The ATP
The results confirmed our hypotheses about significant group differences in the attitudes towards prisoners, as held by prisoners, prison employees and college students. While ATP in the whole sample differed significantly according to age, gender and group, in multivariate analysis only group remained significantly correlated with ATP, when controlling for age and gender.
As expected, the most positive attitudes were held by the prison inmates themselves. If not of an unrealistic nature, such positive attitudes should be able to enhance the effectiveness of the various relapse prevention programs offered in the correctional setting. It might, however, be problematic if those intended to help the inmates in the rehabilitation process, i.e. the prison officers, hold substantially different attitudes than the prisoners themselves do.
Remanded prisoners held less positive attitudes than convicted prisoners. Maybe remanded prisoners identify less with prisoners in general and accordingly hold attitudes more equivalent to those found in the general population.
Among prison employees, a couple of important associations emerged and were retained in the subsequent multivariate analyses. Firstly, attitudes varied significantly relative to the gender of the inmates, so that employees in female-only prisons held more positive attitudes than those employed in male-only prisons. As the female prisons were smaller than the male prisons, we can not dismiss the possibility that the more positive attitudes held by employees in the female prisons at least partly were due to prison size. The differences are contrary to Murphy and Brown who in their general population study from 2000 [
15] found no significant difference in attitudes towards male and female offenders, as measured by the ATP.
The prison officers' attitudes were fairly negative and comparable to the attitudes held by a number of college students. Within the group of college students, however, those studying business economics held significantly more negative attitudes than those studying nursing. Thus, students that had chosen a "caring" profession held the most positive attitudes. Again, our findings are in disagreement with the findings of Murphy and Brown [
15]. They compared mean ATP scores in people with "female", "masculine" and "neutral" occupations and found no significant differences in attitudes.
Although the original ATP scale was developed in another country, the USA, and more than 20 years ago [
2], those results and ours are remarkably similar. In their original work Melvin et al found a mean ATP score of 109.5 among prisoners; we found 106. They found a mean ATP score of 90.7 for correctional officers; we found 90. The original work had included a group of people working in various prison reform and rehabilitation settings; on average, these respondents scored 108.3 on the ATP. Our more heterogeneous group of prison employees scored somewhat lower, on average 98, but even so significantly higher than the prison officers. Melvin et al's student sample averaged 90.5 on the ATP; ours averaged 91. The close correspondence between the two studies lends support to the validity of the ATP.
The ATP has also been used in countries outside the USA. Hogue [
10] used it to assess the attitudes held by correctional staff in Great Britain. Although the relative ranking of the mean ATP scores in the different groups investigated were similar to ours, the British scores were overall somewhat lower. A Spanish study [
12] demonstrated substantially lower scores in correctional officers than in prison rehabilitation team members. The same Spanish study found students' mean ATP scores similar to those found in our student sample, while a British population study [
15] found somewhat lower mean ATP scores.
The above studies have all used the ATP. However, a number of studies using other instruments than the ATP merit attention. These studies have demonstrated that older prison officers seem to focus more on the rehabilitative aspects of their work [
7,
14,
22] and hold more optimistic attitudes towards the inmates than younger officers do [
8]. The findings regarding the effect of the prison officers' gender on attitudes have been inconsistent; with either no effect [
8,
9] or female officers holding more positive attitudes [
7]. In the multivariate analyses carried out in our study, age and gender did not contribute significantly. While other studies have reported inconsistent results regarding the effect of the amount of previous correctional work experience [
4,
6‐
9,
11], we did not find any significant effect of work experience on the mean ATP score.
There is a paucity of studies reporting on the attitudes of correctional personnel in administrative positions [
13]. It seems to be fairly unique that our study is able to report on the attitudes held by prison officers and other employees in the same prison facilities. As already stated, we found that prison officers held more negative attitudes than employees in other work positions. It might be that working with prisoners on a day-to-day basis leads to more negative attitudes. As the main task of the prison officers is to supervise and control the inmates and their behaviour, they must to a certain degree focus on negative aspects of the individual. Other prison employees, particularly teachers and people running rehabilitation programs, are expected to focus more on the inmates' resources and potentials. In addition, they have at their disposal programs and treatments that may provide a sense of empowerment and optimism. Together, these circumstances may result in more positive attitudes.
It is important to acknowledge that attitudes do not necessarily translate into action [
18]. Furthermore, some of the attitudes conceived as negative according to the ATP may in fact represent a fair evaluation of some of the prisoners. Some offenders may be better served by a realistic assessment of their problems rather than by unjustified optimism as to their capacity for change. Hence, both extremely low and extremely high ATP scores may be counterproductive, at least in a forensic setting. Even so, the extremely wide gaps between the highest and the lowest ATP scores within all groups investigated (Table
1) are noteworthy. Intuitively, it seems more reasonable that students would present a wider range of scores than prison officers who have had more uniform experiences with prisoners through their work.
