Effectiveness of Correctional Programs With Ethnically Diverse Offenders

Abstract

Numerous studies have examined the effects of cognitive-behavioural therapy (CBT) on criminal recidivism, and several meta-analyses have confirmed the overall effectiveness of this approach. Few studies, however, have examined the efficacy of these programs specifically with adult offenders from diverse ethnic backgrounds. The present research uses meta-analytic techniques to examine the outcomes for Canadian federal offenders participating in correctional programs according to self-identified ethnic group (Caucasian, Aboriginal, Black, and Other). Correctional programs within the Correctional Service of Canada adhere to the Risk, Need, Responsivity principles outlined in the effective correctional literature. Within-group analyses compared offenders from the same ethnic background who participated in correctional programs with a nontreatment comparison group. Odds ratios ranged from 1.36 to 1.76, indicating significant reductions in recidivism for offenders participating in correctional programs, regardless of ethnic status. Furthermore, the difference in effect size magnitude between ethnic groups was nonsignificant suggesting offenders from a wide variety of ethnic backgrounds can benefit from correctional programs rigorously developed and implemented using a CBT framework.

Keywords

correctional program evaluation ethnically diverse offenders cognitive-behavioural treatments recidivism RNR principles

Incarcerated populations across many jurisdictions in North America, Western Europe, and Australasia are a heterogeneous group representing offenders from varying ethnic and racial backgrounds. Canada’s federal offender population increasingly reflects a broad range of ethnic and cultural groups. According to the Department of Public Safety Canada (2007, 2010), the proportion of federal offenders who do not identify as Caucasian has been increasing steadily over the past decade. Currently, 65.6% of Canada’s federal offender population identify as Caucasian, 17.9% Aboriginal, 7.9% Black, 4.9% Asian, 0.9% Hispanic, and 2.8% Other (Public Safety Canada, 2010). Regardless of the ethnic composition of their populations, correctional agencies have a responsibility to provide effective rehabilitation for all offenders under their care.

Although sociologists make an important distinction between race and cultural or ethnic associations, the basis for researchers classifying individuals as members of a given race or cultural group is often unclear. Race generally refers to a human population that is believed to be distinct based on physical differences such as skin color or facial characteristics, whereas ethnicity refers to groups that share social traits such as tribal affiliations, culture, and traditions (Bhopal, 2004). In Canada, the term visible minority is often used to describe individuals of non-Caucasian ethnic backgrounds; minority being used in the statistical sense (Aspinall, 2002). In this article, we use the term ethnicity as it allows for a broader inclusion of cultural identification.

There has been limited research on the effectiveness of correctional programs for different ethnic groups and, historically, few studies have disaggregated outcomes by ethnicity. Although more recent Canadian research has examined specific treatment outcomes for Aboriginal offenders, the effectiveness of correctional programs with a variety of different ethnic groups has not been explored.

In Canada, adult offenders serving sentences of 2 years or more are under the jurisdiction of the Correctional Service of Canada (CSC), a federal department legislated with the provision of rehabilitation and transitioning services for all offenders who require them across the country (CSC, 2003). CSC has a mandate to “provide a range of programs designed to address the needs of offenders and contribute to their successful reintegration into the community” (Corrections and Conditional Release Act [CCRA], 1992, para. 76). Nationally recognised correctional programs within CSC are based on a cognitive-behavioural therapy (CBT) model of intervention in which programming aims to address maladaptive cognitions and behaviours, while emphasizing training on skills for prosocial living. CSC correctional programs are designed to address a range of criminogenic needs and to adhere to the principles of effective correctional treatment (i.e., Risk, Need, Responsivity [RNR]; see Andrews & Bonta, 2010). Meta-analyses and reviews have shown that correctional programs that incorporate these principles into program design and implementation deliver significant treatment effects beyond those found for correctional programs that do not incorporate these principles (Andrews & Dowden, 2006; Gendreau, Little, & Goggin, 1996; Hanson, Bourgon, Helmus, & Hodgson, 2009; Landenberger & Lipsey, 2005; Lowenkamp, Latessa, & Holsinger, 2006; MacKenzie, 2006).

Most correctional programs in CSC have been accredited by an international panel of experts against criteria related to evidence-based practice identified in the effective corrections literature. That is, all programs are reserved for higher risk offenders as measured by static and dynamic risk indicators, and the highest risk offenders are referred to the highest level of intensity (Risk principle); all programs target dynamic needs related to their criminal history (Need principle: for example, substance abuse is targeted for offenders who have moderate to high need in this area and their offence history involves substance abuse; violent offender programs target factors empirically related to aggression, such as hostile attributions and impulsivity; sex offender programs target sexual deviance and cognitive distortions, etc.); programs are designed, and facilitators are trained on techniques to maximise offender engagement and understanding (Responsivity principle: for example, facilitators are trained in the application of motivational interviewing; treatment goals are set collaboratively; and individual sessions are included in all programs). Recent revisions to the programs have included a relapse prevention framework requiring all participants to personalise a plan to address their specific criminal risk factors. All programs are skills based allowing time for practice and overlearning. All program facilitators are specialists in program delivery and are trained on each specific program they deliver and are provided with local and regional oversight to ensure program integrity. Aboriginal-specific correctional programs were developed in collaboration with Aboriginal Elders and experts and are primarily delivered by Aboriginal facilitators. These programs also adhere to the RNR principles and apply CBT techniques, but supplement the approach with culturally specific teachings, scenarios, and ceremonies and include involvement with Aboriginal Elders.

CBT is widely viewed as the most effective form of treatment for offenders and has repeatedly demonstrated the largest effect sizes for reductions in recidivism (see, for example, Landenberger & Lipsey, 2005; Losel & Schmucker, 2005; Pearson, Lipton, Cleland, & Yee, 2002; D. B. Wilson, Bouffard, & Mackenzie, 2005). Despite this growing consensus, the traditional CBT approach has been criticized for overlooking specific aspects of participants’ culture and ethnic background, and some advocates from ethnic communities have called for the development of culturally specific correctional programs (e.g., Hwang, 2006; King, 2011). CBT programs, particularly those that are manualised, have been criticized on the basis of their inappropriateness of vocabulary, examples, and scenarios (Holleran Steiker, 2008); choice of skills that may not be applicable or adapted to the cultural role expectation of participants (Lin, 2001); and perhaps particularly relevant to offender groups, failure to take into consideration unique pressure and circumstances some ethnic groups face in their daily lives (McCormick, 2000). These programs are also criticized for not taking advantage of the power of evoking the offenders’ cultural roots through appealing to symbols, metaphors, and practices and religious affiliations that hold meaning for many participants (King, 2011). Some Aboriginal critics of CBT object to the extent to which the individual is forced to adopt an offender label and made responsible for the thinking, feeling, behaviour link that ignores the social and historical context for why dysfunctional behaviours developed in the first place (Kendall, 2002). Others, however, argue that CBT is flexible and can be adapted to a wide range of cultural groups with appropriate training of facilitators and consideration of culture-sensitive material and approaches (Hays, 2006).

