«FAMILIES OF NATIONS, VICTIMISATION AND ATTITUDES TOWARDS CRIMINAL JUSTICE PAUL A. NORRIS* School of Social and Political Science, University of ...»
Returning to Figure 1, a second distinct cluster can be identified involving Austria, Finland, Switzerland, Norway, Sweden and Denmark. Notably, this cluster involved all four nations which the Families of Nations Typology suggests can be taken together to represent a 'Nordic' family. Figure 3 illustrates that this cluster is characterised by low levels of crime, a low level of concern about crime, less punitive attitudes and possibly less evidence of individuals taking private precautions against crime (the proportion of households purchasing special locks is particularly low). These findings appear to support the argument, made earlier, that the Nordic countries with their high regard for universalism, equality and sense of social cohesion are in many ways the polar opposite of the 'English-Speaking' jurisdictions discussed above.
The remaining clusters, shown in Figure 1, are less distinct than those identified so far. However, they do still appear to fit with the theoretical expectations of the Families of Nations typology. The Netherlands, Germany, Italy, France and Belgium group together. While Austria is not present in this cluster, its membership broadly reflects the Continental Europe family suggested by Castles (1998). Similarly, while soon joining with the Continental Europe family, Portugal, Spain and Greece are shown to cluster together, supporting the expectation that their more recent transition to democratic politics and resultant differences in economic and social conditions may be responsible for a distinct pattern of outcomes relating to crime and justice. The characteristics of these two clusters are shown in Figures 4 and 5. The characteristics of the 'Continental Europe' cluster appear unremarkable, broadly following the means for the sample as a whole. This suggests that these nations sit mid way between the 'English-Speaking' and 'Nordic' jurisdictions, a finding which reflects not only the work of Cavadino and Dignan (2006) concerning penal policy, but also the way these nations are perceived in social policy more generally (centrist politics, moderately generous welfare states and differences in terms of income equality which while present are not extreme).
Comparing the characteristics of Greece, Portugal and Spain (Figure 5) with the 'Continental Europe' Family (shown in Figure 4) supports the expectation that these three jurisdictions may form a separate, 'Southern European', family.
Figure 3. Attributes of the Nordic Cluster
While it is important not to overstate the difference between these two groups, for instance they appear very similar in terms of their attitudes towards punishment; there are some clear differences in terms of experience of victimisation, worry about crime and attitudes towards the police. Two points are particularly worthy of note. Firstly, in contrast to the three clusters discussed so far, the Southern European cluster exhibits a marked difference between the level of victimisation (generally low compared to other countries) and the extent to which people worry about crime (high compared to other countries). This suggests that other, wider social or cultural, issues may be influencing the perception of risk in these countries. Secondly, the 'Southern European' family exhibits a lower level of satisfaction with local policing than any of the other clusters identified in this analysis. Once again, this attitude appears at odds with the countries' experiences of victimisation and it seems plausible that wider social or cultural issues may be at play. For instance, it could be the case that attitudes toward the police in these countries are coloured by memories of their relatively recent authoritarian regimes. Attitudes towards law enforcement may therefore reflect wider concerns around the authority of the state or the role of the criminal justice system within society.
Figure 1 suggests Japan does not fit well with any of the other jurisdictions considered in this study (it is the last jurisdiction to cluster together with other cases). This provides evidence that it may indeed be an outlier in terms of the cases considered in this article, a point reinforced by Figure 6. Generally, a Figure 4. Attributes of the Continental Europe Cluster Figure 5. Attributes of the Southern Europe Cluster jurisdiction's scores in terms of punitiveness and worry about crime appear related to levels of victimisation (compare Figures 2–4). In contrast, respondents in Japan appear to hold punitive attitudes and be highly concerned about the risk of victimisation even though the country exhibits very low levels of victimisation. The finding that Japan does not fit with the pattern linking attitudes to victimisation levels may support the argument that Japanese society encompasses a unique set of cultural values which influence the way crime is perceived. For instance, while the public's punitive attitude may not reflect the risk of victimisation, it could be attributable to how the Japanese often see the needs of society as more important than the needs of the individual, and place high importance on the obedience of authority. Those who commit crime can be seen to have broken both of these conventions and hence are viewed in a poor light. This point may be especially pertinent because the ICVS question refers to a 'repeat burglar'; hence the offender has transgressed important social conventions on more than one occasion.
The overlap between the theoretical families (outlined in Table 1) and the clusters identified in Figure 1 is very strong suggesting that the Families of Nations typology is relevant to understanding differences in victimisation and attitudes towards crime and justice. It is important to note that differences in the levels of victimisation and attitudes associated with each family would appear consistent with the expectations outlined earlier, and can be seen as attributable to the differences between families in terms of culture, economy and politics
Figure 6. Attributes of Japan
providing a theoretical explanation of why particular jurisdictions cluster together.
