Abstract
The concepts punitiveness and rehabilitation orientation in the general public are generally measured by rather broad attitude items that are not directly related to probation. In this study, two separate attitude scales were used that were tailor-made for the probation context and therefore have a high ecological validity. These ‘ecological scales’ were each analysed with the same predictor set. Cognitive emotive variables showed incremental prediction above demographics. Higher knowledge of probation and more satisfaction with society are related to a higher preference for rehabilitation. Less knowledge of probation and a higher feeling of victimization are related to a more punitive attitude.
Keywords
Introduction
It is important that citizens have some understanding of effective criminal justice practices. As has already been mentioned by Maruna and King (2004), a degree of public acceptance and confidence in the criminal justice system is crucial for its well-functioning. Also the role of probation should not be neglected here, as this institution forms an important interface between the criminal justice system and society. To shed some light on these matters, we conducted a survey study in which we probed respondents' opinions on a range of issues that are relevant in probation in the Netherlands today.
During the last decades, a number of studies have been conducted on public attitudes with respect to criminal justice attitudes in general (e.g. Cullen et al., 2000; Roberts and Indermauer, 2007; Roberts and Stalans, 1997). Maruna and King (2004) indicate that the general conclusion from these studies seems to be that the general public has limited knowledge about the workings of the criminal justice system, crime rates and the basics of criminology. With respect to probation in particular, they mention a survey conducted in England and Wales (National Probation Service, 2002) indicating that 43 per cent of the respondents consider themselves to be hardly informed about what the probation service does. Only 7 per cent of the public indicated they have extensive knowledge about the goal and activities of the probation service.
As for the situation in the Netherlands, along with the advent of populist politics (e.g. Kriesi and Frey, 2008), we have seen an increasing demand for harsher sentences in the media in this country. Thus, the retributional standpoint with respect to issues of criminal justice seems to be on the rise in Dutch public opinion. On the other hand, professional probation organizations and the academic world have continued to focus on rehabilitation, because they consider this angle a more effective approach in reducing crime levels. See also Andrews and Bonta (2006) who emphasize in their Risk, Need, Responsivity model that it is important to target criminogenic needs and provide treatment to reduce recidivism. Also, these authors indicate that interventions should be embedded as much as possible in the regular context of the offenders’ lives, which is a plea for a community-based approach of crime prevention in which the offender is not treated in isolation, but also families, schools, social networks, mental health services and the labor market take up responsibility in reducing criminal behavior. Therefore, we believe that despite a strong current of penal populism (see Pratt, 2007) that may exist nowadays, at least an undercurrent of penal welfarism or the rehabilitation side of criminal justice will be present in the public opinion.
Also in the Netherlands, professional attention is more aimed now at the person of the offender and on underlying causes of criminal conduct. Thus, the focus has shifted from a crime-related to a more person-focused response, taking into account elements of the What Works approach (Ministerie van Justitie, 2010). Because this approach aims to reduce risk factors and enhance protective factors, it will lead to increased effectiveness of sanctions.
As opposed to the above treatment or rehabilitation-based views on crime, we have seen for decades already a tendency for criminal policy to become more victim-centered. Since the 1960s and 1970s, there has been a more punitive and hard-line approach to crime. This development is described by several authors as the ‘punitive turn’. As is described for example by Demker and Duus-Otterström (2008), political parties in several European countries have politicized criminal policy and called for policies that are tough on crime. Garland (1996, 2000, 2001) argues that we have witnessed the abandonment of the model of penal welfarism, i.e. the pragmatic and treatment-orientated attempt to improve the social conditions of offenders. After this punitive turn the focus is now more on the consequences of crime than on its causes.
Roberts et al. (2003) argue that we see a rise in both the rhetoric and practice of severe punishment. According to these authors, these criminal justice policies are designed to appeal to the public appetite for punitiveness, but do little to reduce crime. Because of its emphasis on increased imprisonment and severe sentences, penal populism is also referred to as ‘the New Punitiveness’. It should be noted here that there is very little evidence that increased incarceration has reduced crime (e.g. Reis and Roth, 1993). Thus, we see a discrepancy between the person-focused and rehabilitation-oriented approaches based on the Risk, Need, Responsivity model that are advocated and practiced by criminological experts and the rhetoric driven by penal populism that has an important role in politics and the media. The person-focused approaches are evidence-based and have a potential to reduce crime, because they focus explicitly on the criminogenic factors that explain the development of criminal behavior, whereas penal populism is not focused on the causes or the prevention of crime but mainly on winning votes. As is indicated by Pratt (2007) and Demker and Duus-Otterström (2008), it largely ignores actual crime rates and, instead, concentrates on the political and ideological discourse about crime. McNeill (2011) gives an overview of the implications of the current social conditions for probation agencies.