The literature on therapeutic effectiveness in general often lacks a focus on the impact of interventions based on participants’ cultural background (Nagayama Hall, 2001) and there is concern that evidence-based treatments may not generalize to minority populations (Miranda et al., 2005). While attending to responsivity issues related to cognitive style, cultural relevancy, and learning deficits is one of the key principles of effective correctional treatment (Andrews, Bonta, & Hoge, 1990), research conducted on CBT program effectiveness often includes a primarily Caucasian sample (Miranda et al., 2005). Differential outcomes for minority clients are rarely examined, or, if examined, small sample sizes reduce statistical power (Huey & Polo, 2008). Recidivism rates may also be compared with those of the entire sample instead of the offenders from their ethnic group. For Aboriginal offenders in Canada, this practice can neutralize treatment effects because their base rates for recidivism rates are higher than those of the non-Aboriginal population (Usher & Stewart, 2010).

Although specific research on recidivism reductions resulting from CBT with non-Caucasian adult offenders is sparse, the literature on the effectiveness of CBT with minority young offenders and, more generally, CBT with ethnically diverse samples can be drawn upon. A large-scale meta-analysis examining intervention programs for juvenile delinquents reports positive gains for minority youth and Caucasian youth (S. J. Wilson, Lipsey, & Soydan, 2003). The authors computed mean effect sizes from 141 studies with predominantly minority samples and 164 studies with predominantly Caucasian samples. The minority samples generally consisted of African American and Hispanic youth. Both the minority and Caucasian youth showed positive treatment gains, with no significant difference in overall effect size between the two groups; however, the proportion of programs delivered in a CBT format is unclear, as education and employment programs were also included in this meta-analysis. As such, the results should be interpreted with caution. The results do, however, provide preliminary evidence that correctional programs for minority offenders are effective.

Studies using African American and Hispanic samples have generally found significant and meaningful treatment gains when using CBT to treat a variety of psychological disorders such as depression and anxiety. A number of well-designed studies with large African American, Latino, and Puerto Rican samples and random assignment conditions all report that individuals assigned to the CBT condition demonstrated greater improvements over other treatment modalities (e.g., Miranda et al., 2003; Rossello, Bernal, & Rivera-Medina, 2008).

Huey and Pogo (2008) examined the effectiveness of evidence-based programs (EBP) with minority youth. Several of these studies involved conduct disordered or criminal justice involved youth in programs that are cognitive-behaviourally based. They found that several of these programs produced improvements greater than nontreatment groups, and in some cases, better than treatment as usual options. When they examined culturally adapted EBPs, however, they were unable to establish that these programs produced better results than standard treatments.

Arroyo, Miller, and Tonigan (2003) conducted a study comparing a 12-step facilitation treatment for alcohol abuse, and CBT or motivational enhancement therapy (MET). Clients were randomly assigned to one of the three treatment options and followed for 12 months. Outcomes for non-Hispanic White and Hispanic clients were analyzed to determine the efficacy of the three differential treatments. The authors concluded that Hispanics did equally well as non-Hispanics on the CBT and MET treatment options. In another well-controlled study, Gil, Wagner, and Tubman (2004) studied the outcomes of participants in a controlled clinical trial that evaluated the effectiveness of a brief motivational, cognitive-behavioural intervention with multiethnic adolescents who had some juvenile justice system involvement and who had problems with alcohol and drug use. Participants were randomly assigned to the individual format of a guided self-change program, the family involved format of guided self-change, or a waiting list control condition. The results indicated that there were significant reductions in alcohol and marijuana use for all ethnic groups for both formats of the intervention compared with the controls.

Research has also demonstrated the effectiveness of CBT with Asian populations. CBT has been described as largely compatible with treatment approaches preferred by Chinese clients because of its directive and supportive orientation (Lin, 2001). For example, 30 elderly Chinese Americans were assigned to an eight-session CBT treatment group or a waitlist control. The CBT group showed significant improvements as measured by the Hamilton Depression Scale, whereas the control group showed no improvements (Dai et al., 1999). A pilot study examining CBT treatment for posttraumatic stress disorder with 12 Vietnamese refugees demonstrated significant improvements post-treatment in the CBT group compared with the waitlist controls (Hinton et al., 2004).

Few studies have examined the effectiveness of CBT with Aboriginal, Indigenous, or Native American populations. A study measuring the preference and suitability of different aspects of CBT indicates that certain components of CBT are appropriate for these ethnic groups. Eighty-two American Indian and European Americans (ethnic designations used by the study’s authors) were administered the CBT–Applicability Scale (CBT-AS). The CBT-AS is a self-report measure that rates preference for characteristics consistent with CBT, including active stance, focused in-session behaviour, and structured therapeutic relationship. Both groups indicated an equally high rating for the active stance domain of CBT, although the European American group indicated a stronger preference for a structured therapeutic relationship (Jackson, Wenzel, Schmutzer, & Tyler, 2006); however, caution should be taken when generalizing based on these findings. The effectiveness of CBT may also be mitigated by the cultural competence of the therapist and the client’s level of acculturation (Renfrey, 1992).

The purpose of the present study is to examine the effectiveness of correctional programs with Canadian federal offenders who have self-identified as belonging to diverse ethnic groups. All the correctional programs in CSC apply the CBT format and adhere to the RNR principles stipulated in the effective corrections literature (Andrews & Bonta, 2000; Gendreau & Goggin, 1996). Based on literature suggesting that CBT is equally effective in addressing psychological problems with minority clients as it is with Caucasian clients, it is hypothesized that CBT-based correctional programs will be effective in reducing criminal recidivism for a range of ethnic groups.

Method and Procedure

Selection of Studies

To examine the effectiveness of CSC correctional programming with federal offenders of various ethnicities, a meta-analytic approach was applied. One of the main advantages of a meta-analysis is that it will detect effects that may be difficult to quantify in other approaches to summarizing research, such as narrative summaries or literature reviews. Furthermore, a meta-analysis calculates an effect size for each study and pools those estimates across studies, thereby providing an overall effect estimate with considerably more statistical power than individual studies (Lipsey & Wilson, 2001). As a result, studies with smaller sample sizes, which are common with ethnic minority samples, can be grouped to produce a more meaningful evaluation of treatment effectiveness.