The questionnaire
As shown in Tables
3 and
4, the three groups investigated differed significantly in their answers to a number of the questions. Even so, some questions received remarkably similar answers, like the ones about the percentage of crimes that are punished and the high expected re-imprisonment rate. The latter may testify to a general distrust in rehabilitation programs across all groups. The prisoners gave the highest estimates about the number of the general population that has been incarcerated. This may partly be explained by imprisonment being more prevalent in their home environment.
All, including the prisoners themselves perceived prisoners as a psychologically deviant group with frequent drug and alcohol abuse and other mental problems. These beliefs have implications for the attitudes held towards prisoners, particularly among prison officers [
16].
Both prisoners and students estimated the proportion of prisoners that had committed sexual offences to be higher than the prison officers did, with the prison officers' estimates being closer to the national average of about 7% [
23]. The groups' mean estimate of the percentage of prisoners having committed a violent crime, 28–34%, is slightly above the national average of about 25% [
23].
Not surprising, most inmates were of the opinion that the crime punishment level in Norway is too severe. However, as many as 1 in 3 found it about right and some even too mild. Most prison employees found the punishment level adequate or too mild, while the students expressed the most punitive opinion: more than 3 in 4 found the punishment level too mild. Unsolicited, a fair number of respondents from all groups had indicated in the margin of the questionnaire that they found the punishment level to be too severe in cases of drug related crimes and too mild in relation to sex offences. This may indicate that attitudes towards prisoners most likely vary, according to the crimes they have been convicted of.
Although prison inmates and prison employees, both of them groups that hold first-hand knowledge about the issue, agreed that our prisons do not function as they should, only a small minority in both groups believed that we would be better off if all prisons were shut down. The vast majority of students expressed satisfaction with the way our prisons function. This may be taken to indicate that, as long as our prisons keep offenders locked up, most people are of the opinion that our prisons function according to intention.
A number of the answers to the questionnaire (Tables
3 and
4) correlated strongly with the ATP, particularly in prison officers, and invariably in such a way that those who expressed more pessimistic and punitive views had a lower ATP score. This finding testifies to the close relationship between a person's attitudes towards prisoners and the same person's general beliefs regarding prisoners and crime and punishment in society.
Limitations of the study
In assessing the results of the present study, a number of limitations must be taken into account.
A comprehensive study of attitudes towards prisoners should adopt a broader, more diversified approach. In particular, more than just one instrument should be used.
As always in this kind of studies, one is confronted with the question about statistically vs. clinically significant findings. We have no way of stipulating what would constitute clinically significant group differences in the ATP scores; all we can do is caution against making too strong inferences regarding the results presented.
While attrition was very low among college students, it was higher in the prisons, particularly among inmates in one of them. Unfortunately, we had no opportunity to conduct proper analyses of attrition. Some prisoners were on leave, either for social reasons or because they had to appear in court. In addition, we have missed prisoners with mental or somatic conditions rendering them unable to participate and all prisoners that were not able to answer in English or Norwegian. Hence, the sample is not representative for those groups of inmates.
The differences in attrition rates between prisons could be problematic if there was a link between negative attitudes and high attrition rates, with employees in some prisons being of the opinion that it is a waste of time to do research on prisoners. We do not have any indication that this was the case in our study.
Correctional institutions may differ in a number of ways: some are well run, some are chaotic; some have a stronger focus on rehabilitation, others on punishment. Such factors may influence attrition rates and the attitudes held by employees and inmates. These issues have not been addressed by us but should be explored in future studies.
Although ATP scores for a group of college students may be representative for young people, they should not be taken to indicate the attitudes held towards prisoners in the population at large. Ideally, respondents should be drawn at random from the total population and be large enough to allow the exploration of how gender, age, socio-economic status, crime victimization and education influence people's attitudes towards offenders. However, a group of college students is an interesting comparison group to prison inmates. They may in some respects be considered complementary groups: as opposed to young prisoners, most students have been successfully integrated into a society from which the prisoners have been forcefully excluded, but will have to re-enter after serving their sentence. On the other hand, one would expect college students and prisoners to have markedly different socio-economic backgrounds. This could be an important confounder.
Lastly, the ATP was developed for use in the USA. Even though our results are remarkably similar to those of Melvin et al's original work, we have to take into account that there might be differences in culture and crime policy between the USA and Norway that makes the scale less suitable to measure what it originally intended.