning, K. R., & Frueh, B. C. (1996). Cognitive–behavioral treatment of incarcerated
offenders: An evaluation of the Vermont Department of Corrections' cognitive self-
change program. Criminal Justice & Behavior, 23, 523−541.
Hogue, T. E., Jones, L., Talkes, K., & Tennant, A. (2007). The Peaks: A clinical service for those
with dangerous and severe personality disorder. Psychology, Crime & Law, 13, 57−68.
Hollin, C. R. (2006). Offending behaviour programs and contention: Evidence-based
practice, manuals, and program evaluation. In C. R., & E. J. (Eds.), Offending
behaviour programmes: Development, application, and controversies (pp. 33−67).
Chichester, UK: Wiley.
Hollin, C. R. (2009). Treatment manuals: The good, the bad, and the useful. Journal of
Sexual Aggression, 15, 133−137.
Hollin, C. R., McGuire, J., Palmer, E. J., Bilby, C., Hatcher, R., & Holmes, A. (2002).
Introducing Pathfinder programs into the Probation Service: An interim report (No.
Home Office Research Study 247). London: Home Office.
Hollin, C. R., & Palmer, E. J. (2003). Level of Service Inventory-Revised profiles of violent
and nonviolent prisoners. Journal of Interpersonal Violence, 18, 1075−1086.
Hollin, C. R., & Palmer, E. J. (Eds.). (2006). Offending behaviour programmes:
Development, application, and controversies. Chichester, UK: Wiley.
Holmqvist, R., Hill, T., & Lang, A. (2009). Effects of aggression replacement training in
young offender institutions. International Journal of Offender Therapy and
Comparative Criminology, 53, 74−92.
Howells, K., & Day, A. (2002). Grasping the nettle: Treating and rehabilitating the
violent offender. Australian Psychologist, 37, 222−228.
Howells, K., & Day, A. (2003). Readiness for anger management: Clinical and theoretical
issues. Clinical Psychology Review, 23, 319−337.
Howells, K., & Day, A. (2007). Readiness for treatment in high risk offenders with
personality disorders. Psychology, Crime & Law, 13, 47−56.
Howells, K., Day, A., & Thomas-Peter, B. (2004). Changing violent behaviour: Forensic
mental health and criminological models compared. The Journal of Forensic
Psychiatry and Psychology, 15, 391−404.
Howells, K., Day, A., Williamson, P., Bubner, S., Jauncey, S., Parker, A., et al. (2005). Brief
anger management programs with offenders: Outcomes and predictors of change.
Journal of Forensic Psychiatry and Psychology, 16, 296−311.
Johansson, P., & Høglend, P. (2007). Identifying mechanisms of change in psychother-
apy: Mediators of treatment outcome. Clinical Psychology & Psychotherapy, 14, 1−9.
Kirsch, L. G., & Becker, J. V. (2006). Sexual offending: Theory of problem, theory of
change, and implications for treatment effectiveness. Aggression and Violent
Behavior, 11, 208−224.
Laws, D. R. (2002). Owning your own data: The management of denial. In M. McMurran
(Ed.), Motivating offenders to change: A guide to enhancing engagement in therapy
(pp. 173−192). Chichester, UK: Wiley.
Lösel, F., & Schmucker, M. (2005). The effectiveness of treatment for sexual offenders: A
comprehensive meta-analysis. Journal of Experimental Criminology, 1, 117−146.
Loza, W., Loza-Fanous, A., & Heseltine, K. (2007). The myth of offenders' deception on
self-report measure predicting recidivism: Example from the Self-Appraisal
Questionnaire. Journal of Interpersonal Violence, 23, 671−683.
Maden, A., Williams, J., Wong, S. C. P., & Leis, T. A. (2004). Treating dangerous and severe
personality disorder in high security: Lessons from the Regional Psychiatric Centre,
Saskatoon, Canada. The Journal of Forensic Psychiatry and Psychology, 15, 375−390.
Mann, R. E. (2009). Sex offender treatment: The case for manualization. Journal of
Sexual Aggression, 15, 121−131.
Marques, J. K., Wiederanders, M., Day, D. M., Nelson, C., & van Ommeren, A. (2005).
Effects of a relapse prevention program on sexual recidivism: Final results from
California's sex offender treatment and evaluation project (SOTEP). Sexual Abuse: A
Journal of Research and Treatment, 17, 79−107.
Marshall, W. L. (1996). Assessment, treatment, and theorizing about sex offenders:
Developments during the past twenty years and future directions. Criminal Justice
and Behavior, 23, 162−199.
 D.L.L. Polaschek / Aggression and Violent Behavior 16 (2011) 20–35
Marshall, W. L. (2005). Therapist style in sexual offender treatment: Influences on
indices of change. Sexual Abuse: A Journal of Research and Treatment, 17, 109−116.
