Success of anger management treatment with individuals who have intellectual disabilities convicted of assault-related offenses has not been verified. We employed a single case design with repeated measures with 6 such men. Recidivism is reported at least 4.5 years and up to 10 years. Modified anger management training incorporating cognitive restructuring and arousal reduction was employed. Participants showed no uniform reductions in emotional or behavioral systems of anger and aggression. Although several retained significant anger feelings, there were reductions in the extent to which they would act in an aggressive fashion. Five have not re-offended; 1 re-offended within 6 months but not in the subsequent 4 years. Anger management treatment seems effective for men with intellectual disabilities in the community who have committed socially and legally unacceptable acts.
Researchers have found that between 10% and 24% of offenders with intellectual disabilities have committed offenses involving violence (Day, 1993; Gibbons & Robertson, 1983; Lund, 1990). In a study of 75 incarcerated offenders with intellectual disability in Australia, Klimecki, Jenkinson, and Wilson (1994) found that assault constituted 22% of offenses, and there was a recidivism rate of 42% among those who had committed assault over the 3.5 years covered by the study. The authors felt that this may have been an underestimate because a number of individuals who had committed assault were still incarcerated and, therefore, unable to re-offend. When compared to offenders who did not have intellectual disability, those with intellectual disability had higher rates of serious and more violent crimes (Baroff, 1996). Noble and Conley (1992) reported violent crime rates of up to 38% in a range of studies on prisoners with intellectual disability. These authors are careful to point out that the general picture is complex and does not necessarily indicate a greater proclivity for violence in this population. For example, policies towards incarceration may affect prison populations, with greater percentages of offenders with intellectual disability receiving prison sentences as a result of more restrictive policies. When considering community programs, Noble and Conley reported that the majority of offenders with intellectual disability were convicted of minor offenses and misdemeanors. However, these studies do suggest that a significant problem is present, and Noble and Conley noted the importance (and paucity) of appropriate rehabilitation and educational programs.
Over the last 10 years, there have been some reports on the in-patient management of offenders with intellectual disability. Murphy and Clare (1991) and Clare and Murphy (1992) described an in-patient treatment unit specifically designed to assess, treat, and manage individuals with mild intellectual disability and offending behavior. This was a mixed group of offenders, and results for violent offenders were reported as broadly positive (Clare & Murphy, 1992). Xenitidis, Henry, Russell, Ward, and Murphy (1999) reported that 82% of individuals treated under such a regime are returned to community placements.
In terms of the re-offending rates for individuals who have completed treatment programs, there are no data for those with intellectual disabilities. However, it is interesting to note re-offending rates for probation and day attender projects in the mainstream offending population. Brownlee (1995) reported the effects of intensive probation with young adult offenders. With a group whose offenses included assault, theft, driving violations, vandalism, drug- and alcohol-related charges (but excluding sex offenders), he found re-offending rates of 42% after 1 year and 73% after 2 years. In a study of probation day centers, Mair and Nee (1992) found that 24 months after initial conviction, 63% of offenders had been re-convicted of another offense at least once. Therefore, reconviction rates for individuals without intellectual disability who are given probation sentences appear to be over 40% and are possibly as high as 70%.
The understanding and treatment of problems associated with anger and aggression have advanced significantly over the past 2 decades. Novaco (1975, 1986) has developed an influential theoretical analysis of the way in which people appraise their circumstances to determine whether they are likely to become angry and/or aggressive. Thoughts and intentions attributed to others, along with the way people construe any particular situation, will affect the way they perceive a threat in that situation. Therefore, in Novaco's approach to treatment, cognition, expectations, and appraisal of situations as well as the combination of external circumstances, physiological arousal, and cognitive mediation are crucial in situations involving anger and aggression. Several authors have adapted and simplified these methods and have also incorporated relaxation and arousal-reduction methods to make them understandable and suitable for individuals with intellectual disability (Black, Cullen, & Novaco, 1997; Black & Novaco, 1993; Jackson & Altman, 1996; Lindsay, Overend, Allan, Williams, & Black, 1998). Rose, West, and Clifford (2000) compared 25 individuals with intellectual disabilities who completed anger management training with 19 individuals from a waiting list control group. Anger management intervention produced significant reductions in expressed anger that were maintained at 6 and 12 month follow-ups. Although these results lead to optimistic hypotheses on the effects of treatment, the authors did not clearly explicate the extent to which their participants were actually offenders.