Some potential limitations exist, including the “file drawer” problem where nonsignificant results are less likely to be published and therefore, accessed by researchers. There is also the potential for unfavourable studies to be excluded from a meta-analysis if the researcher is personally invested in the result. To increase the relevancy of the conclusions of this study to a Canadian federal offender context, the studies included in this meta-analysis involved those conducted with CSC samples only. A search of all research previously undertaken by CSC on outcomes of correctional programs was conducted. The final selection of studies for inclusion was based on the following criteria.

Intervention modality

The correctional programs evaluated were delivered by CSC in a federal institution or parole office and were a variant of CBT or was substantially similar to the principles and interventions used in CBT. CSC defines a correctional program as “a structured intervention that addresses the factors directly linked to offenders’ criminal behaviour” (CSC, 2011, para. 5). CSC correctional programs target criminal behaviour by using cognitive-behavioural skills training. This definition excludes vocational, educational, or leisure programs.

Participants

The program participants as well as offenders in the comparison groups were male and female offenders serving a federal sentence (i.e., a sentence of 2 or more years administered by CSC). Reports evaluating programs for female offenders that met the selection criteria were included in this study so as to represent as closely as possible the range of correctional programs provided by CSC. Women offenders make up approximately 4% of the Canadian federal offender population and there are currently six correctional programs specifically targeted to women (CSC, 2010). Although these programs provide gender-specific treatment (Covington & Bloom, 2006), they are also based on a CBT and RNR framework. The literature suggests that there is little evidence that gender significantly affects CBT outcomes. Landenberger and Lipsey’s (2005) meta-analysis of CBT programs for offenders found that the gender mix of recipients showed no relation to effect size. Within CSC, offenders who meet the program referral criteria as determined through a thorough intake assessment of their risk level and criminogenic needs are referred to one or more correctional programs. Recent research conducted by CSC demonstrated that the proportion of offenders from Black, Aboriginal, Caucasian, or other ethnic groups who met the referral criteria for a correctional program did not differ from the proportion from each group who are eventually enrolled in at least one correctional program (Stewart & Wilton, 2012). Thus, we determined that program participation rates reported in the present research were not biased toward specific ethnic groups.

To be selected for the meta-analysis, the study sample must have a range of ethnic groups, including, but not limited to, Caucasian, First Nations/Aboriginal, Black, South American, Asian, and South/East Asian. Where ethnicity was not a variable included in the original study, the ethnic background variable was obtained from archival data in CSC. CSC maintains a database of demographic information for all federal offenders. The self-reported ethnic status of offenders is recorded during the intake process and is described in more detail below.

Outcome measures

The studies used in this meta-analysis must have reported readmission to custody subsequent to participation in a correctional program. All reports used readmission to custody as an outcome measure. Readmissions included violations of terms of conditional release and new criminal offences. Separate outcomes were presented for a variety of ethnic groups. If the publication did not provide enough information to code effect sizes for a variety of ethnic groups, the authors were contacted and complete data sets were requested.

Methodology

The studies selected used a randomized or matched-control design that compared the treatment condition with a comparison group that did not receive the correctional program. Studies that did not use a comparison group were excluded. All but one of the studies included in this analysis used a matched-control design. The Cognitive Skills Training Program had a policy of randomly assigning eligible offenders to either treatment or a waitlist group until the practice was phased out in 1993 (Robinson, 1995). Five reports used an intent-to-treat design that included dropouts in the treatment sample. The three remaining reports included only those offenders who successfully completed the program.

Source

Studies were undertaken by, or overseen by, CSC and published as research reports by the department. Studies with both significant and nonsignificant findings were included. Of note is CSC’s (2009) Evaluation Report: Correctional Service Canada’s Correctional Programs, which consisted of outcome evaluations of every nationally recognised correctional program currently being delivered by CSC. Although compiled into one extensive document, results from each correctional program were conducted as a separate evaluation. Please see Table 1 for a detailed description of all reports included in the present study.

Table 1.

Description of Correctional Program Evaluation Studies.

Author	Program being evaluated	Reported ethnic group breakdown	Sample size	Study design	Outcome measure	Reported length of follow-up
Cortoni, Nunes, & Latendresse (2006)	Violence Prevention Program	Aboriginal, non-Aboriginal	T = 500, C = 466	T = intent-to-treat (completers, non-completers, dropouts) C = matched control (SIR, need, offence, age, ethnicity)	Readmission to custody (violations and new offences)	Mean time-at-risk: T = 1 yr, C = 1.6 yrs
Dowden, Blanchette, & Serin (1999)	Anger Management	Aboriginal, Black, Caucasian, Other	T = 110, C = 110	T = successful program completers C = matched control (SIR, age, offence)	Readmission to custody with a new offence	3 year follow-up
Kunic &Varis (2010)	Aboriginal Offender Substance Abuse Program	Aboriginal	T = 312, C = 100	T = intent-to-treat (completers, non-completers, dropouts) C = control (all other Aboriginal offenders admitted to custody during study)	Readmission to custody (violations and new offences)	40 month follow-up
Robinson (1995)	Cognitive Skills Training	Aboriginal, non-Aboriginal	T = 3,531, C = 541	T = intent-to-treat (completers, dropouts) C = random assignment to waitlist*	Readmission to custody (violations and new offences)	1 year follow-up
Correctional Service Canada (2009)	Family Violence	Aboriginal, Black, Caucasian, Other	T = 998, C = 923	All programs evaluated by this author follow the same study design: T = intent-to-treat (completers, non-completers, dropouts) C = matched control (risk, time, opportunity)	All programs evaluated by this author use the same outcome measure: Readmission to custody (violations and new offences)	All programs evaluated by this author report the same length of follow-up: 3 year follow-up
	Aboriginal Family Violence	Aboriginal	T = 15,C = 1
	Living Skills	Aboriginal, Black, Caucasian, Other	T = 6,671C = 3,277
	Circles of Change	Aboriginal	T = 22,C = 2
	Aboriginal Basic Healing	Aboriginal	T = 20,C = 7
	National Substance Abuse Program	Aboriginal, Black, Caucasian, Other	T = 1,499,C = 962
	Aboriginal Offender Substance Abuse Program	Aboriginal	T = 38,C = 9
	Aggressive BehaviourControl	Aboriginal, Black, Caucasian, Other	T = 315,C = 160
	Violence Prevention Program	Aboriginal, Black, Caucasian, Other	T = 629,C = 792
	In Search of your Warrior	Aboriginal, Caucasian	T = 284,C = 155
	Spirit of a Warrior	Aboriginal	T = 45,C = 4
	Sex Offender Program	Aboriginal, Black, Caucasian, Other	T = 1,061,C = 293
	Aboriginal Sex Offender Program	Aboriginal, Black, Caucasian, Other	T = 193,C = 61
Stewart, Hamilton, Wilton, Cousineau, &Varrette (2009)	Tupiq	Aboriginal	T = 61,C = 115	T = intent-to-treatC = control (all other Inuit sex offenders in custody during study period)	Readmission to custody (violations and new offences)	Mean time-at-risk: T = 4 yrs,C = 4.3 yrs
Trevathen, Moore, & Allegri (2005)	In Search of your Warrior	Aboriginal	T = 112,C = 64	T = program completersC = matched control (ethnicity, offence, age, sentence length, risk)	Readmission to custody (violations and new offences)	1 year follow-up
Yessine & Kroner (2004)	Counterpoint	Aboriginal, Black, Caucasian, Other	T = 332,C = 313	T = program completersC = matched control (risk, prior treatment)	Readmission to custody (violations and new offences)	Mean time-at-risk: T = 1.4 yrs, C = 1.4 yrs

Note: T = treatment group; C = comparison group.

