Examining the Impact of Environmental Factors on Quality of Life Across Massachusetts*

Abstract

Several studies indicate that there are significant relationships among quality of life, green vegetation, and socioeconomic conditions, particularly in urban environments. The purpose of this research is twofold: (1) to compare two weighting and aggregation techniques, data envelopment analysis (DEA) and principal components analysis (PCA), in the development of a socioeconomic index; and (2) to test for and explore spatial variation in the relationship between socioeconomic index and environmental variables using geographically weighted regression (GWR). The analysis was conducted at the census block group level in Massachusetts. First, DEA and PCA were used to generate two separate socioeconomic indexes. Second, the relationship between these indexes and environmental variables including percentage impervious surface, percentage industrial land use, percentage land used for waste, and traffic density was modeled using ordinary least squares (OLS) regression and GWR. The GWR models explained more variance in the relationship than the OLS models and indicated that there is considerable spatial variation in the character and the strength of this relationship. The results of the GWR analyses were similar between the models generated using DEA- and PCA-derived indexes, indicating that the results were corroborative. The study concludes that the environmental variables are generally a strong predictor of the socioeconomic conditions at the scale of census block group; however, there is substantial geographical variation in the strength and the character of this relationship. The results of this study also suggest that various weighting and aggregation methods should be tested in every study that uses or creates composite indicators.