In summary, the extant literature highlights the importance of the issue of violent offenses among individuals with intellectual disabilities and the need for effective treatment. However, there are limited data on the outcomes of treatment for those in this population who commit violent offenses, particularly in terms of recidivism rates. Finally, although effective programs exist that are focused on anger management, they have limited application to individuals convicted of violent offenses.
In the present study we employed the adapted methods of anger management to treat 6 men with intellectual disabilities convicted of violent personal assaults. The men continued to live at home in the community and were treated while serving probation sentences. All attended four weekly treatment as out-patients and were followed-up for at least 4 years. On the basis of re-offending rates reported in studies on offenders without intellectual disabilities, we conducted this study to test the hypothesis that between half and two thirds of the sample were likely to re-offend within 2 years of conviction, despite the intervention.
Mr. A. was 38 years old and had an IQ of 64 (Wechsler Adult Intelligence Scale-Revised—WAIS-R). (All of the information, including ages, refers to participants at the point at which they were referred to us for treatment.) He had lived in a hospital for people with intellectual disabilities for much of his life and had been discharged in the late 1980s, following a lengthy program of rehabilitation. Within a few months he had committed two assaults and a series of thefts in local shops. The first two incidents were not prosecuted, and the police referred Mr. A. back to the intellectual disabilities system. However, subsequent assault offenses through 1990 and 1991 were prosecuted, and Mr. A. faced a series of charges for assault, breach of the peace, theft, and alcohol-related offenses. He was given a series of probation sentences for these assaults and faced subsequent charges of breach of probation when he committed further assaults and thefts. During this time he lived in a series of Salvation Army hostels and other similar homes. He continued to prefer to stay in these establishments. Following the series of charges for assault, theft, and breach of the peace, he was accepted for treatment pursuant to yet another 3-year probation order in May 1991. Mr. A. never committed assaults under the influence of alcohol; rather, the assaults were to obtain money for buying alcohol.
Mr. B. was 42 years old and had an IQ of 65 (WAIS-R). He had attended a local school for students with intellectual disabilities, and after leaving school he had no further contact with the intellectual disabilities service system. He married, had two children, and lived in an innercity neighborhood. Mr. B. had frequent arguments with neighbors, which resulted in verbal aggression, and in the year prior to his referral had begun to attack local children in the street with a shovel. He was charged on two occasions with breach of the peace and assault on several of the children in his neighborhood.
Mr. M. was 18 years old. His IQ was 69 (WAIS-R). He lived with his mother and saw his father fairly frequently. Mr. M. had a long history of verbal and physical aggression at school. Although he had attended a mainstream school, he was placed in a section with extra resources for slow learners. Mr. M. had also been diagnosed as having attention deficit disorder. He was well known to local police and had gone to the Children's Panel (Juvenile Court in Scotland) because of aggressive incidents that occurred both at school and in his local town. Upon reaching the age of 18, he continued with aggressive incidents and was charged with breach of the peace for walking up the middle of a street shouting and swearing at passersby and kicking passing cars. Following his conviction he was sentenced to 2 years probation with a condition that he attend our treatment program to address his aggressive tendencies.
Mr. C. was 21 years old; his IQ was 68 (WAIS-R). He lived with his parents, had always attended mainstream school with learning support, and had been expelled from several schools for aggression towards other children. He came to the attention of the intellectual disabilities service system when, with little or no provocation, he assaulted two young males, one in the street and one on a bus. Mr. C. was charged with both incidents, convicted, and received a probation sentence of 2 years. He was then referred to us for treatment, and it became clear that he was an extremely verbally aggressive individual both to staff members and to other clients. Indeed, it became an unresolved puzzle as to why he had stayed out of trouble from the time he left school until he was 21 years of age.