Offenders initially assigned to the waitlist group who eventually received treatment were included in the treatment group

Coding Procedures and Statistical Analyses

Coding for ethnic group

All incoming federal offenders participate in the Offender Intake Assessment (OIA), a comprehensive assessment that provides background demographic information as well as an assessment of their static and dynamic risk factors. One of the demographic questions asks offenders to select the ethnic group with which they identify.

Based on the data available, separate analyses were undertaken for four different ethnic groups. These included Caucasian (generally European descent), Black (generally Caribbean and African descent), Aboriginal (Inuit, Innu, North American Indian, and Métis), and Other. The Other category consisted of all remaining ethnic groups. Due to low sample sizes, separate analyses could not be undertaken individually for the remaining ethnic groups, and were therefore combined into the Other category. The authors acknowledge that while there are likely finer distinctions to be made with respect to ethnicity, for the purposes of this study, ethnic background was categorized by racial identification.

Whenever possible, complete data sets for each program evaluation were examined to extract data pertaining to all ethnic groups. If the data sets were not available, effect sizes were calculated based on the information provided in the written reports. Studies that did not provide specific information on ethnic composition and whose data sets were not available were necessarily excluded from this report.

Meta-analysis and effect size calculation

The outcome data consisted of 2 × 2 tables containing the readmission outcomes of the treatment and comparison groups. The effect size measure chosen for this study was the odds ratio (OR), which is widely recommended for dichotomous data (Cooper, Hedges, & Valentine, 2009; Lipsey & Wilson, 2001). The OR compares the odds of an event between two groups. In this case, the event in question is success in the community on release (i.e., not recidivating) and the two groups refer to a treatment group and a comparison group within each ethnic category. For example, within the Aboriginal ethnic group, Aboriginal program participants were compared with Aboriginal offenders who did not attend the program. This was an important consideration because previous research had indicated that base rates of reoffending for offenders participating in correctional programs varied significantly based on their ethnic background (Usher & Stewart, 2010). An OR of 1.0 indicates no difference between groups. An OR greater than 1.0 indicates a positive effect for the treatment group, and an OR less than 1.0 indicates a positive effect for the comparison group. For each correctional program evaluated, separate effect sizes were calculated for each of the four ethnic groups. A summary statistic was then calculated for each of the four ethnic groups studied. Statistical analyses were conducted on the natural logarithm of the OR and were weighted based on the inverse standard error of the individual effect sizes. Procedures for the OR effect size calculations can be found in Cooper et al. (2009) and Lipsey and Wilson (2001).

Homogeneity of variance was tested using the Q statistic, which determines the extent to which the effect sizes varied across studies. Homogeneity is rejected if Q exceeds the critical value (α = .01) for a chi-square with k − 1 degrees of freedom, where k is the number of effect sizes (Lipsey & Wilson, 2001). If the test for homogeneity of variance was nonsignificant, a fixed-effects model was used. Otherwise, a random-effects model was used. If heterogeneity was found, the I² statistic was calculated to measure the level of inconsistency across studies (Higgins & Thompson, 2002).

In four cases, ORs could not be calculated from the data provided in the published reports alone. In these instances, the original data sets were consulted and effect sizes were calculated through a reanalysis of the data. The advantage of this approach is that ORs were calculated consistently across all studies, and therefore no conversions from other effect sizes were necessary. A limitation of this approach, however, is that should readers seek to independently consult these studies, the ORs reported in this report may not be found as they may not have been included in the original analyses.

Results

An initial search yielded 21 technical research reports published by CSC evaluating the effectiveness of CBT correctional programs. The final sample consisted of eight research reports that met the final criteria for inclusion, yielding more than 50 distinct effect sizes distributed across the four ethnic groups. These reports are marked with an asterisk in the reference section. Separate analyses were conducted for each of the four ethnic groups in question and are categorized according to the correctional program being evaluated.¹

Overall Mean Effect of CBT Programming on Recidivism

The mean OR representing the average effect of CBT-based correctional program interventions on Canadian federal offenders as a whole (i.e., regardless of ethnicity) was 1.65 (p < .001). This indicates that the odds of not recidivating for individuals in the treatment group were more than one and half times greater than those in the control group. We now turn to an examination of intervention effects according to ethnic group, which is the primary purpose of this study.

Caucasian Offenders

A total of 18 distinct effect sizes (k) were obtained from three reports, yielding an overall sample size (n) of 12,221 Caucasian offenders. The weighted mean effect size for this ethnic group was 1.76, 99% confidence interval (CI) = [1.65, 1.87]. This result indicates that the treatment group was significantly more successful on release than the comparison group. Specifically, among Caucasian offenders, the odds of not recidivating were 1.76 times greater for program participants than for nonparticipants. Homogeneity of variance was obtained, Q(17) = 26.01, p = .07, indicating that the dispersion of individual effect sizes around the mean is no greater than would be expected from sampling error alone. Table 2 displays the effect sizes for each correctional program as well as the overall mean effect size.

Table 2.

Effect Sizes for Caucasian Offenders by Correctional Program.

