Geographies of Care: Space, Place and the Voluntary Sector

tion. Consider briefly three constructs that many health geographers might consider relevant to the understanding of relationships between health indicators (e.g., disease incidence) and area-based measures of the neighborhood environment (e.g., poverty). Tobler’s (1970) first law of geographyFeverything is related to everything else, but near things are more related than distant thingsFimplies that spatial dependencies are almost certainly present in the area-based measures used to evaluate neighborhood-health relationships. Yet the notion of spatial correlation and how it might affect analytical frameworks receives little mention. For example, correlation coefficients for area-based data are presented without any correction for spatial autocorrelation, even though it is widely recognized that both correlation coefficients and probability values are biased when the variables are autocorrelated. Similarly, the Modifiable Areal Unit Problem (MAUP) is a concept central to the interpretation and understanding of how the geographic scale of analysis influences bivariate correlations in area-based measures. But although correlations between area-based measures at different geographic aggregation scales (e.g., census tracts, zip codes) are presented in the book, the implications of MAUP for their interpretation is not discussed. Finally, the ‘‘small numbers problem’’ is often encountered when populations in geographic subunits such as neighborhoods are subdivided by race and ethnicity. Disease rates as well as many area-based measures of deprivation and SES include population size in the denominator and are unstable when their variance increases as the denominator decreases. A discussion on approaches to rate stabilization that might be used when working with neighborhood data would be a useful addition to this otherwise highly comprehensive collection. Although certain constructs of medical geography might have received better treatment, the sociological paradigm as presented in this book, especially the role of the neighborhood as a mediator for psychosocial effects in individual and population health, is currently not well represented in medical geography and spatial epidemiology. The need for an integration and a synthesis of the sociological and geographic approaches in studying the role of neighborhoods in human health is therefore substantial. This book makes a major contribution to our understanding of the health effects of neighborhoods. Its depth and clarity of presentation is exceptional, and it is a ‘‘must read’’ for those who research relationships between health and place.