Marshall, W. L. (2009). Manualization: A blessing or a curse? Journal of Sexual
Aggression, 15, 109−120.
Marshall, W. L. (in press). The treatment of sexual offenders: A positive approach.
Washington, DC: American Psychological Association.
Marshall, W. L., Anderson, D., & Fernandez, Y. M. (1999). Cognitive behavioural treatment
of sexual offenders. Chichester, UK: Wiley.
Marshall, W. L., & Burton, D. L. (2009). The importance of group processes in offender
treatment. Aggression and Violent Behavior, doi:10.1016/ j.avb.2009.08.008.
Marshall, L. E., & Moulden, H. M. (2006). Preparatory programs for sexual offenders. In
W. L. Marshall, Y. M. Fernandez, L. E. Marshall, & G. E. Serran (Eds.), Sexual offender
treatment: Controversial issues (pp. 199−210). Chichester: Wiley.
Maruna, S. (2001). Making good: How ex-convicts reform and rebuild their lives.
Washington, DC: American Psychological Association.
McDougall, C., Clarbour, J., Perry, A. E., & Bowles, R. (2009). Evaluation of HM Prison
Service Enhanced Thinking Skills Program: Report on the implementation of a
randomised controlled trial. London: Ministry of Justice Research Series 4/09.
Downloaded February 21 2010 from http://www.justice.gov.uk/publications/
enhanced-thinking-skills-programme.htm.
McGuire, J. (2002). Integrating findings from research reviews. In J. McGuire (Ed.),
Offender rehabilitation and treatment: Effective programs and policies to reduce re-
offending (pp. 3−38). Chichester, UK: Wiley.
McGuire, J. (2005). Social problem solving: Basic concepts, research, and applications.
In M. McMurran, & J. McGuire (Eds.), Social problem-solving and offenders
(pp. 3−29). Chichester, UK: Wiley.
McGuire, J. (2006). General offending behaviour programs: Concept, theory, and
practice. In C. R. Hollin, & E. J. Palmer (Eds.), Offending behaviour programs:
Development, application, and controversies (pp. 69−111). Chichester, UK: Wiley.
McMurran, M. (2009). Motivational interviewing with offenders: A systematic review.
Legal and Criminological Psychology, 14, 83−100.
Mills, J. F., & Kroner, D. G. (2006). Impression management and self-report among
violent offenders. Journal of Interpersonal Violence, 21, 178−192.
Nadesu, A. (2007). Reconviction patterns of released prisoners: A 36-months follow-up
analysis.Department of Corrections Policy Development unpublished report.
New Zealand Department of Corrections (n.d.). Programs for offenders. Retrieved
March 10 2010, from http://www.corrections.govt.nz/about-us/fact-sheets/man-
aging-offenders/programs-for-offenders.html.
Novaco, R. W. (1977). Stress inoculation: A cognitive therapy for anger and its
application to a case of depression. Journal of Consulting and Clinical Psychology, 45,
600−608.
Olver, M. E., Wong, S. C. P., Nicholaichuk, T., & Gordon, A. (2007). The validity and
reliability of the Violence Risk Scale—Sexual Offender Version: Assessing sex
offender risk and evaluating therapeutic change. Psychological Assessment, 19,
318−329.
Palmer, E. J., McGuire, J., Hatcher, R. M., Hounsome, J. C., Bilby, C. A., & Hollin, C. R.
(2009). Allocation to offending behavior programs in the English and Welsh
Probation Service. Criminal Justice and Behavior, 36, 909−922.
Polaschek, D. L. L. (2006a). National Violence Prevention Unit Evaluation Report III:
Psychometric and recidivism analyses for the retrospective evaluation sample.
Wellington, New Zealand: Department of Corrections unpublished report.
Polaschek, D. L. L. (2006b). Violent offender programess: Concept, theory and practice.
In C. R. Hollin, & E. J. Palmer (Eds.), Offending behaviour programmes: Development,
application, and controversies (pp. 113−154). Chichester, UK: Wiley.
Polaschek, D. L. L. (2010). A treatment theory for high-risk violent offenders.
Manuscript in preparation.
Polaschek, D. L. L., Bell, R. K., Calvert, S. W., & Takarangi, M. K. T. (2010). Cognitive–
behavioural rehabilitation of high-risk violent offenders: investigating treatment
change with explicit and implicit measures. Applied Cognitive Psychology, 24,
437−449.
Polaschek, D. L. L., & Collie, R. M. (2004). Rehabilitating serious violent adult offenders:
An empirical and theoretical stocktake. Psychology, Crime and Law, 10, 321−334.
Polaschek, D. L. L., & Ross, E. C. (2010). Do early therapeutic alliance, motivation, and
change readiness predict therapy outcomes for high risk violent prisoners?
Criminal Behaviour and Mental Health, 20, 100−111.
35
Prisoners of politics: Canada's criminal-justice policy (2009, December 19). The
Economist. Retrieved 20 January 2010 from. http://www.economist.com/displayS-
tory.cfm?story_id=15127510.