Mr. I. was 33 years old and had an IQ of 68 (WAIS-R). He had lived on his own or with girlfriends for the last 14 years. In his early teenage years, Mr. I. had attended a residential school for persons with behavioral problems, but because of his aggression, he had been expelled and transferred back home. His history was typified by physical and verbal aggression towards others and a large number of theft offenses. He had 15 offenses for assault and breach of the peace and many more for theft. Mr. I. spent four short periods in prison as a result of these offenses preceded by numerous sentences of probation and fines. At the time of contact with us, he had been charged with two assaults: attacking a store security officer during a theft and committing a serious assault on a girlfriend during an argument.
Mr. S., who was 18 years old, had an IQ of 70 (WAIS-R). He had lived with his parents all of his life, had attended mainstream school with learning support, and had a history of aggression at school, with frequent referrals to Child Psychiatry Services, Child Clinical Psychology Services, and mandatory attendance at the Children's Panel. The aggressive incidents continued after he was 16 years old. He had previously been charged with breaking into a car and aggravated assault and after a further charge of aggravated assault (he had set off fireworks in a public place, resulting in an injury to a child), he was referred to our treatment program. He received a 3-year probation sentence along with the treatment.
Our criteria for inclusion in this report were reasonably cautious given the recidivism information mentioned in the introduction. Participants were required to (a) have an IQ less than 75, (b) have been convicted in an adult court of assault-related offenses, and (c) have the majority of their aggressive incidents committed in the absence of intoxication through alcohol. Because Brownlee (1995) found that re-offending rates almost doubled from 1 to 2 years post conviction, we thought it best to include only those individuals for whom at least 2 years of follow-up data were available. In fact, over 4 years postconviction data were available on all participants.
Many more individuals were treated during the time period of this study. However, others did not fulfill the criteria of inclusion (i.e., they had not been convicted of an assault-related offense; although the individuals may have shown problems related to anger, they had never been taken to court as a result of anger), they committed many offenses while intoxicated by alcohol, they had an IQ over 75, there was not sufficient follow-up data, they dropped out of treatment or had incomplete data sets, or they were treated for significant periods of time as in-patients).
This series of case studies began in 1990, and the same measures have been used throughout. These measures are based on Novaco's (1986) analysis of anger and are similar to, but not identical with, assessments standardized more recently (Jones, Thomas-Peter, & Trout, 1999; Novaco, 1994). Three forms of assessment were used: an anger inventory, anger-provoking role-plays, and daily self-reports of anger. Information on re-offending was also collected through our extensive contacts with the police, criminal justice services, social work services, and health services in addition to criminal records.
The Anger Inventory is split into six sections that cover anger related to frustration, disappointment, jealousy, embarrassment, anger towards self, and direct assault. Each section has a number of questions related to that emotion. For example, “How would you feel if your girlfriend/boyfriend went out with another boy/girl?” Following the initial response, the participant would be asked, “Would you feel jealous or angry?” Following this response, the respondent is asked “How angry would you feel—not angry, a little angry, very angry, tremendously angry (scored 0 to 3)? Respondents are given a diagrammatic representation of degrees of anger—four histograms, from a simple line to a large rectangle indicating increasing degrees of anger. Verbal responses to all questions were recorded and the information used later during treatment. Scores were recorded in relation to each question for later analysis.
The respondent is then asked. “What would you do?” Answers are scored as follows: 2 points for an answer indicating physical violence, 1 point for an answer indicating verbal violence, 0 for an answer indicating avoidance of violence. This series of answers is recorded for each question in the inventory. There are 20 questions in total, leading to a maximum anger score of 60 for the question “How angry would you feel?” and a maximum score of 40 for the question “What would you do?” Although this questionnaire has been used consistently throughout the study period, its psychometric properties have not been established.
