Please use this identifier to cite or link to this item:
|Title:||Spatio-thematic accuracy in the evaluation of the English Safer Cities Programme|
|Authors:||Law, Ho Chung.|
|Presented at:||University of Leicester|
|Abstract:||The Safer Cities Programme in England as a whole implemented over 3,600 crime prevention schemes in 20 cities between 1988-1995 (total costing Â£30 million). The large-scale evaluation of the Programme's impact on domestic burglary has estimated that, overall, schemes of the Safer Cities Action reduced burglaries by 56,000 and were cost-effective (a saving of about Â£31 million). Using two cities: Bristol and Coventry within the Safer Cities Programme as a case study, this research aims to explore some of the accuracy issues in the GIS processing involved in the evaluation. This thesis a) describes how spatio-thematic accuracy can be estimated using Monte Carlo and dasymetric methods within the context of the Safer Cities Programme Evaluation, b) thereby provides a precise quantitative statement on the errors involved in the geographical data processing; and c) examines how spatial errors may affect the conclusion of the Evaluation using multi-level modelling. On average, the results show that the overlay method used in the Evaluation has over-estimated the household counts by 3.6% and 5% for Bristol and Coventry respectively. Subsequently, the Safer Cities Programme Evaluation has underestimated the action intensity by -0.8 and -9% and the burglary risk by -7% and -5% (for Bristol and Coventry respectively). Multi-level modelling shows that the mean errors due to the spatial interpolation estimated by the Monte Carlo dasymetric method are -.5%, 2.3% and 0.7% for Coventry, Bristol and the two cities combined respectively. In all cases, these are well within the standard errors generated by the overlay method. It is concluded that spatial and thematic errors have no significant impact upon the conclusions of the Safer Cities Programme Evaluation. However, spatial analyses show that potential burglary hot spots might have been missed as a result of such errors in crime pattern analysis. The analysis of the error distribution shows that a geographical area would have a higher error rate if it has: dense population; is near the city centre; or has an irregular geographical boundary. The implications in GIS applications, and crime prevention for decision and policy makers are discussed..|
|Rights:||Copyright © the author. All rights reserved.|
|Appears in Collections:||Theses, Dept. of Geography|
Items in LRA are protected by copyright, with all rights reserved, unless otherwise indicated.