The implicit hypothesis that increased sentences lead to less crime does not seem to be borne out. As is indicated by Roberts et al. (2003), by the time that research has demonstrated the futility of many of these ‘solutions’ of crime, the parade has moved on and the election is over. They also highlighted that, ironically, the trend towards harsher sentencing has occurred in a period in which crime rates have been declining in the five countries they investigated.
The literature also shows an ongoing discourse on the factors that explain punitive attitudes. As is indicated by Pratt (2007), penal populism can to some extent be explained because criminals and prisoners are assumed to be favored at the expense of crime victims and the law-abiding public in general. According to Garland (2001), the new punitiveness is a result of the greater personal freedom and lower social control associated with late-modern societies, combined with a greater fear of becoming victimized by crime among middle-class people. Maruna and King (2004) investigated whether punitive public attitudes are derived from direct experience of crime and from victimization. They concluded from their large-scale survey that factors such as ‘concerns about the economy’ and the ‘state of the youth today’ account for a substantial proportion of the effect of actual crime concerns on punitiveness. Crime-related factors, such as victimization experiences or anxieties about crime do not appear to be strong predictors of punitiveness in their sample.
Johnson (2009) and Costelloe (2002) conclude that punitiveness also seems to have an affective component. Anger about crime is a significant predictor of punitive attitudes in their studies. On the other hand, the findings of Van Kesteren (2009) concerning the victimization-punitiveness link are negative. He reports a modest increase in support for imprisonment among those who had been a victim of either a burglary over the past five years or a contact crime. His data showed no relationship with previous victimization when controlling for external factors. In general, the literature seems to yield a mixed picture of the relationship between victimization and punitive attitudes. On the one hand, there seems to be little support for a relationship between direct experiences as a victim of crime and punitiveness. On the other hand, Maruna and King (2004) suggest that general concerns about society, which can in our opinion also be crime-induced, play an important role in the prediction of punitive attitudes.
As is indicated by Pratt (2007), the role of expert knowledge has changed during the shift from penal welfarism or the rehabilitation approach to crime to the penal populist standpoint. In the past, the criminological debate was dominated by a highly educated elite of experts who based themselves on scientific argumentation and statistical data. In line with increasing populism, during the last decades the general public has been wanting to be involved in opinion formation themselves, rather than allowing elites to do this for them. This shift is reflected in the media that, in the increasing competition for audiences, need to make programs that are attractive to a mass audience.
Still, Roberts and Indermauer (2007) concluded that criminal justice knowledge is one of the most important predictors of punitive attitudes in their survey. They report that the strongest predictors of punitiveness were criminal justice attitudes, among which they classify crime perceptions. These variables increased the amount of variance in punitive attitudes that could be explained and accounted for more variance in punitive attitudes than demographics, political orientation, religious attendance or media use. Accordingly, they report that inaccurate perceptions of crime are related to more punitive attitudes. They give the following example of false perceptions of crime. In Australia 70 per cent of survey respondents believe that crime has increased whereas, in reality, it has stabilized or decreased (Indermauer and Roberts, 2005). Also, the public has false expectations about what sentencing can deliver in terms of crime reduction and containment (Roberts and Stalans, 1997).
Far less studies have been done to investigate the determinants of rehabilitation orientation. We could only find the study by Maruna and King (2004) who predicted attitudes with respect to community sanctions (i.e. a form of rehabilitation) based on criminal justice attitudes and core beliefs and values (about crime).
In this article we aim to approach the domain of penal attitudes from an integral framework. This means that, contrary to the above mentioned studies, we do not want to restrict ourselves to predicting punitiveness or rehabilitation-focus separately. We were inspired by the results of Applegate at al. (2000) who show that when forgiving attitudes are measured beside punitive attitudes, the former will prevail. In this sense, we adhere to a bifurcated approach of criminal justice attitudes. We conducted a survey study in which opinions were collected on a wide array of topics that are relevant for probation in the Netherlands at this moment. Another difference with the above authors is that their predictors are in part closely related to their criteria, because they used attitudes about crime to predict other attitudes about crime. In our opinion this creates artificial overlap between the predictors and the criteria. Accordingly, apart from demographics, we took into account the role of victimization, satisfaction with society and knowledge as relatively un-confounded predictors of probation-related attitudes. In this article we will address two main research questions:
What accounts for individual differences in punitive and rehabilitation-orientated attitudes towards probation? More specific, what is the role of feelings of victimization, knowledge about probation and satisfaction with society? Is the picture similar for punitiveness and rehabilitation-orientation?