Anger-provoking role plays
Role plays included two situations that were considered to be generally anger-provoking and one that was specific to the participant involved. The “general” role plays were someone jumping in front of you in a queue and someone calling you names. The specific role plays included being ignored when waiting to buy something in a shop, someone standing in front of you when you are watching television, someone not returning a greeting, and someone criticizing your football team.
All role plays were videotaped and later rated by two observers blind to the conditions of the study. They were rated simply on the amount of anger shown by each participant on a scale from 0 to 4, with 4 representing tremendously angry. Agreement between observers was consistently over 90% (average 92.3%, calculated by the number of agreements divided by the total number of ratings). The maximum score on the role-play assessment was 12.
Self-reports of anger experienced each day
Participants were encouraged to complete an Anger Inventory on how they felt during each day. This was in the form of a weekly diary with a line to be rated for each day of the week on a scale of 0 to 3, with 0 representing no anger today; 1, a wee bit angry today; 2, pretty angry today; and 3, very angry all day. Each line had a pictorial representation of the amount of anger in the form of histograms (similar to those used for the Anger Inventory). If participants had felt particularly angry on any day, they were encouraged to make either a pictorial or written note (using one simple word to help them remember) so that the situation could be discussed at the next session. Unfortunately, not all of them were able to complete these forms consistently for the duration of the treatment period, and, therefore, the information should be interpreted with caution. This was especially true during the follow-up periods, when some participants completed only 3 or 4 days per week of recording.
All assessments were completed at baseline, prior to the implementation of the treatment program: the Anger Inventory over at least a 2-week period, the three role plays, and over 2 weeks for the self-report measures. All measures were repeated after 3 months and 6 months of treatment and at the end of treatment at 9 months. Measures were then repeated at 3 months, 9 months, and 15 months follow-up after cessation of treatment. Fifteen-month follow-up represents at least 2 years after admission to treatment. For Mr. A. and Mr. M., there were further follow-up assessments.
Information on re-offending
Information was also collected on re-offending. As previously noted, we have a very comprehensive network of contact with the police and the criminal justice service. We also receive reports from social workers and community nurses who visit clients' homes. This information was reviewed routinely every 6 months, except when indicated otherwise in the Results section. Therefore, there was a very close monitoring of both offenses and incidents that may have occurred but were not prosecuted.
Treatment lasted 9 months, and sessions were conducted weekly (around 40 sessions). Each session lasted 40 to 60 minutes, depending on the concentration levels of participants. Participants were seen in groups, although not all were in the same group. Because of the intellectual deficits of the participants, we felt that it would be easier and more sensible to begin treatment with the relaxation/arousal reduction phase, which has high face validity and less dependence on awareness of the concepts involved in treatment of emotion. Behavioral relaxation (Lindsay & Olley, 1998) was used for the first two treatment sessions, followed by abbreviated progressive relaxation. Once the methods and skills of relaxation were acquired by each individual (typically 3 to 4 weeks), the feelings of relaxation were then linked to images of anger-provoking situations:
You are in a supermarket queue and you are in a hurry. You look around to make sure you have got everything and somebody moves in front of you in the queue; you begin to feel extremely annoyed and angry and you want to shout at them to get to the back of the queue. Now I want you to continue to feel relaxed and calm as you are just now.
Next, feelings of relaxation were increasingly linked to anger-arousing situations specific to each participant. The arousal reduction phase lasted from Sessions 1 through 8.
Discussions on appropriate and inappropriate anger and the link between behavioral, cognitive, and physiological systems of emotion were then introduced (Sessions 8 through 12). During the earlier sessions, there were simple discussions about the purpose of anger, when anger is appropriate, when anger is inappropriate, how your body feels when you are angry (physiological systems of emotion), and what you do when you are angry (behavioral system). These discussions began around Session 5 and at first were illustrated with fairly concrete examples. As the concepts became more complex, the different systems of emotion were linked. Participants considered “angry thoughts” that they held about people and places and how these thoughts might “get them worked up,” that is, make their body feel agitated (cognition and physiology). If you start swearing and shouting or kick the furniture, you will feel agitated (behavior and physiology) and might be thinking angry thoughts about someone (behavior and cognition). All of these discussions were recorded and illustrated using a flipchart.
