Correctional Service Canada
Symbol of the Government of Canada

Common menu bar links

FORUM on Corrections Research

Warning This Web page has been archived on the Web.

Anger Management Program Outcomes

Increasingly during the last decade, correctional systems in North America have felt that anger management programs would benefit their clients. Many offenders are thought to have difficulty dealing with anger, aggression and hostility. Such difficulties are assumed to have a causal relationship with offenders' antisocial and criminal behaviour, and with the level of violence in correctional institutions.

The content of anger management programs has been described in some detail by Novaco(2)
and Ellis.(3) Cognitive-behavioural programs specifically for use with offender populations have been described by Fink,(4) Kitchner and Kennedy,(5) and Samuel(6) However, with the notable exception of Novaco's 1978 monograph there has been a dearth of studies evaluating the effect of such programs.

The results of the few evaluation studies published are not always clear. Typically, studies show short-term treatment success using measures of self report, but show no differences in ratings of institutional behaviour. Long-term evaluation of the effectiveness of anger management programs is notably absent from the literature. Yet long-term evaluation is particularly crucial with offender populations, where it is assumed that anger management programs will have some impact on future criminal behaviour. These assumptions affect decisions regarding future risk and conditional release programs for offenders. Parole boards and case management officers have been reluctant to grant or recommend conditional releases to inmates who have received no appropriate help for anger problems.

The present study evaluates the short- and long-term impact of a cognitive behavioural anger management program on a group of adult male offenders. Program Description The program consisted of 12 weekly two-hour sessions in which a combination of educational and experiential material was used to address three basic issues:
  1. Understanding the concept of anger and why and when to control anger. Techniques used included arousal awareness, anger recognition, basic moral reasoning and concepts of self-interest.
  2. Reducing anger cognitively through the use of coping selfstatements, problem-solving exercises, a sense of proportion and humour, and the basic tenets of rational-emotive therapy.
  3. Modifying and improving behavioural coping skills through relaxation training, assertiveness training and role playing different behavioural responses.
The program emphasized role playing, both to model new behaviours and to allow inmates to practise new skills. In a group setting, this also allowed clients to receive feedback, in a relatively "safe" situation, on how their behaviour determined the behaviour of others. Thus, participants were taught to use assertive behaviours in place of aggressive or under assertive responses when angry feelings were unavoidable. The program also taught participants to reappraise external events cognitively to reduce the frequency and intensity of angry feelings.

While the focus and aims of each session were strictly adhered to, the content of the role playing remained flexible. Appropriate coping skills depend very much on the situation and may best be taught within the context of the inmate's own problems. These coping skills typically include how to respond to accusations, how to complain effectively and how to keep out of fights.

The anger management group was jointly run by an experienced clinical psychologist with advanced training in rational-emotive therapy and a drama teacher from a local university, who was assisted by a drama student. The latter two were well versed in the analysis of behaviour such as eye contact, body posture and voice. Dramatic acting involves skills in observing environmental factors that shape thoughts, feelings and behaviour. Thus, the drama teacher and student had considerable credibility in teaching clients how to act in different and more appropriate ways. Study Method A study was carried out on one group of offenders to evaluate the short- and long-term impact of this cognitive behavioural anger management program. This group was compared with a second group of offenders, who were referred and assessed as suitable clients for the program, but who did not start or did not complete the program for various reasons. Initial Assessment Over a period of 24 months, 79 male offenders incarcerated in Kingston Penitentiary were referred and assessed as suitable clients for an anger management group program. All were administered an extensive battery of assessments which included the Ammons Quick Test, the Beck Depression Inventory, the Over controlled Hostility Scale from the Minnesota Multiphasic Personality Inventory, the IPAT Anxiety Inventory, the Interpersonal Behavior Survey, the Jones Irrational Beliefs Test and questionnaires relating to physical symptoms of anger and anger-provoking situations.

A total of 52 offenders attended at least six group sessions (i.e., half of the program) and were deemed to have received treatment. Another 27 offenders were deemed not to have received treatment, including 8 who had started the program but dropped out after one or two sessions. The remaining 19 decided not to participate in the program for a variety of reasons, such as work priorities, imminent transfer to another institution or lack of interest. This comparison group did not differ significantly from the offenders who completed the program on any of the initial psychometric assessments. Immediate Postprogram Follow-up Measures Within four weeks of completing the program, all participants were again given the two anger questionnaires, the Interpersonal Behavior Survey and the Jones Irrational Beliefs Test.

Attempts to retest non-participants (those who were referred but who did not complete the program) were largely unsuccessful. There were, therefore, insufficient data to report on this comparison group.

Participants in the last two group programs were also put through three standardized role-playing tests, both before and after the program. Similar procedures to those described by Novaco(7) were used, and participants were asked to rate their level of anger immediately after each role-playing situation. Long-Term Follow-up Measures During a nine-month period, four years after completion of the last program, an attempt was made to locate all 79 offenders who had been referred to the program. Where possible, subjects were asked to report how many hours of psychological or other professional treatment they had received since completing or being referred to the program.