Program	OR	SE	p
Family Violence – Moderate Intensity (CSC, 2009)	1.73	0.148	< 0.01
Family Violence – High Intensity (CSC, 2009)	1.63	0.227	< 0.01
Living Skills – Anger and Emotion Management (CSC, 2009)	1.75	0.077	< 0.01
Living Skills – Anger and Emotion Management (CSC, 2009)^a	0.58	0.431	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)	1.62	0.098	< 0.01
Living Skills – Reasoning and Rehabilitation (CSC, 2009)^a	1.30	0.666	ns
National Substance Abuse – High Intensity (CSC, 2009)	1.81	0.240	< 0.01
National Substance Abuse – Moderate Intensity (CSC, 2009)	1.73	0.124	< 0.01
Aggressive Behaviour Control (CSC, 2009)	1.00	0.269	ns
Violence Prevention Program (CSC, 2009)	1.93	0.195	< 0.01
In Search of your Warrior – High Intensity (CSC, 2009)	2.19	0.984	ns
Sex Offender Program – High Intensity (CSC, 2009)	2.60	0.324	< 0.01
Sex Offender Program– Moderate Intensity (CSC, 2009)	2.41	0.245	< 0.01
Sex Offender Program – Low Intensity (CSC, 2009)	5.19	0.418	< 0.01
Aboriginal Sex Offender Program (CSC, 2009)	1.52	0.800	ns
Sex Offender Program – Special Needs Offenders (CSC, 2009)	2.83	0.715	< 0.05
Counterpoint (Yessine & Kroner, 2004)	2.04	0.192	< 0.01
Anger Management (Dowden, Blanchette & Serin, 1999)	3.74	0.460	< 0.01
Overall	1.76	0.043	< 0.01

Note: OR = odds ratio; SE = standard error; ns = non-significant.

Women offender program.

Aboriginal Offenders

In total, eight reports were examined and comprised a total 5,755 offenders (n) who self-identified as Inuit, Innu, North American Indian, or Métis. These reports yielded 28 separate effect sizes (k). The weighted mean effect size for this group was 1.45, 99% CI = [1.27, 1.63]. In other words, the odds of not recidivating were almost one and half times greater for Aboriginal offenders who participated in correctional programs than those who did not participate in programs. There was, however, a significant amount of variability across effect sizes, Q(27) = 50.93, p < .01. The I² statistic was then calculated to determine the amount of variability that may be attributed to heterogeneity. The level of heterogeneity was found to be 47%, which is considered moderate (Higgins & Thompson, 2002). This result indicates that the treatment effects may have been moderated by other variables or that Aboriginal offenders are a more heterogeneous group than other ethnic groups. Table 3 displays the OR effect sizes calculated for each correctional program as well as the overall mean effect size for this group.

Table 3.

Effect Sizes for Aboriginal Offenders by Correctional Program.

Program	OR	SE	p
Family Violence – Moderate Intensity (CSC, 2009)	1.50	0.278	ns
Family Violence – High Intensity (CSC, 2009)	1.11	0.424	ns
Aboriginal Family Violence – High Intensity (CSC, 2009)	0.25	1.817	ns
Living Skills – Anger and Emotion Management (CSC, 2009)	1.35	0.158	< 0.05
Living Skills – Anger and Emotion Management (CSC, 2009)^a	0.81	0.749	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)	1.23	0.209	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)^a	0.25	1.436	ns
Circles of Change (CSC, 2009)^a	1.5	1.652	ns
Aboriginal Basic Healing Program (CSC, 2009)	3.23	1.177	ns
National Substance Abuse – High Intensity (CSC, 2009)	1.87	0.496	ns
National Substance Abuse – Moderate Intensity (CSC, 2009)	1.32	0.233	ns
Aboriginal Offender Substance Abuse Program (CSC, 2009)	2.28	0.868	ns
Aggressive Behaviour Control (CSC, 2009)	1.03	0.505	ns
Violence Prevention Program (CSC, 2009)	1.57	0.385	ns
In Search of your Warrior – High Intensity (CSC, 2009)	1.05	0.240	ns
Spirit of a Warrior (CSC, 2009)^a	1.22	1.201	ns
Sex Offender Program – High Intensity (CSC, 2009)	0.77	0.609	ns
Sex Offender Program – Moderate Intensity (CSC, 2009)	2.06	0.481	< 0.05
Sex Offender Program – Low Intensity (CSC, 2009)	8.28	0.587	ns
Aboriginal Sex Offender Program (CSC, 2009)	1.42	0.355	ns
Sex Offender Program – Special Needs Offenders (CSC, 2009)	1.60	1.857	ns
Counterpoint (Yessine& Kroner, 2004)	12.5	0.796	< 0.01
Anger Management (Dowden, Blanchette & Serin, 1999)	11.63	1.056	< 0.01
Cognitive Skills Training (Robinson, 1995)	1.00	0.325	ns
In Search of your Warrior (Trevethan, Moore &Allegri, 2005)	0.57	0.363	ns
Violence Prevention Program (Cortoni, et al, 2006)	4.37	0.385	< 0.01
Tupiq (Stewart, Hamilton, Wilton, Cousineau & Varrette, 2009)	3.25	0.381	< 0.01
Aboriginal Offender Substance Abuse Program (Kunic et al 2010)	1.80	0.232	< 0.01
Overall	1.45	0.070	< 0.01

Note: OR = odds ratio; SE = standard error; ns = non-significant.

Women offender program.

Studies were further separated into Aboriginal-specific correctional programs and generic correctional programs. The overall mean effect size for Aboriginal offenders participating in Aboriginal-specific programming was found to be 1.39, 99% CI = [1.06, 1.72] and 1.48, 99% CI = [1.27, 1.70] for Aboriginal offenders participating in generic programs. Both effect sizes were significant and, although the magnitude of success for Aboriginal offenders who participated in the generic programs was slightly greater than for those who attended Aboriginal-specific programs, the difference between the two types of programs as computed by a Z test was not significant.

Black Offenders

For the Black ethnic group, a total of 16 effect sizes (k) were calculated based on three reports. This combination produced an overall sample size (n) of 1,150 offenders. The weighted mean effect size for this group was 1.36, 99% CI = [1.02, 1.71], which means that of the offenders who self-identified as Black, those who participated in a correctional program had odds of success that were 1.36 times greater than the nontreatment comparison group. In this case, homogeneity of variance was obtained, Q(15) = 6.14, p = .977. Table 4 displays the effect sizes calculated for each correctional program as well as the overall mean effect size.

Table 4.

Effect Sizes for Black Offenders by Correctional Program.

Program	OR	SE	p
Family Violence – Moderate Intensity (CSC, 2009)	1.50	0.456	ns
Family Violence – High Intensity (CSC, 2009)	1.60	0.684	ns
Living Skills – Anger and Emotion Management (CSC, 2009)	1.19	0.201	ns
Living Skills – Anger and Emotion Management (CSC, 2009)^a	0.5	2.236	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)	1.65	0.308	< 0.05
Living Skills – Reasoning and Rehabilitation (CSC, 2009)^a	4.00	1.414	< 0.05
National Substance Abuse – High Intensity (CSC, 2009)	0.33	1.354	ns
National Substance Abuse – Moderate Intensity (CSC, 2009)	1.47	0.713	ns
Aggressive Behaviour Control (CSC, 2009)	0.80	0.922	ns
Violence Prevention Program (CSC, 2009)	1.69	0.524	ns
Sex Offender Program – High Intensity (CSC, 2009)	0.25	1.732	ns
Sex Offender Program– Moderate Intensity (CSC, 2009)	1.57	1.111	ns
Sex Offender Program – Low Intensity (CSC, 2009)	12.00	1.780	< 0.01
Aboriginal Sex Offender Program (CSC, 2009)	2.00	2.236	ns
Counterpoint (Yessine& Kroner, 2004)	1.61	0.503	ns
Anger Management (Dowden, Blanchette & Serin, 1999)	0.67	1.555	ns
Overall	1.36	0.133	< 0.01

Note: OR = odds ratio; SE = standard error; ns = non-significant.

