FORUM on Corrections Research

Case need domain: "Attitude"

This article presents the findings of a meta-analysis that examined the predictive potency of criminal and antisocial attitudes in 32 studies. The review is organized around the principal components, subcomponents and indicators that make up the attitudinal domain of the Case Needs Identification and Analysis (CNIA). The meta-analysis yielded 112 correlations with recidivism/ misconducts. Overall, the meta-analysis indicated that the justice, violence and lifestyle components of the CNIA are moderately related to recidivism. In contrast, the society component was only weakly related to recidivism. The predictive power of the property component could not be ascertained given that no studies examined its relationship to future criminal behaviour. Although the strongest CNIA indicator was non-conforming attitudes, the weakest CNIA indicators were attitudes toward employment and school, marital and family, interpersonal relations, and personal and emotional stability. Interestingly, the review identified neutralizations denial and minimization techniques employed by offenders to minimize the severity and nature of their offending an established construct in the literature but not identified in the CNIA, as a significant predictor.

Definition and constructs

A single definition of an attitude is somewhat elusive as definitions have shifted and redefined themselves throughout the decades. However, the following definition has stood the test of time: "an attitude is a relatively stable pattern of beliefs, feelings and behaviour tendencies toward some object."² Moreover, attitudes are learned, malleable entities that directly influence behaviour. These hypotheses have been well corroborated in the published literature across several disciplines.

The importance of attitudes in offender rehabilitation was first recognized almost a century ago. Since then, at least 168 studies, including a recent meta-analytic review, confirm the ability of attitudes to predict criminal behaviour for both institutional and community adjustment.³ Attitudes are also productive targets for intervention, with changes in values and beliefs resulting in marked changes in behaviour.⁴

In general, attitudinal constructs, both pro-social and antisocial, have not been used consistently in criminal behaviour prediction literature (n = 168) compared with the measurement of personality factors (n = 621) and psychological distress (n = 226). Yet antisocial attitudes demonstrate a stronger relationship with criminal behaviour (r = .22) than either of the aforementioned variables (r = .21 and r = .08 respectively).⁵

Despite the relative empirical disinterest, the construct of criminal attitudes is strongly rooted in several prominent theories of criminal conduct that have emerged during the last 60 years. Even the various criminological theories that compete or even directly contradict one another, such as Freud's psychodynamic approach emphasizing innate mechanisms such as the id and Sutherland's differential association theory that emphasizes social learning, all agree on the utility of measuring antisocial attitudes, values and beliefs.

Overview

In 1990, the Correctional Strategy incorporated multi-method and multi-predictor assessment techniques and systematic reassessments into the rehabilitation regime. The Strategy established, among other things, the primacy of offenders' criminogenic needs in directing offender programming to ensure maximal post-release adjustment. In 1994, the Offender Intake Assessment (OIA) process was implemented across the Correctional Service of Canada. This process yields a comprehensive and systematic evaluation of each offender on admission to federal custody. Moreover, the OIA provides a baseline evaluation for routine reassessments during rehabilitation, and directs treatment and services during incarceration.

This meta-analysis provides a comprehensive overview of the literature on the relationship between the CNIA attitudinal indicators and criminal behaviour. The review will also identify those attitudes currently not assessed by the CNIA but that are present in the literature.

Methodology

The predictive efficiency of the CNIA attitudinal indicators was evaluated by a quantitative meta-analysis involving 32 of 645 studies originally identified as potentially relevant. A meta-analysis is a statistical technique that allows researchers to objectively aggregate the magnitude of a relationship between two variables across numerous studies in the form of an effect size (e.g., a Pearson r correlation coefficient). In the present meta-analysis, 32 studies produced 112 effect sizes with recidivism. Recidivism was broadly defined to include prison misconducts, conditional release technical violations, arrests, charges and reconvictions. These studies were restricted to those that used a longitudinal research design with a specified follow-up duration, provided sufficient statistical information necessary for the meta-analytic calculations and used offender samples. Further, an average inter-rater reliability estimate of .92, whereby all the studies were coded by two individuals and 92% of the time they interpreted and coded the study exactly the same, demonstrated an acceptable level of coding reliability.

A qualitative evaluation was also conducted. This component provided a narrative summary of psychometric and post-dictive classification studies that were not predictive. These studies were reviewed in light of possible future contributions to the attitude domain of the CNIA.

Key findings

The meta-analysis revealed several descriptive findings. First, 84% of the effect sizes were based exclusively on male offenders. Second, 60% of the studies were conducted in Canada with 64% of these coming from federal institutions. Third, the most commonly adopted assessment approach was the pencil-and-paper self-report questionnaire format (62%). Information pertaining to age (59%), employment status (90%), education level (78%), ethnic origin (72%) or marital status (90%) was often not reported.

The meta-analysis left little doubt that attitudes in general are potent predictors of future behaviour of offenders. The weighted and unweighted mean effect sizes for each principal component, subcomponent and indicator are displayed in Table 1. Weighted effect sizes were calculated to account for magnitudinal differences accorded by the sample size of the study. Overall, the justice, violence and lifestyle components of the CNIA were moderately related to recidivism with weighted mean effect sizes ranging between .12 and .17. In contrast, the society component was only weakly related to recidivism (weighted mean effect size = .06). The predictive power of the property component could not be ascertained given that no studies were found that examined its relationship to future criminal behaviour. Although the strongest CNIA indicator was non-conforming attitudes (weighted mean effect size = .21), the weakest CNIA indicators (weighted mean effect sizes < .10) were attitudes toward employment and school, marital and family, interpersonal relations, and personal and emotional stability. There was a moderate relationship between neutralizations and recidivism.

