Home

Methods

To investigate if some factors such as population size and socio-economic status increase the likelihood of a crime, specifically residential burglary, we examined general trends between the census tracts. Each of the parameters were first mapped out on the census tracts, using census data provided by American FactFinder and 2004 crime statistics provided by the San Diego Police Department.

The study area was determined by selecting from California census tracts for which the FBI's Uniform Crime Reporting (UCR) statistics were available from the San Diego Police Department for the year 2004. This information was collected and published under the Clery Act, which requires post secondary institutions to present crime data in and around their campuses. As a result, our data and analyses did not cover the entire San Diego city region. Other data collections exist from the San Diego Police department that summarise crime statistics, however, these are based on less well-defined areal units such as neighbourhoods. Census tracts # 3800, 5500, and 9902 showed anomalies in the data, perhaps leading to errors in analyses, the reasons for which will be discussed later. It should also be noted that there were 77 residential burglaries for which the location was reported as "unknown" by the San Diego Police Department. This could have a severe impact on the representation of the data, as the maximum number of residential burglaries in known census tracts was 86.

Click on each image to enlarge

Number of Burglaries
	This map shows a summary of the number of residential burglaries per census tract. These totals were then correlated to the values obtained for the other parameters using the regression tool (regress.dll by M. Sawada); the outcomes can be seen in the results section.

Total Population
	It may be expected that the number of crimes occurring would be proportional to the size of the population, so the total population per census tract was assessed and displayed. We expected that those census tracts with high populations would correspond with a greater number of burglaries. The population in each census tract ranged from a minimum of 63 people to a maximum of 10,000.

Income

We also expected that neighbourhoods which typically represent higher income brackets my be targeted, as those with more disposable incomes may furnish their homes with valuables such as high end electronics. The median household income per census tract was provided by American FactFinder. The census tracts appear to be segregated into blocks based on the income values, with the highest found in the northern area, and the lowest being in the south central area.

Housing Units

The housing units were examined on the assumption that there would be a higher prevalence of burglaries in multiple unit homes, such as apartments and condominiums, than for single unit homes. Apartments are often inhabited by lower income residents, and may not always offer the security that single unit homes do. Higher end condominiums, on the other hand, may be less likely to be burglarised, due to security features such as video cameras, security personnel, and special locking mechanisms in the lobby. The number of single unit homes was normalised by the total number of homes to produce a percentage for that census tract. The multiple unit homes were assessed in the same way. Mobile housing units such as RVs and boats were included in the housing unit totals, but were not individually analyzed for the sake of simplicity.

Home Ownership : Owned vs Rented


Finally, the occupant status was mapped, where we predicted that similarly to the apartment building trend, those who rent their homes are more likely to be burglarised than those who own them. Again, renters may have a lower income, and may be unable to install security measures if they do not own their home. The number of both types of occupants were normalised by the total number of occupants.