Women offender program.

Other Offenders

The remaining ethnic groups that make up CSC’s offender population were combined into one group. Offenders entering CSC who do not identify as Aboriginal, Caucasian, or Black can choose to identify themselves as Arab/West Indian, Asiatic, East Indian, Hispanic, Chinese, Filipino, Japanese, Korean, Latin American, South Asian, South East Asian, or Other. Because these samples were small, it was impossible to calculate effect sizes specific to each of these ethnic groups. They were therefore combined to form one group referred to as Other.

The total sample size (n) for this Other group was 884. A total of 16 distinct effect sizes (k) were calculated based on three reports. The overall mean effect size for this group was 1.53, 99% CI = [1.15, 1.91]. This finding means that offenders in the remaining ethnic group categories who participated in programs had greater odds of success after release than those in the comparison group who did not participate in programs. Homogeneity of variance was also obtained, Q(15) = 6.43, p = .97. Table 5 displays the effect sizes calculated for each correctional program along with the overall weighted mean effect size for this group.

Table 5.

Effect Sizes for the Other Ethnic Group by Correctional Program.

Program	OR	SE	p
Family Violence – Moderate Intensity (CSC, 2009)	3.60	0.136	< 0.01
Family Violence – High Intensity (CSC, 2009)	0.70	0.236	ns
Living Skills – Anger and Emotion Management (CSC, 2009)	1.51	0.241	< 0.05
Living Skills – Anger and Emotion Management (CSC, 2009)^a	0.50	1.732	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)	1.34	0.316	ns
Living Skills – Reasoning and Rehabilitation (CSC, 2009)^a	2.00	2.236	ns
National Substance Abuse – High Intensity (CSC, 2009)	6.00	0.289	< 0.01
National Substance Abuse – Moderate Intensity (CSC, 2009)	1.00	0.132	ns
Aggressive Behaviour Control (CSC, 2009)	1.25	0.236	ns
Violence Prevention Program (CSC, 2009)	1.64	0.158	ns
Sex Offender Program – High Intensity (CSC, 2009)	1.40	1.531	ns
Sex Offender Program – Moderate Intensity (CSC, 2009)	2.80	0.685	< 0.01
Sex Offender Program – Low Intensity (CSC, 2009)	6.00	1.472	< 0.01
Aboriginal Sex Offender Program (CSC, 2009)	1.33	1.528	ns
Counterpoint (Yessine& Kroner, 2004)	1.29	0.164	ns
Anger Management (Dowden, Blanchette & Serin, 1999)	1.33	0.447	ns
Overall	1.53	0.148	< 0.01

Note: OR = odds ratio; SE = standard error; ns = non-significant.

Women offender program.

When comparing across ethnic groups, results indicate that all ethnicities showed treatment gains over nontreatment comparison groups. Figure 1 summarizes the effect sizes for each ethnic group. The square denotes the effect size and the bars indicate 99% CIs. A Z test was used to compare effect sizes between ethnic groups. After applying the Bonferroni correction for multiple comparisons (p = .05/6), no significant differences were found between ethnic groups with respect to their response to correctional programming.

Figure 1.

Comparison of effect sizes across ethnic groups.Note: OR = odds ratio.

Discussion

The primary aim of this study was to compare the effectiveness of correctional programs that adhere to the principles of RNR and are based on a CBT treatment modality in reducing recidivism for a range of ethnically diverse Canadian federal offenders. Results indicate that each of the four ethnic groups examined demonstrated significant treatment gains as a result of participating in CBT-based correctional programming. In other words, participation in correctional programming significantly reduced the likelihood of readmission to custody, regardless of offenders’ ethnic background.

This finding is consistent with research indicating that many different ethnic groups respond well to CBT interventions (Miranda et al., 2005). Perhaps equally as important, these findings speak to the utility of attending to RNR principles when implementing correctional programs as these results should not be interpreted as minimizing the importance of attending to ethnic and cultural differences. The effective corrections literature specifies that correctional interventions should “maintain respect for, and attention to, diversity in both people and programming” (Andrews et al., 1990, p. 20). All program facilitators within CSC are trained to respect and incorporate aspects of offender culture into delivery of a program. For example, facilitators who deliver the Aboriginal programs incorporate cultural ceremonies, teachings, and approaches within the standard CBT approach. Aboriginal Elders may be involved in counselling Aboriginal offenders who attend both the culturally specific and the generic programs. Aboriginal offenders were found to show treatment gains from participating in both generic programming and Aboriginal-specific programming. Again, the implication of these findings is that CBT-based correctional programming can effectively attend to the responsivity needs of ethnic minority offenders and can address issues around cultural differences by adaptations within the CBT framework.

Homogeneity across effect sizes was found for all ethnic groups except Aboriginal offenders. Although the overall mean effect size was determined to be significant for this group, moderate levels of heterogeneity were found. Heterogeneity can be interpreted in a number of different ways. It can be considered as an indication of clinical variability in the participants or the interventions under study, or it may point to fundamental differences in the design and methodology of the studies being grouped (Higgins & Green, 2009). Given that, for the most part, the same studies were examined across all four ethnic groups, it is likely that a portion of the variability can be attributed to the differences among the participants and the quality of the implementation of the programs. Aboriginal offenders may simply be a more diverse group than the other ethnic categories, yielding a wider range of variability in treatment responsiveness.

It is important to note that although the overall difference in treatment effect sizes between ethnic groups was found to be nonsignificant, there remains the possibility of differences in recidivism outcomes between Caucasian and Aboriginal offenders. Recall that the 99% CIs were nonoverlapping between these two groups prior to Bonferroni correction. A review of the growing debate regarding the validity of the Bonferroni correction is beyond the scope of this article; suffice to say that there may be legitimate counterarguments to its use (Nakagawa, 2004; Perneger, 1998). Although small, the difference in effect size between these two groups may in fact be significant particularly given that this was found in spite of the heterogeneity among Aboriginal offenders. This outcome is nonetheless cause for concern, as a reduced treatment effect for this group will translate into a greater number of Aboriginal offenders who eventually return to custody. This finding warrants attention and should be the focus of further research for this agency. We would encourage readers to interpret this outcome with caution, as this finding may not be generalizable beyond CSC populations.