The identification of separate 'English-Speaking', 'Nordic', and 'Continental European' families marks an interesting contrast with the results of Smit et al.
(2008) where these countries were argued to represent a single cluster. It is not possible to pinpoint the exact reason for this difference; it is most likely a combination of the differences between the two pieces of analysis, for instance the cluster techniques they employ and the indicators they chose to consider.
However, perhaps the single most plausible explanation is that the two pieces of research involve very different samples. The sample considered by Smit et al.
includes not only most of the long-term industrialised democracies considered in this article but, also the new democracies of Central and Eastern Europe.
When considered in this wider context, the criminal justice situation in Scandinavia, Western Europe and North America may appear relatively similar.
However, when the countries of Central and Eastern Europe are excluded from the analysis more subtle differences between the remaining nations become apparent1.
Typologies are likely to be of most use for comparative research when the groupings they suggest are highly internally homogenous, yet distinct from each other. One measure of the extent to which a range of cases are similar is the Coefficient of Variation. This is calculated as the standard deviation of the sample divided by its mean. The smaller a Coefficient of Variation, the greater the similarity between cases. For each of the indicators considered in this analysis, Table 4 shows the Coefficient of Variation for the sample as a whole, and for each of the 'families' suggested by the previous cluster analysis. If the Families of Nations Typology does a good job of explaining the variation between jurisdictions then it should be the case that the Coefficients of Variation for the separate families will be smaller than that for the sample as a whole.
On average, the Coefficients of Variation for the individual 'families' are around 25% lower than those for the sample as a whole (last row of Table 4).
However, this finding is not universal across either 'families' or indicators. The mean Coefficient of Variation for the 'English-Speaking' cluster is the lowest in Table 4, reflecting how this cluster is the most distinctive in Figure 1. For each of the remaining 'families' it is possible to identify at least one indicator where the family's Coefficient of Variation is greater than that for the sample as a whole. For instance, Table 4 suggests that the 'Nordic' cluster is generally very homogeneous (several of its Coefficients of Variation are smaller than those associated with the 'English-Speaking' cluster), however members of this cluster appear much less cohesive in terms of their level of burglary, or their use of burglar alarms. Similarly, the 'Continental European' countries appear to vary substantially in terms of their level of robbery and the 'Southern European' nations are particularly varied with regards to their rates of attempted burglary.
These results illustrate how, while the Families of Nations typology may TABLE 4 Coefficients of Variation for Indicators of Victimisation and Attitudes Towards Crime and Justice Across Different Families
provide a useful starting point for summarising and understanding differences around victimisation and attitudes towards justice, there will always be exceptions which do not fit the overall pattern. Identifying those examples which do not fit expectations is likely to provide useful guidance as to where further, more detailed, research should be focused. For instance, why is it that Austria and Switzerland cluster with the Nordic nations rather than associating with the 'Continental Europe' cluster as is suggested by the Families of Nations typology?
Are the Three Jurisdictions in the United Kingdom Really That Different?
This paper concludes with two brief examples of how the grouping of jurisdictions suggested by the Families of Nations approach may help to shed new light on existing questions in comparative criminology. As outlined earlier, the United Kingdom is made up of three separate jurisdictions each of which has a distinct history, legal definitions, and criminal justice practices. Figure 7 suggests that, across the different dimensions considered in this article, there may be some salient differences within the United Kingdom. Most notably, victimisation rates within England and Wales appear higher than those in Scotland or Northern Ireland, while the population of England and Wales also appears to express greater levels of worry with regards to crime. Less stark differences can also be seen when comparing Scotland and Northern Ireland; for instance the victimisation rates for personal crimes appear higher in Scotland than in Northern Ireland. However, Figure 7 also suggests that it is important not to overstate the differences within the UK. This is best illustrated by the responses to those questions concerning attitudes towards the police and punishment where all three jurisdictions provide very similar responses.
The results of the previous cluster analysis (Figure 1) can help to put the apparent differences across the UK in context. The dendrogram suggests that Northern Ireland and Scotland are very similar (they are placed side by side in Figure 1 and fuse together very early in the clustering process). In contrast, England and Wales appears slightly separate but still joins the 'English-Speaking' cluster relatively quickly. This suggests that when considered in the context of differences between other industrialised democracies, the differences between the jurisdictions which make up the UK, at least, concerning the issues considered in this article, may be relatively small.
This conclusion does not mean that differences between jurisdictions within the UK are unimportant, or that they should not be treated as separate cases in a comparative analysis. It may however provide a useful perspective in which to discuss the salience of differences within the UK.
Figure 7. Radar Graph Comparing the Attributes of England and Wales, Scotland and Northern Ireland Is the United States an Exceptional Case?