We did not use existing, general questions for the measurement of probation-related attitudes. Instead, we explicitly aimed to take into account the probation context by incorporating real-life issues from the work domain into the attitude scales. Therefore, we conducted qualitative sessions with a group of experts on probation. From the qualitative material potential survey items that are meaningful for the probation context were derived. Based on multivariate analyses, custom-made or ecological scales with a high face validity and ecological validity (see Brewer, 2000) were constructed. Thus, we used a combination of qualitative and quantitative research methods (e.g. Butter, 2011).
As is shown in the conceptual model (Figure 1), we expect that the same sets of demographic and cognitive emotive variables are related to both punitive and rehabilitation-related attitudes towards probation. Because of the exploratory nature of this study, we did not have more specific hypotheses on the exact constellation of the relations.
Conceptual model.
Method
Construction of the ecological punitiveness and rehabilitation-orientation scales
The constructs punitiveness and focus on rehabilitation were measured in an ecologically valid way. This means that we wanted to approximate as much as possible the real-life situation that is under investigation, in this case the ‘world of probation’. It will be clear that the general attitudinal items found in the literature, such as for example, the following items by Roberts and Indermauer (2005) did not fit into our approach:
The death penalty should be the punishment for murder; People who break the law should be given stiffer sentences; Judges should reflect public opinion about crimes when sentencing criminals
It should be noted here that the above example highlights that measuring punitivity is not without its difficulties, as is shown by Kury and Shea (2011). We organized a series of qualitative group sessions to develop the ecological criminal justice attitude scales. The group consisted of seven experts on probation issues, either working in the probation context, or involved in teaching on this topic, or both. The facilitator was an expert in ecological scale construction without inside knowledge of probation.
In the first session the probation experts independently proposed themes that are relevant for the context of probation in the Netherlands. Next, the individual contributions were discussed in the group and any overlap was removed. Eight themes were chosen and each expert generated possible scale items for a subset of the themes with which he/she was most familiar. As with the themes, the items were discussed in the group and a selection was made. The following themes were selected. Within parentheses an example item for each theme is given.
Viability of probation (More probation means less fallback into crime); Advisory task of probation (It is important that probation informs the judge of the suspect's personal situation); Supervisory and control role of probation (Probation is not there to help thugs, but to protect law-abiding citizens); Supervision and help with changing behavior (Training to handle money or to find and keep a job enable a smooth return in society); Supervision and morale (Probation should teach people to respect others); Community work orders (The work place should be adjusted as much as possible to the person's qualities); Help after prison (Probation should not pamper persons released from jail; let them find a home and a job themselves) ; Responsibility of probation for their clients' behavior (If a client commits a crime, probation did not do its job properly). Theme 2: advisory task The first task of probation is to advise the court (at least in the Dutch situation). If so requested by the court, the probation worker talks to the person who is suspected of a crime. The person in question, the circumstances and the risk of re-offending are taken into account. Based on this information, probation writes a report advising on the best approach for this specific person.
Each theme was covered by five to six items. Prior to each set of items that cover a theme, a short introduction of the theme was given to enable respondents lacking in knowledge on probation to answer the items. As an example, we give the introduction for the theme advisory task of probation.
As the items were developed with the aim of maximizing the coverage of each theme and not minimizing the overlap between themes, we conducted an explorative factor analysis to shed light on possible structures underlying the 46 items. Oblique rotation was used. The analysis resulted in a two-factor structure, explaining respectively 18.9 and 8.5 per cent of the total variance. After item analysis on the items with the highest factor loadings we constructed two ecological scales, consisting of respectively 12 and 13 items. The first scale is derived from the first factor and could be interpreted as a rehabilitation-orientation scale (α = .83). The second scale is derived from the second factor and could be interpreted as a punitiveness scale (α = .87). The items of both scales and their descriptive statistics are shown in the Appendix.
By comparing for example the Roberts and Indermauer (2007) items that were shown above to the items in the Appendix, it will be evident that the tailor-made scales are much more reflective of the probation context. This means that their ecological validity is relatively high. Hence their denotation as ‘ecological scales’ (Butter, 2009; Butter and Born, 2007, 2011).
