Data for crimes detected by the police were acquired for the county of Dorset UK for the 5-year period January 1, , to December 31, In the United Kingdom, crimes detected by the police include both primary detections, which are solved through investigative effort and, in contrast to many other countries, those that are taken into consideration TIC. The latter are crimes that an offender will volunteer information about to the police at the time of arrest. If the details supplied by the offender can be verified, and the information supplied by the offender satisfies other legal criteria, the offender will accept those offenses.
One incentive for an offender to volunteer such information at the point of arrest is that if they are convicted for a series of offenses at the same time, the sentence for each offense may, under some conditions, run concurrently; an offender cannot be convicted of the same offense more than once, so admitting to these offenses may ultimately be beneficial. The legal validity of such crimes means that detection rates for some crimes may be higher in the United Kingdom than those in other countries. Only those offenses where geographical coordinates were available for both the offense and offender, home locations were analyzed.
For some offenses Most studies that apply the discrete choice approach exclude such offenses e. This leads to considerable attrition in the data that may be problematic here, as older and younger offenders may engage in co-offending to differing degrees. For this reason, where more than one offender was involved in an offense, we adopt the approach taken by Bernasco et al.
As with Bernasco et al.
Testing Ecological Theories of Offender Spatial Decision Making Using a Discrete Choice Model
However, the results were identical, and so we discuss them no further. This is a limitation of the method that applies to all studies of this kind but it is a necessary one. The reason is that the approach to analysis requires that all alternatives in the choice set areas that are and are not selected be enumerated for the purposes of parameter estimation. Including all possible choices outside of the study region would require more data potentially all areas in the United Kingdom than were available for analysis.
A total of crimes committed by offenders were included in the analyses. The dependency in the data associated with reoffending has the potential to lead to errors of statistical inference. Specifically, if they are treated as independent choices, the decisions of prolific offenders can disproportionately influence parameter estimation and lead to downward bias in the standard errors. For this reason, as in previous work e.
The unit of analysis selected was the U. For the study area, there were LSOAs, each with a population of around 1, people and about residential households. LSOAs are somewhat smaller than the areas used in most but not all; for example, Bernasco et al. For example, in the Bernasco and Nieuwbeerta study, the areas selected had an average population of 4, and around 2, households.
Thematic map of theft from vehicle TFV aggregated offense and offender home locations for the U. Census Lower Super Output Areas in the study area. Table 1 provides descriptive statistics for the independent variables used. To compute the distances for each of the possible origin-destination flows, we calculated the distance between each centroid and every other. Where the two areas were the same, as would be the case where an offender lived within the LSOA within which they offended, in line with Bernasco and Nieuwbeerta , we estimated the distance that they would have traveled using the Ghosh correction of 0.
Connectivity can be measured in a number of ways. For example, it could be operationalized by examining variation in the extent to which areas have major roads running through them see White, Here, we measure it more specifically by determining which combinations of areas are directly connected via the network of major roads.
- Integrating mainstream criminological theory into the biosocial perspective: An empirical analysis.
- Deals on Classics In Environmental Criminology | Compare Prices & Shop Online | PriceCheck;
- Passing of the Frontier, the;
- Classics In Environmental Criminology!
- Properties of Solid Polymeric Materials, Part ATreatise on Materials Science and Technology!
- Neuromarketing: Exploring the Brain of the Consumer.
In the event that two areas considered i. In addition, as connectivity so derived could simply indicate that two areas are next to each other, we also included an adjacency variable so that we could isolate the influence of connectivity after controlling for adjacency. For mass transit stations, we identified those areas in which there was a rail station.
There were six in the study area. Population turnover was calculated by computing the number of people who had moved into the area in the previous 12 months, divided by the total population for the area, multiplied by This scaling has no effect on the statistical significance of the parameter estimates and merely serves to make the parameters easier to understand.
There was little variation in ethnicity across the study area and for this reason, we did not attempt to estimate how ethnic heterogeneity might influence offender spatial decision making. The index can be interpreted as indicating the probability that any two people selected at random from an area will belong to different socioeconomic groups, with larger values indicating more heterogeneity.
- Sociology International Journal!
- Public Service Improvement: Theories and Evidence?
- Criminology Essays.
- Mobile Channel Characteristics (The Springer International Series in Engineering and Computer Science).
- Charmed Destinies.
- Joel-Peter Witkin: Forty Photographs.
- No. 3 in A Minor, Op. 34, No. 2.
In this case, the index was derived using data from the U. Census, and indicated the extent to which people in an area belonged to the same socioeconomic groups. The data were classified into six groups: a managerial or other professional occupations; b intermediate occupations and small employers; c lower, semi- routine occupations; d never worked or long-term unemployed; e full-time studies; and f other. As with the index of population turnover, to ease interpretation, and for consistency with previous research e. As two basic indicators of opportunity, we included the number of car parks—derived using U.
Department for Transport data—and the number of registered cars and vans in an area, estimated using data from the U. Table 2 shows the results of the conditional logit model, organized by theoretical perspective. It should be noted that with respect to overall model fit, the pseudo R 2 values associated with the conditional logit model are always much lower than those associated with for example ordinary least squares regression models. In fact, McFadden states that R 2 values above. Note that distance measures shown are logged values.
