Data Manipulation in Excel

The Summary Files downloaded from the Census Bureau contained information that needed manipulation in several ways in order to meet the meeds of this study. The relevant themes included population (as well as male and female populations), age broken down into 31 age categories, education broken down into seven categories, employment status among males broken into four categories, family status broken into seven categories, and lastly per capita income, all subject to tract ID. The population of males and females was used to derive a ratio of males to females as an increasing trend exists between the male to female ratio and the occurence of sexual offences. The age categories below 16 years were added to provide a cumulative count of children under 16 as this age group represents the majority of victims of sexual offenders. It is usually quite difficult to predict the age of offenders as new research has shown that the average offender practices 16 years prior to arrest. The education categories below post-GED were added as there is an inverse relationship between education and sexual offences. Research has also shown that repeat offenders have little known career and familial attachments and are 97 % male (this estimate can vary depending on the source). Of the four categories for male employment status, I added the categories for "unemployed" and "not in the labor force", as sexual offenders might not always declare themselves "unemployed" with governmental agencies. Furthermore, the family types "married with children", "married without children", and "male with children" were compared with the "non-family" category to produce a family to non-family ratio. Population and per capita income were left as they were.

Conversion from vector to raster

The excel file was then exported as a databse file (*.dbf). The Census tracts shapefile was joined with the above mentioned excel file in ArcView and using the Spatial Analyst extension, each theme was exported as a a grid. Next the theme attribute grids were exported as ASCII files. IDRISI imported the ASCII theme files using ARCRASTER and converted them to raster files (*.rst) for raster analysis. Then the residences of sex offenders were digitized and converted to raster with POINTRAS.

Weighted Linear Combination was used to compare factors and thay were overlayed with the constraint file created from digitizing.