The discussions about emotion became more complex as sessions progressed (Sessions 13 through 30), leading to an analysis of the emotion of anger in accordance with that described by Novaco (1986). As participants became more familiar with the topic, the relationship between anger and different emotions, such as jealousy, embarrassment, disappointment, frustration, and anxiety, was discussed more fully. This analysis was conducted by talking about the way in which certain situations could be misconstrued as anger-provoking when another emotion or the control of aggression might be more appropriate. At no time during the course of treatment did the group leader suggest that anger was a bad emotion. Feelings of anger were always said to be appropriate in certain situations. However, when the consequences of anger for participants was discussed, the group usually concluded that physical aggression was not helpful.
Occasionally (from approximately Session 13, depending on how well the participants had grasped the principles of anger), we used pictorial or filmed aids to help them identify emotions. One particularly useful method was to tape an episode of a popular soap opera. Having tried a number of them, we found that the most useful were the ones where the emotion is more salient and somewhat less subtle. Where there was an angry or aggressive response, the situation could be analyzed in terms of how appropriate the anger response was.
Role plays of general situations and specifically relevant situations were also used to illustrate and understand aggression (Sessions 16 through 40). Role play situations were organized, conducted, analyzed, and practiced in a similar way to the problem-solving and identification of emotion situations. If participants experienced high rates of anger, then the role play would be stopped and relaxation procedures and discussions would ensue, with an emphasis that this is precisely the emotion that the individual has to control if he is going to stay free from trouble in the future. The emphasis in the discussion is not that these feelings are bad or unnatural, but that if they lead to an excess of aggression, then this will not be helpful to the individual or conducive to adaptation in his home or community setting. In this way each participant was encouraged to develop personal, practical skills for controlling aggressive emotions.
Results and Outcomes
The scores on Table 1 show the responses to the question, “How angry would you feel?” The total possible score for this scale is 60. At baseline, it can be seen that the scores ranged from 21 to 46 (average = 34.8). Following the relaxation phase (up to 3 months), there were reductions of around 10 points in response to the Anger Inventory for all participants (average = 23.3). This trend continues to assessment at 6 months, with no further significant reductions on this aspect of the Anger Provocation Inventory throughout the remaining posttreatment and follow-up assessments.
Although there is a general pattern of a decrease in anger scores on the Emotion Scale, it was not uniform across participants. The scores of Mr. M. fell by 3 months and continued at a low level for all further assessments during treatment and follow-up. On the other hand, the scores of Mr. A., Mr. C., and Mr. S. showed a more gradual reduction, without falling to very low levels at any point during treatment or follow-up assessments. This would suggest that the men were simply not responding in a socially acceptable fashion, with 3 of them indicating that they would continue to become angry in certain situations. For Mr. A. and Mr. M., assessment at 21 and 27 months showed no significant change to the pattern established during the previous follow-up assessments.
Table 1 also shows the responses on the Anger Inventory to the question, “What would you do?” The total possible score on this scale was 40, and the average baseline across participants was 18.1. This reduces to 10.2 at the 3-month assessment, with no further significant reductions throughout subsequent assessments. For Mr. A. and Mr. M., scores at 21- and 27-month follow-ups are similar to scores in previous follow-up assessments.
Table 2 shows the amount of aggressive responding made by participants during videotaped role plays at each phase of the study. It can be seen that they responded with scores in the upper half of the scale at baseline (maximum score is 12). There was not a notable reduction after the relaxation phase of the study, except for Mr. M. (3-month assessment). Thereafter, the participants seemed to respond with some aggression to each situation, but, in general, the responses were rated as being in the lower end of the scale. Therefore, they were not responding with the high degrees of aggression that they had at baseline and assessment at 3 months.