Ross, R. R., Fabiano, E. A., & Ewles, C. D. (1988). Reasoning and rehabilitation.
International Journal of Offender Therapy and Comparative Criminology, 32, 29−36.
Ross, E. C., Polaschek, D. L. L., & Ward, T. (2008). The therapeutic alliance: A theoretical
revision for offender rehabilitation. Aggression and Violent Behavior, 13, 462−480.
Serin, R. C., Lloyd, C. D., Helmus, L., Derkzen, D., & Luong, D. (2010). Assessment and
conceptual issues in understanding offender change. Canada: National Institute of
Corrections unpublished report.
Serin, R. C., & Mailloux, D. L. (2003). Assessment of sex offenders: Lessons learned from
the assessment of non-sex offenders. In R. Prentky, & M. Seto (Eds.), Understanding
sexually coercive behavior (pp. 185−197). New York: Academy of Sciences.
Serin, R. C., Kennedy, S. M., Mailloux, D. L., & Hanby, L. J. (in press). The origins of
treatment readiness. In A. Day, C. S., T. Ward, K. Howells & J. Vess (Eds.), Transition
to better lives: Offender readiness and rehabilitation. Cullompton, UK: Willam.
Serran, G. A., Fernandez, Y., Marshall, W. L., & Mann, R. E. (2003). Process issues in
treatment: Application to sexual offender programs. Professional Psychology:
Research and Practice, 34(4), 368−374.
Skeem, J. L., Eno Louden, J., Polaschek, D. L. L., & Camp, J. (2007). Assessing relationship
quality in mandated community treatment: Blending care with control. Psycho-
logical Assessment, 19, 397−410.
Skeem, J. L., Polaschek, D. L. L., & Manchak, S. (2009). Appropriate treatment works, but
how? Rehabilitating general, psychopathic, and high risk offenders. In J. L. Skeem, K.
Douglas, & S. Lilienfeld (Eds.), Psychological science in the courtroom: Controversies
and consensus (pp. 358−384). Washington, DC: American Psychological
Association.
Smith, P., Gendreau, P., & Swartz, K. (2009). Validating the principles of effective
intervention: A systematic review of the contributions of meta-analysis in the field
of corrections. Victims and Offenders, 4, 148−169.
Toderov, A., & Bargh, J. A. (2002). Automatic sources of aggression. Aggression and
Violent Behavior, 7, 53−68.
Vermont Department of Corrections (2002). A manual for delivery of cognitive self
change. Downloaded on 12 March 2010 from. http://www.doc.state.vt.us/pro-
grams/violent/cscmanual.pdf October.
Violence Prevention Program (n. d.). Retrieved 4 March 2010 from http://www.csc-scc.
gc.ca/text/prgrm/vp-eng.shtml.
Ward, T., Day, A., Howells, K., & Birgden, A. (2004). The multifactor offender readiness
model. Aggression and Violent Behavior, 9, 645−673.
Ward, T., & Maruna, S. (2007). Rehabilitation. London: Routledge.
Ward, T., Melser, J., & Yates, P. M. (2007). Reconstructing the risk, need, responsivity
model: A theoretical elaboration and evaluation. Aggression and Violent Behavior,
12, 208−228.
Ward, T., Polaschek, D. L. L., & Beech, A. R. (2006). Theories of sexual offending.
Chichester: Wiley.
Ward, T., & Stewart, C. A. (2003). The treatment of sex offenders: Risk management and
good lives. Professional Psychology: Research and Practice, 34, 353−360.
Wilson, T. D., Lindsey, S., & Schooler, T. Y. (2000). A model of dual attitudes.
Psychological Review, 107, 101−126.
Wong, S., & Gordon, A. (2000). Violence Risk Scale. Canada: University of Saskatchewan.
Wong, S., Gordon, A., & Gu, D. (2007). Assessment and treatment of violence-prone
forensic clients: An integrated approach. British Journal of Psychiatry, 190(suppl.
49), s66−s74.
Wong, S. C. P., Vander Veen, S., Leis, T. A., Gu, D., Liber, E. U., & Middleton, H. L. (2005).
Reintegrating serious violent and personality-disordered offenders from a super-
maximum security institution into the general offender population. International
Journal of Offender Therapy and Comparative Criminology, 49, 362−375.
Wormith, J. S., Althouse, R., Simpson, M., Reitzel, L. R., Fagan, T. J., & Morgan, R. D.
(2007). The rehabilitation and reintegration of offenders: The current landscape
and some future directions for correctional psychology. Criminal Justice and
Behavior, 34, 879−892.
Yochelson, S., & Samenow, S. (1977). The criminal personality, vol 2: The change process.
New York: Jason Aronsen.
Zamble, E., & Quinsey, V. L. (1997). The criminal recidivism process. Cambridge, UK: CUP.