The President’s doctoral Scholar Award of the University of Manchester has financed a research project aimed at gaining an insight into the application of small area estimation techniques to know the distribution of variables pertaining to crime surveys within a small geographical area. The project will allow David Buil Gil to develop his doctoral thesis under the direction of Juan José Medina Ariza and Natalie Shlomo.
Those involved in the project are convinced that when public policies are designed to focus on crime prevention and the increase in security to be perceived by the public, criminality detected by law-enforcement agencies cannot be assessed on its own, as their data only amounts to a part of the crime committed. Furthermore, other variables of interest must be considered which are only detectable with crime surveys, like fear of crime, perceptions of disorder or the level of trust on public administration.
The problem arises as small samples used for these surveys do not allow for an observation of the existing differences between small areas of cities, impeding the design of public policies focused and based on the existing differences between areas. In other words, the samples used for the crime surveys include an interview, on average, with at least one person from a neighbourhood and, therefore, with the use of conventional methods the differences between small areas cannot be known.
According to the researchers, the use of a new generation of statistical techniques, known as small area estimation, allows for the calculation of the estimation of the variables of interest regarding small geographical areas. Starting with a combination of data gathered for the surveys with detailed information from the census and other public data, statistical models are generated which contribute to explain, with the use of estimations, the differences between small areas. As mentioned previously, although statistical models exist to make estimations in small areas both on an individual and an area basis, in the case of crime surveys most studies will have to be based on models on an area basis, as data is rarely published on an individual basis due to factors linked to anonymity and confidentiality.
Specifically, the project carried out at the University of Manchester will facilitate estimations in the city of Manchester, in terms of a small area, with some of the variables gathered by the Crime Survey for England and Wales. As members state with confidence, the resulting estimations of the project will allow for the design of both public policies and crime prevention strategies which discern the differences in the distribution of a range of variables concerning very small areas, allowing for a scientifically based approach to social problems detected in different areas of cities.
Links of interest
Article about Small Area Estimation for Targeting Hot-Spot Policing Interventions published by the British Society of Criminology Policing Network.