Some important limitations of this study should be mentioned. Because offenders often participate in multiple programs as part of their correctional plan, it is difficult to determine the specific effects of a single program. The individual effect sizes calculated for each program may have been influenced by participation in other programs. Nevertheless, the overall mean effect sizes calculated for each ethnic group can still be considered as an indication of general correctional program effectiveness. It should also be noted that another consequence of participation in multiple programs is the possibility of sample overlap between studies. Although unlikely given the publication dates of the included reports, it was impossible to determine the extent to which offenders may have been participants in more than one of the studies examined in this report.

Sampling methodology was not consistent across reports included in this study. While most used an intent-to-treat design, program evaluations including only successful completers in the treatment group risk inflating treatment effect. Samples of convenience were used in cases where the pool of potential offender participants was quite low, particularly for Inuit offenders qualifying for a sex offender program and Aboriginal offenders in need of a substance abuse program. Failing to use a matched-control design reduces methodological rigor and presents an additional limitation to the generalizability of those studies in question. Follow-up time was also variable across studies, ranging from 1 to 4 years.

Finally, certain correctional programs offered by CSC were not included in this meta-analysis due to lack of available data and small sample sizes, notably the women offender programs. For example, three reports detailing evaluations of women offender programs were excluded on the basis that they did not use recidivism as an outcome measure. The generalizability of these findings to women offender programs as a whole may therefore be limited. Other studies that did not provide sufficient information on the ethnic composition of their sample or did not present separate outcomes for each ethnic group were excluded from this report unless original data sets could be located. This exclusion may have resulted in a smaller pool of studies from which to draw on; however, outcomes for nearly every correctional program offered by CSC were included in this study.

Conclusion

This study provides support for the use of evidence-based correctional interventions with ethnic minority offender populations. In particular, offender rehabilitation programs that adhere to CBT and RNR principles appear to be effective with the broad range of ethnicities that currently make up the Canadian federal offender population. Outcomes may not be generalizable beyond this population given that the correctional programs examined in this study were restricted to those offered by CSC and adhere to strict guidelines for program development and implementation. Nevertheless, these findings may be used to inform other jurisdictions with diverse offender populations by encouraging the use of CBT-based rehabilitation programs that attend to the responsivity needs of ethnic minority offenders.

Footnotes

Acknowledgements

We wish to thank the Research Branch at the Correctional Service of Canada (CSC) for their support in facilitating this project and the Evaluation Branch at CSC for providing us access to their databases. This report could not have been completed without their efforts. We would also like to thank Shevaun Cory, Julie Blais, and Holly Wilson for sharing their expertise on meta-analytic procedures.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct (5th ed.). Cincinnati, OH: Anderson Press.

Andrews

D. A.

Bonta

Hoge

R. D.

(1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice and Behavior, 17, 19-52.

Andrews

D. A.

Dowden

(2006). Risk principle of case classification in correctional treatment. International Journal of Offender Therapy and Comparative Criminology, 50, 88-100.

Arroyo

J. A.

Miller

W. R.

Tonigan

S. J.

(2003). The influence of Hispanic ethnicity on long-term outcome in three alcohol-treatment modalities. Journal of Studies on Alcohol, 64, 98-105.

Aspinall

P. J.

(2002). Collective terminology to describe the minority ethnic population: The persistence of confusion and ambiguity in usage. Sociology, 36, 803-816.

Bhopal

(2004). Glossary of terms relating to ethnicity and race: For reflection and debate. Journal of Epidemiology & Community Health, 58, 441-445.

Cooper

Hedges

L. V.

Valentine

J. C.

(2009). The handbook of research synthesis and meta-analysis (2nd ed.). New York, NY: Russell Sage Foundation.

Correctional Service of Canada. (2003). Commissioner’s directive 726: Correctional programs. Ottawa, Ontario: Commissioner of the Correctional Service of Canada.

*Correctional Service of Canada. (2009). Evaluation report: Correctional Service Canada’s correctional programs. Ottawa, Ontario: Author.

10.

Correctional Service of Canada. (2010). Demographic profile of the Canadian federal offender population [Unpublished raw data]. Retrieved from Correctional Service of Canada Corporate Reporting System.

11.

Correctional Service of Canada. (2011). Commissioner’s directive 726. Ottawa, Ontario: Author.

12.

Corrections and Conditional Release Act. (1992). Retrieved from http://laws-lois.justice.gc.ca/eng/acts/C-44.6/

13.

*Cortoni

Nunes

Latendresse

(2006). An examination of the effectiveness of the Violence Prevention Program (Research Report No. R-178). Ottawa, Ontario: Correctional Services Canada.

14.

Covington

S. S.

Bloom

B. E.

(2006). Gender responsive treatment and services in correctional settings. Women & Therapy, 29(3-4), 9-33.

15.

Dai

Zhang

Yamamoto

Belin

T. R.

Cheung

Hifumi

S. S.

(1999). Cognitive behavioral therapy of minor depressive symptoms in elderly Chinese Americans: A pilot study. Community Mental Health Journal, 35, 537-542.

16.

*Dowden

Blanchette

Serin

(1999). Anger management programming for federal male inmates: An effective intervention (Research Report No. R-38). Ottawa, Ontario: Correctional Service of Canada.

17.

Gendreau

Goggin

(1996). Principles of effective correctional programming. Forum on Correctional Research, 8(3), 38-41.

18.

Gendreau

Little

Goggin

(1996). A meta-analysis of the predictors of adult offender recidivism: What works! Criminology, 34, 575-607.

19.

Gil

A. G.

Wagner

E. F.

Tubman

J. G.

(2004). Culturally sensitive substance abuse intervention for Hispanic and African American adolescents: Empirical examples from the alcohol treatment targeting adolescents in need (ATTAIN) project. Addiction Supplement, 99(2), 140-150.

20.

Hanson

K. R.

Bourgon

Helmus

Hodgson

(2009). A meta-analysis of the effectiveness of treatment for sexual offenders: Risk, need, and responsivity (Corrections Research User Report No. 2009-01). Ottawa, Ontario: Public Safety Canada.

21.

Hays

P. A.

(2006). Developing culturally responsive cognitive therapies. In Hays

P. A.

Iwmasa

G. Y.

(Eds.), Culturally responsive cognitive-behavioral therapy: Assessment, practice, and supervision (pp. 3-20). Washington, DC: American Psychological Association.

22.