Considering the CPT variables first, as predicted Hypothesis 1 , there was a significant negative effect of distance, with offenders being more likely to target areas that were closer to where they lived. In Table 2 , distance is measured on a logarithmic scale. The rationale for transforming the data in this way is that for example every additional 1 km traveled is likely to be perceived as more important for shorter than for longer trips, and using a log transformation accounts for this. However, the coefficients associated with the logged distances are a little difficult to interpret as one has to think in terms of logged distances, and so for the purposes of illustration, we consider the coefficients for the untransformed data.
In this case, the odds of an adult juvenile offender selecting an area decreases by a factor of 0. In line with Hypothesis 2, the presence of schools appeared to influence the spatial decision making of the younger offenders, for whom they would be potential routine activity nodes. As predicted Hypothesis 2a , this type of facility had little or no influence on the spatial decision making of adult offenders.
In contrast, the odds of an adult offender targeting an area increased the closer it was to the city center. The opposite pattern was observed for juvenile offenders.
Download Classics In Environmental Criminology
This is not entirely inconsistent with Hypothesis 3a, but it is surprising that for the younger offenders there was a trend in favor of them targeting areas that were located further away from the city center, rather than toward it. This could be explained by a variety of factors and we will return to this issue in the discussion section. As predicted, all else equal, the presence of a train station in an area increased the odds that an offender would commit an offense in that area.
However, while the coefficients were positive for both adult and juvenile offenders, only that for adult offenders was statistically significant.
Against Authoritarianism and Punishment
Considering connectivity, as predicted Hypotheses 5 and 5a , the odds of an adult offender targeting an area increased by a factor of about two, if it was connected to the area in which they lived by the network of major roads. In the case of the juvenile offenders, the estimated effects were in the right direction but not statistically significant. The adjacency variable was in the expected direction, but non-significant. Turning to our two measures of social cohesion, the findings also provide support for social disorganization theory.
Considering the two control variables, the number of car parks in an area did not appear to influence offender target choice. The number of registered vehicles in an area did, however, have a small but significant effect, with offenders favoring areas with more registered vehicles. In the present article, we examined offender spatial decision making for a high-volume acquisitive crime that has received little attention in the literature, TFV.
The aim of so doing was twofold. First, to test theories of spatial decision making for a different type of crime that is committed under different conditions to those examined hitherto, thereby testing the generality of the theoretical perspectives considered. Second, to examine particular expectations suggested by theoretical models that had not been tested so far. Assuming that offender spatial decision making for this type of crime is non-random, according to CPT one would expect offenders to be most likely to target areas that are within their awareness spaces and that are the most accessible.
As this type of crime generally occurs outside, where offenders can potentially be seen during the commission of an offense, one would also expect that offender perceptions of risk, such as that associated with social cohesion, might also have a part to play in their selection of areas within which to offend. Using a discrete choice framework, we find evidence to support both theoretical perspectives.
For our sample at least, offenders were more likely to target areas that were close to where they lived, and that were likely to include routine activity nodes of importance to their age group—schools in the case of juveniles, the city center and rail stations in the case of adults. As predicted, given their relatively increased likely mobility, although they tended to commit offenses close to their home area, adult offenders were found to travel further distances to commit offenses than their younger counterparts.
Their spatial decision making also appears to be influenced by how easy it would be to travel from their home location to potential destinations via the network of major roads, as predicted. That this was not the case for the juvenile offenders warrants further attention. One potential explanation for this finding is that, for our sample at least, the crime trips made by juveniles were simply too short for the road network to play an important part in shaping their target choices.
The influence of the road network on adult offender spatial decision making is a particularly novel result and this finding, along with those concerned with the impact of routine activity nodes on offender spatial decision making, has potential policy implications for those involved in the development of urban spaces.
This is the case insofar as the results suggest that the design of environments, in terms of where routine activity nodes are located and how areas are connected, may not only shape opportunities for crime to occur but also which opportunities offenders are most likely to exploit. To elaborate, the findings suggest that all other things being equal, when faced with the choice of which of a set of areas to target, an adult offender is more likely to target those that are more accessible via the street network.
The results also suggest that the placement of new facilities that might attract people to an area could increase the probability with which crime will occur in them. Thus, our findings suggest that urban planners involved in the building of new developments, or in extending or making changes to the road network, should consider the potential influence of their decisions on the risk of crime in an area.
Conducting a formal crime impact assessment for a further discussion, see Bowers, ; Ekblom, , that is informed by findings such as those presented here would be one way of doing this. The fact that younger offenders did not appear to exhibit a preference for targeting areas that were closer to the city center, but appeared to prefer those located further away is a potentially puzzling finding. Therefore, the pattern observed could simply be the result of younger offenders being more likely to commit crime closer to their homes confirmed by our analyses and to also live in areas that are further away from the city center.
Follow-up analyses provide some support for this by showing that on average younger offenders lived further 4. However, this finding deserves further exploration in future research. In line with social disorganization theory, it appears that offenders were more likely to target those areas in which residents have the least potential to form social ties, either because population turnover is relatively high, or because the residents come from different socioeconomic backgrounds or both.
This result is in line with those of Bernasco and Nieuwbeerta , who find that for residential burglary the social composition of an area may influence offender decision making by deterring offenders from targeting those locations in which social cohesion is most likely.