In addition, each offender's case management officer was asked to rate the subject's ability to cope with problematic emotions (e.g., anger and anxiety) and to cope in various life domains (e.g., employment, family and friends). The likelihood of the offender remaining crime-free was also rated.

For those offenders who had been released, Canadian Police Information records and Correctional Service of Canada records were obtained to learn whether the subject had reoffended, and if 50, the type of crime and the latency to rearrest (i.e., the length of time that had elapsed between release and rearrest). Results Table 1 summarizes the differences in preprogram and postprogram scores for the questionnaires and psychometric tests. The results include only those offenders who completed the anger management program. As mentioned above, insufficient information was available on the comparison group of non-participants. Furthermore, the almost 50% reduction in the number of subjects, from the 52 who had completed the program, resulted from subjects' refusal or unavailability to complete the postprogram measures.

Table 1
Pre and post program Scores * for Offenders Completing the Anger
Management Program
Physical Symptoms (N=25)
Anger Inventory (N=26)
Irrational Beliefs (N=21)
Interpersonal Behavior Survey (N=27)
   GGR (Aggression)
   SGR (Assertion)
   Conflict Avoidance
   Impression Management


Role Playing Self Ratings (N=9)
* All differences in scores are highly significant statistically, with the exception of
scores on the measure of conflict avoidance

With one exception, all differences were highly significant, and in the expected and desired direction. That is, participants performed significantly better on almost ah measures after completion of the program. The one exception is in conflict avoidance where any change in participants' scores after treatment could have been construed as undesirable.

Table 2 summarizes the long-term follow-up measures for program participants and non-participants.

Table 2
Comparison of Follow-up Measures* for Program Participants and Non-Participants
Case management officer adjustment ratings**
33.58 (N=19)
40.09 (N=42)
Treatment hours since program
11.12 (N=17)
22.28 (N=38)
Latancy to rearrest (months)**
5.55 (N=11)
22.04 (N=12)
Any evidence of recidivism
No Yes
31.3% 68.8%
No Yes
44% 56%
Conviction for violent crime
No Yes
31.3% 65.8%
No Yes
60% 40%
* Results measured at the time of follow-up
** Difference in scores is statistically significant

Measures of recidivism were restricted to offenders who had been released. Of the total number of subjects, 39 had never been released and most of these were serving life sentences. Two of those never released died in prison during the period of the long-term follow-up. One individual released into the community also died. None of the deceased were included in the long-term follow-up figures.

Differences between the two groups on measures of recidivism are unclear, making interpretation difficult. Four out of five measures show a trend in the desired direction. However, the results for "latency to rearrest" and "treatment hours" should be viewed with caution, since the scores within each group varied 50 much.

With only two exceptions, offenders in this study had histories of violence. This included a number of offenders convicted of murder, and murder offenders typically have very low rates of recidivism. However, of the 41 released offenders, only one had been serving a life sentence for murder; 37 had a history of multiple incarcerations when they were referred to the program. The released subjects may have therefore been considered a high-risk group for violent reoffending. Thus, it was important to distinguish between violent reoffending and general recidivism which included breach of parole conditions and outstanding charges which had not resulted in reincarceration. While program participation showed no effect on general recidivism, the effect on convictions for further violent crime did approach significance. Discussion Changes resulting from the program, as measured by self-report and other psychometric tests administered immediately after the program, can clearly be demonstrated.

Similar changes in Interpersonal Behavior Survey scores to those reported here have been reported by Bellemare and McKay.(8) However, such changes may not be entirely trustworthy, as evidenced by the significant increase in "impression management" in Table 1. Subjects have many reasons for professing a change in how they feel, and how they would hypothetically behave, after completion of the program. Even with the best of intentions, improved skills and self-efficacy, the demands of a provocative situation may overwhelm the subject and elicit ingrained dysfunctional reactions. It would be reassuring to have more behavioural measures of change.

For several reasons, such behavioural measures are difficult to obtain. In this study, an attempt was made to use the number of institutional charges as a measure of behaviour. However, charges were so rare as to render the measure useless. Also, a case manager's rating of an offender's coping skills, some years after program completion, is somewhat tenuous as a measure of behaviour. Nonetheless, the difference in ratings between program participants and non-participants is significant and in the desired direction.

The need for follow-up or maintenance treatment to bring about long-term behavioural control has been eloquently argued by those involved in sex offender treatment and research (e.g., FORUM, Vol. 3, No. 4, December 1991, on sex offender program-ming). Maintenance treatment and follow-up are equally essential for correctional clients with any long-standing pattern of dysfunctional behaviour. Unfortunately, many offenders, as well as case managers, fail to accept the need for follow-up treatment; they believe the program somehow cures or fixes the problem. This attitude 15 typified by the offender who, in no uncertain terms, angrily rejected the suggestion of taking an anger management treatment program. He said that he had already done anger management and had all the notes in a box somewhere. Unfortunately, the program content was neither in his head nor in his behavioural repertoire.