Table 1

Meta-analysis Results: Weighted (Mz') and unweighted (Mr) Mean Pearson Coefficients with Corresponding Confidence Intervals by Predictor
Predictor(k)	N	Mr	Cl	Mz'	Cl
Justice (35)    Laws - Negative toward law (17)    Enforcement - Negative toward police (4)    Judicial system - Negative towards courts (2)    Corrections (12)       Negative towards corrections (10)       Negative towards community supervision (0)       Negative towards rehabilitation (3)	4,873 3,472 706 464 2,099 2,210 0 524	0.18 0.18 0.18 0.14 0.16 0.14 - 0.27	0.12-0.22 0.16-0.25 - - 0.05-0.22 0.05-0.22 - -	0.12* 0.12* 0.17 0.11 0.10* 0.10* - 0.13	0.09-0.15 0.09-0.15 - - 0.06-0.14 0.06-0.14 - -
Society (32)    Convention (30)       Employment/education has no value (3)       Marital/family relations have no value (4)       Interpersonal relations have no value (16)       Values substance abuse (0)       Basic life skills have no value (2)    Personal/emotional stability has no value (3)    Women/men - Women/men roles are unequal (2)    Minorities (ethic/religion/disabled) (1)	11,780 11,123 1,793 1,579 10,708 0 418 360 0 685 573	0.15 0.15 0.07 0.25 0.13 - 0.24 0.03 - 0.19 0.12	0.09-0.20 0.09-0.21 - - 0.06-0.20 - - - - - -	0.06* 0.06* 0.02 0.06 0.05* - 0.11 0.04 - 0.17 0.12	0.04-0.08 0.04-0.08 - - 0.03-0.07 - - - - - -
Property (0)    Personal - Disrespectful of personal belongings (0)    Communal - Disrespectful of personal belongings (0)    Commercial - Disrespectful of personal belongings (0)	0 0 0 0	- - - -	- - - -	- - - -	- - - -
Violence (6)    Domestic - Supportive of domestic violence (0)    Insrumental - Supportive of instrumental violence (6)	1,025 0 1,025	0.15 - 0.15	0.04-0.26 - 0.04-0.26	0.17* - 0.17*	0.11-0.23 - 0.11-0.23
Lifestyle (39)    Goal directed - Lacks direction (7)    Conforming - Non-conforming (27)    Neutralizations* (5)	7,394 3,585 5,001 1,012	0.16 0.12 0.20 0.14	0.13-0.19 0.05-0.20 0.15-0.25 -	0.16* 0.10* 0.21* 0.10	0.14-0.18 0.07-0.13 0.18-0.23 -

Further analyses revealed that there were no statistically significant differences between the ability of attitudes to predict community versus institutional adjustment. As well, predictive potency did not differ between assessment methods employing a single indicator versus those using multiple indicators simultaneously. However, the results did approach statistical significance in favour of the multiple indicator approach.

Conclusion

The review uncovered an extensive and comprehensive literature base involving attitude assessments with offenders. However, this research lacked the convincing longitudinal predictive studies that definitively substantiate the hopeful claims of preliminary psychometric, correlational and cross-sectional studies. Further much of the research did not examine several CNIA indicators. For instance, not a single study could be located that addressed the aspects of attitudes toward substance abuse, the elderly, disrespect toward any type of property, or intolerance of disabled persons or other cultures. Despite this, 112 predictive effect sizes tapping specific CNIA attitudinal indicators were located with a large majority demonstrating modest predictive capability. In sum, there is empirical support for the presence of the attitude domain in the CNIA.

The obvious strength of the CNIA attitude domain lies in its multi-dimensional nature spanning a broad spectrum of values and beliefs. However, given scarce resources and demanding time pressures, it may be most profitable to consider focusing on those attitudes that have produced at least moderate correlations with recidivism. Another potential amendment to the CNIA process would be to incorporate measures examining changes in attitudes as an offender progresses through the system. Attitudes are dynamic in nature and may easily provide a natural pocket of greater predictive potency if resources were focused on pertinent changes.

Moreover, one might consider adding additional indicators that tap neutralization techniques such as denial of victim injury and denial of responsibility. Lastly, promising new attitudinal concepts such as those contained in the Criminal Self-efficacy Scale⁶ and the Violence Belief Survey⁷ could potentially merit inclusion in the CNIA protocol. However, such amendments require exploration during field consultations.

---

1. 1125 Colonel By Drive, Ottawa, Ontario K1S 5B6.

2. G. W. Allport, "Attitudes," Handbook of Social Psychology, C. Murchison, Ed. (Worchester, MA: Clark University, 1935).

3. P. Gendreau, D. Andrews, C. Goggin and T. Chanteloupe, Prediction of Criminal Behaviour: A Meta-analysis, unpublished paper (University of New Brunswick and Carleton University, 1992). See also P. Gendreau, T. Little and C. Goggin, "A meta-analysis of predictors of adult offender recidivism: What works!" Criminology, 34 (1996): 575607.

4. J. Krosmick, "Attitude importance and attitude change," Journal of Experimental Social Psychology, 24, 3 (1988): 240255. See also J. A. McCarty and L. J. Shrum, "The recycling of solid wastes: Personal values, value orientations and attitudes about recycling as antecedent to recycling behaviour," Journal of Business Research, 30, 1 (1994): 5362.

5. P. Gendreau, D. Andrews, C. Goggin and T. Chanteloupe, Prediction of Criminal Behaviour: A Meta-analysis.

6. S. L. Brown, E. Zamble and P. Nugent, Criminal Self-efficacy Scale, unpublished test (Kingston, ON: Queen's University, 1998).

7. M. Bettman, The Violence Belief Survey, unpublished test (Kingston, ON: Queen's University, 1996).