Table 3 shows the total weekly ratings for each participant (averaged or prorated) and rounded up for simplicity of analysis. Complete data sets were available from participants until the follow-up assessments, where Mr. A., Mr. B., and Mr. C. submitted only 3 or 4 days of ratings for the week's assessment at 15-month follow-up. The total possible score at each data point is 21. It can be seen from the total scores that there was a general drift down in the amount of daily anger reported by the 6 participants.
As has been mentioned, all participants were followed-up routinely and systematically with an extensive network of contacts. The area in which we serve our clients is geographically distinct, and clients tend not to move across service borders. Therefore, any known incidents of aggression would have likely been recorded.
Mr. A. continued to live in Salvation Army hostels and similar establishments. Although we made several attempts to help him control his drinking, we were never successful, and at the time of this writing, he continued to spend all of his available money on alcohol. However, there have been no reports over subsequent years of any violent or aggressive behavior, and he has committed no further offenses of any nature. Based on systematic record searches and anecdotal information, it has now been 10 years since his last offense.
Mr. B. lives on his own in the same area where he previously lived with his wife and children. It is not known when they split up, but they were still married at the time Mr. B. completed his 2-year period of probation. He refused to continue involvement with the intellectual disability service system after his probation finished, but there were no reports of further aggressive incidents during that time. Towards the end of treatment, Mr. B. gained supported employment with the local recreation and parks department. He maintained this job for almost a year, losing it because of unreliability rather than aggression. It is currently 9 years since his referral to the service, and there have been no further reports of either offenses or aggressive incidents.
As of this writing, Mr. M. continues to live with his mother, and there have been no further offenses 5 years after his initial referral to the service. During the treatment phase, he gained and has now maintained supported employment, and although there have been reports of him getting extremely agitated there, he has never been violent.
Mr. C. continues to live with his parents, and there have been no reports of violence or further offenses since his initial referral to the service 10 years ago. At the beginning of treatment, Mr. C. had secured a part-time job in a local farm shop. This is seasonal employment, which he has maintained ever since.
Mr. I. committed two further theft offenses during his period of probation. These offenses occurred within the first year of his referral to the service and did not involve aggression. Around 6 months after the initiation of treatment, he settled with a new girlfriend. This relationship has continued for the last 3.5 years, and they have recently had a child. He also resumed contact with his parents. There have been no reports of violence from his girlfriend, and when talking to social workers, community nurses, and when visiting us, Mr. I. quite freely attributes this to the techniques that he learned during anger management training. It has been 4 years since his conviction and referral for treatment.
Mr. S. has continued to live at home with his parents. During routine visits with the community nurse, his mother reported feeling much more assured about his behavior, but this may simply be due to services being available rather than improvements in Mr. S.. Six months after his conviction, he drank an excess of alcohol, stole his father's car, and committed eight traffic offenses. At the time of this writing, it has been 4 years since this incident, and there have been no further offenses. When treatment finished, Mr. S, went to a local employment agency. He successfully held a job portering in a local institution, to the extent that he began going on social functions with the staff. Unfortunately, he was terminated after a year when changes were made to government benefit regulations, making it financially unfeasible to continue his employment. There were no reports of aggressive incidents at work.