The attempt to collect information on further treatment was not entirely successful in this study. While there was a trend for program participants to continue with more treatment than the comparison group, further review of the "treatment hours" data suggests that this measure was also related to recidivism and further incarceration. There was a significant positive correlation between the number of months of reincarceration and total hours of treatment. Most subjects who were reincarcerated had further treatment after reincarceration, rather than in the community where it might have been of more value. Thus, the tendency for offenders to take part in treatment programs only when incarcerated obscures the role of continued treatment in reducing recidivism. To compound the problem in this study, some data on continued treatment were missing, and most of the missing cases were subjects who were released successfully into the community.

A further complication arises from the kind of additional treatment offered to or taken by the subject. It is reasonable to assume that some treatment may not be compatible with the cognitive-behavioural orientation of the anger management program, with its emphasis on self-management and responsibility. Thus, further treatment may have been counter-productive. In this study, inmates supplied their best guesstimate of the number of hours of professional help they had received. Many subjects were vague and unsure of what treatment they had received. Information on treatment was rarely included in a subject's psychological records, and even if it was, it was not always clear how much or what kind of treatment had been offered.

Like many quasi-experimental studies conducted by clinicians whose attention is divided among competing priorities and pressures, there are a number of methodological design problems. Offenders are not always co-operative in completing lengthy questionnaires. Because of other pressures, some offenders were allowed into the program without completing the preprogram assessment. Others refused to participate in the program because they did not like filling out the preprogram psychometric tests. Needless to say, they were not about to comply with a request to redo the same battery of questionnaires. Missing data were frequently a problem, making numbers too low to perform an adequate analysis.

The movement of offenders within the system is both time consuming and difficult to track. Once past an offender's warrant expiry date, tracking is virtually impossible unless the offender commits another offence. This makes any collection of longitudinal measures difficult. Official records are more readily available but are not always designed to capture the information required. For example, Correctional Service of Canada summaries give sentence commencement dates for new charges. These dates are rarely the same as reincarceration dates, which are more pertinent to a measure of latency to rearrest.

Finally, there 15 the problem of whether to include offenders serving life sentences in this type of study. Long-term program assessment using such traditional measures as recidivism are clearly inappropriate. The length of time to release is unusually long and recidivism rates are low, regardless of treatment. Andrews and colleagues(9) argue against treating low-risk and low-need offenders. Many lifers can be 50 categorized, and yet many demand psychological programs and aim to achieve considerable personal growth from participating in them. Certainly, such programs are a means of passing time profitably and may be a means of reaffirming the non-criminal values held by many lifers. Summary For those offenders in this study who were released during the follow-up period, the rate of overall recidivism was relatively high (61%). Whether or not the subject had completed the anger management program had no significant effect on the overall rate. However, if we examine only reconvictions for a violent crime, program completion had a more noticeable effect (the rate of recidivism was 40%). Certainly, the length of time in the community before rearrest was significantly longer for those who had completed the program.

Once again, we are brought back to the question of expectations of change in high-risk violent offenders. Perhaps any reduction in violent crime and any increase in latency to reoffend should be considered a success.

This study offers some evidence of the positive impact of completing a cognitive-behavioural anger management program on violent offenders. The program was based on the now-accepted principles of effective correctional treatment, which emphasize self-monitoring and responsibility for attitudes and behaviour.

From both anecdotal evidence provided by the many graduates of cognitive-behavioural programs, such as the one under consideration, and the limited results of the present study, I remain optimistic about the value of such programs.
(1)Correspondence may be addressed to Gareth V Hughes, Ph.D., District Psychologist, Eastern & Northern District Parole Office, Correctional Service of Canada, 920 Princess Street, Suite 203, Kingston, Ontario K7L 1H1.
(2)R. Novaco, Anger Control: The Development and Evaluation of an Experimental Treatment (Lexington, Mass.: DC. Heath, 1975).
A. Ellis, How to Live With and Without Anger (Toronto: Fitzhenry & Whiteside, 1977).
E. Fink, A Cognitive-Behavioral Group Therapy Program for Problems of Anger and Aggression in Adult Male Offenders. Doctoral dissertation, Rutgers University, 1980.
E.P. Kirchner and R.E. Kennedy, Leaders' Manual for an Assertive Skills Course in Correctional Settings (University Park, Pa.: Institute for Research on Human Resources, 1978).
R. Samuel, "ABC of Anger Management," Canada's Mental Health, 31,1 (1983): 21.
Novaco, Anger Control.
F. Bellemare and D. McKay. "An Anger Management Program in a Maximum Security Federal Penitentiary." Unpublished poster presentation, 1990.
(9)D.A. Andrews, L Zinger, R.D. Hoge, J. Bonta, P. Gendreau and F. T. Cullen, "Does Correctional Treatment Work? A Clinically Relevant and Psychologically Informed Meta-Analysis," Criminology, 28(1990): 369-404.