Participants and procedure
The data was collected in the spring of 2010 using an online survey. The sample was not representative because our aim in this study was to investigate the relations between variables rather than to provide a picture of the Dutch population in general. A snowball type of sampling was used. The researchers approached respondents from their own personal and professionals networks who also recruited their acquaintances to complete the survey (e.g. Sagalnik and Hekathorn, 2004). We collected responses from 303 participants. After deleting incomplete cases, 280 cases remained.
Measures
The independent measures were completed after the dependent measures (i.e. the items on probation) in order to prevent contamination of the dependent by the independent variables. We collected data on a wide range of independent measures. The descriptives of the demographics were as follows: gender (108 male; 171 female), age in years (M = 44.97, SD = 13.36), marital status (72.5% married or living together), education (86.3% higher education), net monthly income in euros (59.5% > 2000), native country (96.6% the Netherlands), political orientation (62.5% left-wing). Second, based on the literature review, we also measured using a 4-point scale ranging from 1 (totally nothing) to 4 (very much), the ‘amount of knowledge on probation’ (M = 2.36, SD = .71). Third, various aspects of satisfaction were measured using 5-point scales ranging from 1 very satisfied to 5 very dissatisfied. The descriptors were as follows: ‘general satisfaction’ (M = 1.76, SD = .65), ‘work satisfaction’ (M = 2.03; SD = .81), ‘satisfaction with respect to relations’ (M = 1.80, SD =.96), ‘satisfaction with respect to living environment’ (M = 1.65, SD = .73) and ‘satisfaction with society in general’ (M = 2.56, SD = .72). Fourth, we asked for direct experiences as a victim of crime in the past year. The reported numbers were as follows: destruction (22), theft (38), fraud (6), violence (7) and sexual harassment (0). Finally, the subjective feeling of crime victimization was measured using a rating scale ranging from 1 (totally not) to 4 (to a high extent) (M = 1.26, SD = .54).
Analysis
The relationship between the dependent (the punitiveness and rehabilitation-orientation scales) and the independent variables were first explored using bivariate correlation analyses. Next, hierarchical regression analyses (see Cohen et al., 2003) were used to investigate the unique contributions of two sets of explanatory variables (i.e. set 1: demographic variables and set 2: cognitive emotive variables). This technique allows the researcher to assess the additional variance in the prediction of the dependent variables that can be explained by the cognitive emotive variables above the demographics. The same analyses were performed both for punitiveness and rehabilitation-orientation.
Results
The results of the hierarchical regression analysis for punitiveness and rehabilitation- orientation are presented in Table 1. For punitiveness and rehabilitation-orientation, respectively 20 and 17 per cent of variance is explained by the same set of six predictor variables. As can be judged from Cohen (1988) these are medium to large effect sizes, that is, f 2 = .25 and .20 respectively.
Hierarchical regression models for predicting punitiveness and rehabilitation-orientation towards probation.
*p < .05 (two-tailed); **p < .01 (two-tailed); ***p < .001 (two-tailed).
Similar to the findings of Roberts and Indermauer (2007), Table 1 shows only a small effect size for demographic variables. Increased punitiveness is associated with being younger and having a lower education level. For rehabilitation-orientation, the picture is different. An increased rehabilitation-oriented attitude is related to being older. We found no correlation with education level, but a negative correlation with income. Apparently, preference for rehabilitation decreases when income increases.
In the next step, a set of cognitive emotive variables (a combination of questions referring to knowledge, satisfaction and feelings) were included in the prediction model. Adding these variables leads to a significant and sizeable improvement in prediction, yielding moderate to strong effect sizes for punitiveness and rehabilitation-orientation. Increased punitiveness is related to a low level of knowledge about probation and an increased feeling of being victimized by crime. Increased rehabilitation-orientation is associated to a higher level of knowledge about probation and increased satisfaction with society.
Discussion
Applegate et al. (2000) showed that when forgiving attitudes are measured beside punitive attitudes, the forgiving attitudes will prevail. Our study confirms this result in the sense that the factor analysis showed that rehabilitation-oriented attitudes explain more variance than punitive attitudes. This is also reflected in the highly significant difference between the mean scores on rehabilitation-orientation and punitiveness, which are respectively 4.21 and 2.84 on a scale ranging from 1 (fully agree) to 5 (fully disagree) (t = 28.95, df = 2.67, p < .01).