Higgins

J. P. T.

Green

(2009). Cochrane handbook of systematic reviews of interventions (Version 5.0.2). The Cochrane Collaboration. Available from www.cochrane-handbook.org

23.

Higgins

J. P. T.

Thompson

S. G.

(2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21, 1539-1558.

24.

Hinton

D. E.

Pham

Tran

Safren

S. A.

Otto

M. W.

Pollack

M. H.

(2004). CBT for Vietnamese refugees with treatment-resistant PTSD and panic attacks: A pilot study. Journal of Traumatic Stress, 17, 429-433.

25.

Holleran Steiker

L. K.

(2008). Making drug and alcohol prevention relevant: Adapting evidence-based curricula to unique adolescent cultures. Family & Community Health, 31, 52-60.

26.

Huey

S. J.

Polo

A. J.

(2008). Evidence-based psychosocial treatments for ethnic minority youth. Journal of Clinical Child & Adolescent Psychology, 37, 262-301.

27.

Hwang

W. C.

(2006). The psychotherapy adaptation and modification framework: Application to Asian Americans. American Psychologist, 61, 702-715.

28.

Jackson

L. C.

Wenzel

Schmutzer

P. A.

Tyler

J. D.

(2006). Applicability of cognitive-behavior therapy with American Indian individuals. Psychotherapy: Theory, Research, Practice, Training, 43, 506-517.

29.

Kendall

(2002). Time to think again about cognitive-behavioural programmes. In Carlan

(Ed.), Women and punishment: The struggle for justice (pp. 182-198). Cullompton, UK: Willan.

30.

King

(2011). Scaling up the knowledge to achieve Aboriginal wellness. Canadian Journal of Psychiatry, 56, 73-74.

31.

*Kunic

Varis

D. D.

(2010). The Aboriginal Offender Substance Abuse Program (AOSAP): Examination of the effects of successful completion on post-release outcomes (Research Report No. R-217). Ottawa, Ontario: Correctional Service of Canada.

32.

Landenberger

N. A.

Lipsey

M. W.

(2005). The positive effects of cognitive-behavioral programs for offenders: A meta-analysis of factors associated with effective treatment. Journal of Experimental Criminology, 1, 451-476.

33.

Lin

(2001). Application of cognitive-behavioral therapy to counselling Chinese. American Journal of Psychotherapy, 55, 46-58.

34.

Lipsey

M. W.

Wilson

D. B.

(2001). Practical meta-analysis. Thousand Oaks, CA: Sage.

35.

Losel

Schmucker

(2005). The effectiveness of treatment for sexual offenders: A comprehensive meta-analysis. Journal of Experimental Criminology, 1, 117-146.

36.

Lowenkamp

C. T.

Latessa

Holsinger

(2006). Risk principle in action: What have we learned from 13,676 offenders and 97 correctional programs? Crime & Delinquency, 52, 77-93.

37.

MacKenzie

D. L.

(2006). What works in corrections: Reducing the criminal activities of offenders and delinquents. New York, NY: Cambridge University Press.

38.

McCormick

R. M.

(2000). Aboriginal traditions in the treatment of substance abuse. Canadian Journal of Counselling, 34(1), 25-32.

39.

Miranda

Bernal.

Lau

Kohn

Hwang

LaFromboise

(2005). State of the science on psychosocial interventions for ethnic minorities. Annual Review of Clinical Psychology, 1, 113-142.

40.

Miranda

Chung

J. Y.

Green

B. L.

Krupnick

Siddique

Revicki

D. A.

Belin

(2003). Treating depression in predominantly low-income young minority women. Journal of the American Medical Association, 290, 57-65.

41.

Nagayama Hall

(2001). Psychotherapy research with ethnic minorities: Empirical, ethical, and conceptual issues. Journal of Consulting and Clinical Psychology, 69, 502-510.

42.

Nakagawa

(2004). A farewell to Bonferroni: The problems of low statistical power and publication bias. Behavioral Ecology, 16, 1044-1045.

43.

Pearson

F. S.

Lipton

D. S.

Cleland

C. M.

Yee

D. S.

(2002).The effects of behavioral/cognitive-behavioral programs on recidivism. Crime & Delinquency, 48, 476-496.

44.

Perneger

T. V.

(1998). What’s wrong with Bonferroni adjustments. British Medical Journal, 316, 1236.

45.

Public Safety Canada. (2007). Corrections and conditional release statistical overview: Annual report 2007. Ottawa, Ontario: Public Works and Government Services Canada.

46.

Public Safety Canada. (2010). Corrections and conditional release statistical overview: Annual report 2010. Ottawa, Ontario: Public Works and Government Services Canada.

47.

Renfrey

G. S.

(1992). Cognitive-behavior therapy and the native American client. Behavior Therapy, 23, 321-340.

48.

*Robinson

(1995). The impact of cognitive skills training on post-release recidivism among Canadian federal offenders (Research Report No. R-41). Ottawa, Ontario: Correctional Service of Canada.

49.

Rossello

Bernal

Rivera-Medina

(2008). Individual and group CBT and IPT for Puerto Rican adolescents with depressive symptoms. Cultural Diversity & Ethnic Minority Psychology, 14, 234-245.

50.

*Stewart

Hamilton

Wilton

Cousineau

Varrette

(2009). An examination of the effectiveness of Tupiq: A culturally specific program for Inuit sex offenders (Research Report No. R-213). Ottawa, Ontario: Correctional Service of Canada.

51.

Stewart

Wilton

(2012). Correctional program participation: Are there differences among offenders of various racial backgrounds? (Research Brief No. 12–6). Ottawa, Ontario: Correctional Service Canada.

52.

*Trevethan

Moore

Allegri

(2005). The “In Search of Your Warrior” program for Aboriginal offenders: A preliminary evaluation (Research Report No. R-172). Ottawa, Ontario: Correctional Service of Canada.

53.

Usher

Stewart

(2010). The effectiveness of correctional programs with diverse offenders: A meta-analytic study (Research Report No. R-246). Ottawa, Ontario: Correctional Service of Canada.

54.

Wilson

D. B.

Bouffard

L. A.

Mackenzie

D. L.

(2005). A quantitative review of structured, group-oriented, cognitive-behavioral programs for offenders. Criminal Justice and Behavior, 32, 172-204.

55.

Wilson

S. J.

Lipsey

M. W.

Soydan

(2003). Are mainstream programs for juvenile delinquency less effective with minority youth than majority youth? A meta-analysis of outcomes research. Research on Social Work Practice, 13, 3-26.

56.

*Yessine

Kroner

D. G.

(2004). Altering antisocial attitudes among federal male offenders on release: A preliminary analysis of the counter-point community program (Research Report No. R-152). Ottawa, Ontario: Correctional Service of Canada.