The purpose of this study is to explore how urban residents respond to their social and physical environments--what they define as problems and how they respond to them. I focus on one large, city--Baltimore, Maryland--and then compare two very different neighborhoods within it: Federal Hill, a well-off, and fashionable, area with mostly white residents, contrasted with Sandtown-Winchester, a neighborhood plagued by urban blight and crime, and where the majority of residents are black. I use a geographic information system (GIS) and spatial analyses to explore neighborhood call rates regarding physical and social incivilities, using the traditional sociological framework of "social disorder" as a theoretical lens for exploring similarities and differences in what disorders increase or decrease call rates.

I use more commonly applied stochastic methods for much of the analysis (statistical means and ordinary least squares statistics), but I also explore, in a tentative way, the potential power of spatial methods, which are not widely used or known in sociology, to reveal more about what makes these spaces similar and different and how they affect call rate patterns.

The predictive models demonstrate mixed results when predicting variation in the call rate patterns of the two neighborhoods. Income, education, and population-density effects are consistent, yet weak, positive predictors in both areas, while other indicators (home ownership, number of vacant houses, etc.) exhibit substantive positive effects in the wealthier neighborhood but none in the poorer. Neighborhood homogeneity and stability show negative impacts on rates, but depending on the neighborhood.

I focus on how local variations in action, even under similar circumstances, may depend not only on residents' aggregate capacity to commit to change, but also on how neighborhood space is internalized as a "neighborhood generalized other" as a "community," according to George Herbert Mead, either constraining or enhancing engagement. This within- and between-neighborhood variance in the strength and direction of predictor variables, and in their capacity to predict residents' calling patterns, underscores issues of validity and operationalization regarding indicators traditionally used to measure social disorganization, and how spatial methods can be valuable corrective tools.