Perhaps the first thing to note is that none of the participants has been charged again with an assault-related offense. One man re-offended with two thefts; and another, with driving offenses. Neither has re-offended in the subsequent 3 years. Therefore, although the numbers are small, there is some indication that re-offending may have been less than it would have been given a simple probation sentence without anger management treatment. Therefore, it appears that the hypothesis that two thirds of the individuals would re-offend with an aggression-related offense within 2 years was not supported. Publication of this report has been withheld until sufficient numbers of individuals were available to illustrate the effects of the treatment and to ensure that re-offending information was sufficiently convincing over a long period of time. Therefore, 3 of the participants were followed-up for 10 years following their offense, 1 for 5 years, and the other 2 for 4 years. Within the context of treatment on work for offenders, these are indeed long follow-up periods. None of the individuals has re-offended 3 years following their last conviction for any offense; this result compares favorably with those from the studies outlined in the introduction, in which investigators found re-offending rates of between 40% and 70%. Quinsey (1998), in a study of sex offenders, argued that
from the viewpoint of social policy, measures of offending provide the data of interest. Proximal measures of treatment outcomes are of interest only to the extent that they predict subsequent re-offending. Process measures are similarly of no social policy interest in themselves (p. 66, italics added).
In terms of response to treatment for anger and aggression, all of the men have shown considerable progress in some areas related to aggression. However, it is notable that treatment has not produced compliant, trouble-free individuals. In their self-report data, all of the men continue to note experiencing significant feelings of anger 24 months after the commencement of treatment. Reports of 4 of the individuals at 18 and 24 months follow-up show little difference from baseline levels, suggesting some validity to the data in that participants were not reporting low daily levels of anger as might be expected if they were trying to return socially acceptable responses. These reductions (improvements) in scores could simply be the result of learning how to respond in a socially acceptable fashion either to the Anger Inventory or the role-play situation. This, however, is not reflected in the self-ratings, and scrutiny of the individual profiles does not persuade us that there has been a significant reduction in the feelings of anger experienced by these men.
The behavior of these participants during aggression-inducing role plays indicates that all of them still demonstrate some aggressive responses, although much less than during the baseline assessments. One obvious criticism of such a measure is that role plays are employed during treatment, and participants could simply be becoming familiar with the demands of role-play techniques. However, one might expect their responses to be reduced to zero in such a scenario; nevertheless, this remains a possibility. In addition, results do not suggest that only one measure should be considered to reflect a successful treatment. Rather, a weight of evidence is presented for readers to consider in judging the course of treatment and effectiveness of outcomes. Perhaps the most convincing self-report evidence is that on the Anger Inventory, although caution must be recommended because once again familiarity with the materials may be a confounding variable. All of the participants showed some improvement in their self-perception of the way they respond to anger-provoking situations. Here again the responses are by no means uniform, and all participants retained at least one and, in some cases, several areas where they indicated they would become extremely angry. These findings were consistent throughout treatment and at 24-month follow-up.
Taken together, the results and outcomes of the present study become more convincing. The results of the Anger Inventory are consistent with the reduction in anecdotal reports of aggressive incidents over subsequent years for all participants. The results of the role-play situations are consistent with the fact that no further incidents of serious aggression by the men have been reported 4 years following their last conviction.
Methodological shortcomings limit the confidence that can be placed on the current results. We used an AB experimental design, which does not allow for experimental control using multiple baselines across participants. Because the assessments employed were not standardized, it is difficult to judge the significance of the reported changes. The treatment phases are not “pure,” with aspects of cognitive appraisal being introduced into the relaxation phase. Corroborative evidence (number of convictions and aggressive incidents) is unreliable, although we believe that in this case the sources of evidence are extensive. Although employment success was not one of the original research issues, it is interesting to note that 4 participants were able to obtain supported employment (Mr. B., Mr. M., Mr. C., and Mr. S.).
The self-reports of feelings of anger show that this group of individuals have not changed completely. As mentioned previously, these data should be viewed with some caution, but data do indicate that although these participants continue to experience significant amounts of anger, they are dealing with it in a more adaptive way and are not engaging in behavior that would lead to social and legal recriminations.
Authors: William R. Lindsay, Professor, Head of Clinical Psychology Services, and Chair, Learning Disabilities, Clinical Psychology Department, Wedderburn House, 1 Edward St., Dundee, DDI 5NS, Scotland ( email@example.com). Ronald Allan, Fiona MacLeod, Nicola Smart, and Anne H W Smith. Requests for reprints should be sent to the first author