Our main conclusion is that cognitive emotive variables, that is knowledge, satisfaction and anxiety, have an added value in explaining both punitive and rehabilitation-oriented attitudes towards probation. Punitiveness is related to decreased knowledge about probation, and increased feelings of victimization. Rehabilitation-orientation is related to increased knowledge about probation and a higher satisfaction with society.
Before discussing our results in more detail, we will position our study in relation to the articles by Maruna and King (2004) and Roberts and Indermauer (2007), which are the most important reference studies. We used the same set of predictors for predicting simultaneously the constructs punitiveness as well as rehabilitation-orientation. By contrast, both of the above mentioned studies focus on one construct only. We believe our simultaneous approach gives a better insight into the attitudinal dynamics with respect to probation than is the case in a one-sided prediction effort.
Apart from not taking into account the punitiveness aspect in their study, Maruna and King (2004) also concentrate on the public opinion towards community sanctions. Hence, their study is related to the rehabilitation side of probation, but takes a rather limited perspective towards this context. On the other hand, Robert and Indermauer (2007) focus on predicting punitiveness only. That is, they ignore the rehabilitation side in their study. Besides, they work from a general criminological frame-of-reference. As we already indicated in the introduction, this is reflected by the general punitiveness criterion they use. Also, the word ‘probation’ is never mentioned in their article.
With respect to the variances explained, Maruna and King (2004) show a total R square of .40, based on a model with 16 predictors. Regrettably, they do not mention the exact scale that they used to measure attitudes towards community sanctions. However, a number of their predictors seem closely related to the criterion. In this respect, we mention their core beliefs and values ‘crime is a choice’, and ‘belief in redeemability’, which in our opinion share an underlying progressive orientation with the criterion and, thus, have an a priori conceptual overlap with rehabilitation-oriented attitudes, such as the attitude towards community sanctions. It should be noted here that Maruna and King (2009) also included redeemability as a predictor in a model with punitivity as the criterion, using a broader measurement for the dependent variable than Roberts and Indermauer (2007). In this analysis, the regression weight turned out to be negative.
Roberts and Indermauer (2007) demonstrate a total R square of .38, based on a model with 13 predictors. As was indicated above, they use a general punitiveness criterion that is rather sentence-related. Their set of predictors contains the variables ‘right political spectrum’, ‘Liberal/National supporter’, ‘commercial TV’, ‘confidence in courts’, ‘accurate crime perceptions’, ‘confidence in police’, ‘immigrants don't increase crime’ that seem a priori linked to the three criterion items on severe sentencing. One might argue that these predictor items constitute a conservativeness scale together with the criterion items. Accordingly, we think that both preceding studies, apart from not being directly related to probation, also suffer to some extent from inflated R squares due to autocorrelation. We believe that our predictor set, because of its greater distance from the criteria, is less susceptible to this artifact, implying that our R-squares, which are lower, should not be compared at face value to the above values. Apart from the above arguments, it should also be taken into account that both earlier studies hinge on twice the amount of predictors than the present study, which can therefore be considered as more parsimonious.
In the present study, an ecological approach was followed to construct the probation-specific punitiveness and rehabilitation-orientation scales. These ecological scales were derived in a bottom-up way using qualitative group session with a group of probation experts. This implies that the items are rich in content and cover a wide range of issues that are relevant in probation today. Using an environmental parallel, one could say that the items are derived in a slow process from the ‘natural world of probation’. As a result their ecological validity (e.g. Brewer, 2000; Shadish et al., 2002) is higher, than would have been the case if we had used the general items of, for example, Roberts and Indermauer (2007). Accordingly, we think that the ecological scale construction has a positive impact on the meaningfulness of our conclusions for the context of probation.
Next, the role of each of the three cognitive emotive variables, i.e. ‘knowledge of probation’, ‘satisfaction with society’ and ‘feeling of victimization’ will be discussed.
Knowledge of probation
Roberts and Indermauer (2007) concluded that criminal justice knowledge is one of the most important predictors of punitive attitudes in their survey. These authors operationalize criminal justice knowledge by measuring the accuracy of knowledge about crime trends (Over the last two years, do you think crime has decreased or stayed the same?). They conclude that inaccurate perceptions of crime are related to increased punitiveness. Although our results support this conclusion, two remarks should be made. First, their operationalization of criminal justice knowledge seems rather limited. Second, it is not directly related to knowledge of probation.
