Christopher Greene

This study used 2006 Statistics Canada census data to investigate whether aggregate socioeconomic characteristics, represented geographically by census tract, had explanatory value concerning participation in an urban reforestation program. A two-stage multinomial regression analysis of census data for Etobicoke, Markham, Scarborough and Toronto was designed. Stage one evaluated the strength of a priori assumptions about program participants (developed through a literature review and professional consultation). Stage two employed an exploratory data analysis approach to identify variables that may have differed from a priori assumptions. Results revealed that there was a marked regional difference in both the baseline regression model (based on a priori assumptions), as well as in the variables identified through the exploratory regression analysis. The explanatory ability of the regression models was strongest for Toronto and weakest for Markham. Strong contributing variables included the number of dwellings constructed before 1946 and the number of single-detached dwellings per 10,000 population.