Maruna and King (2004) do mention ‘the widespread and systematic ignorance about crime and justice’ (see also Durham 1993; Hough and Park 2002; Morgan and Russell, 2000; Tarling and Dowds, 1997); however, their prediction model does not contain explicit knowledge related-variables.
Our analyses show that ‘knowledge of probation’ is the most important non-demographic predictor of punitive and rehabilitation-orientated attitudes versus probation. Possessing more knowledge of probation is related to having a less punitive and more rehabilitation-oriented attitude towards probation.
Feeling of victimization
Maruna and King (2004) concluded that crime-related factors, such as victimization experiences or anxieties about crime, do not appear to be strong predictors of the attitude towards community penalties in their sample. Our analyses confirm this result in the sense that feelings of victimization are unrelated to a rehabilitation-oriented attitude with respect to probation. Roberts and Indermauer (2007) mention that research (e.g. Hogan et al., 2005) shows that individual differences in punitiveness are related to victimization, but they discard this variable in their own prediction model. Our analyses yield a differential picture regarding victimization. It is a predictor of punitive attitudes towards probation, but does not play a role in explaining rehabilitation-oriented views. With respect to direct crime experiences, the literature (Maruna and King, 2008; Van Kesteren, 2009) yields little support in relation to punitiveness. Our preliminary analyses confirm this picture, as we found no correlations between direct experiences with destruction, theft, fraud, violence and sexual harassment and the dependent variables.
Satisfaction with society
Roberts and Indermauer (2007) do not pay attention to well-being related factors in their study. Maruna and King (2004: 93) mention 'general concerns about society that might revolve around ontological insecurity or a widespread sense of anxiety driven by the disembedding processes of modernity that have resulted from the erosion of former social certainties'. Under the title ‘expressive variables’, they include the predictors ‘collective efficacy and trust’ and ‘economic pressure’ in their prediction model. The first of these variables turns out to be positively related to a favorable attitude towards community sanctions. This result is confirmed in our analyses, as we found a positive correlation between satisfaction with society and rehabilitation-orientation. Interestingly, they also mention the possibility that people's opinions on crime and punishment might be related to very personal psychodynamic histories (see Maruna et al., 2004). We found no evidence in our study for the predictive role of more personal frustrations. Both for punitiveness and rehabilitation-orientation, there is no correlation with variables measuring satisfaction with work, personal relations, and living environment.
Implications for practice
Our analyses demonstrate that a higher level of knowledge of probation is related to a lower level of punitiveness and a higher rehabilitation orientation with respect to probation. There is room for improvement here, as in our sample 60.4 per cent of the respondents indicate that they know little or nothing about probation. We think that this is a serious problem for the societal legitimacy and support of probation services. As can be deduced from the introduction, evidence-based practices in probation are more rehabilitation-orientated than punitive in character. Investing in knowledge of probation will lead to more support for effective practices and less support for ineffective practices. Accordingly, it is advisable that probation organizations should make more explicit to the public what exactly they do, and why it is the most effective way to prevent crime. See also Maruna and King (2008) and Allen (2008) in this respect. Although they are not the most easy target group for information campaigns, paying special attention to young people with lower education should be considered. Maybe target group ambassadors, i.e. (former) probation clients, can play a role here by visiting schools or youth centers.
Limitations of the study
As we have already indicated, we did not use a random sample in this study but a convenience sample based on a snow-balling method. This approach led to an underrepresentation of lower educated persons in our sample. Left-wing voters were overrepresented. Thus, it is not possible to generalize the results to the entire Dutch population. This was not the aim of our study, however, as we were more interested in relations between variables than in their absolute levels. It should be mentioned here that the survey held by the National Probation Service (2002) in England and Wales showed that only 7 per cent of the public indicate that they know a lot about what the probation service does. In our sample this percentage is 6.8 per cent, suggesting that there is no serious overrepresentation of respondents who are knowledgeable on probation in the sample.
Footnotes
Appendix
Items of the ecological punitiveness and rehabilitation-orientation scales (means and standard deviations are given in parentheses). The items were measured on 5-point Likert scales ranging from 1 (fully agree) to 5 (fully disagree)
Acknowledgements
The authors would like to thank the following people for their help in this study: Ada Andreas; Jacqueline Bosker; Joyce Cordus; Donnalee Heij; Joram Hoogink; Liesbeth Jongerius; Nel Koning; Lous Krechtig; Annelies Sturm; Anke van Stijn; Jaap van Vliet and Marjet Zeegers.