Assessments of access to healthful food frequently use GIS to measure the distance and concentration of food outlets relative to where residents live. These descriptive approaches do not account for food shopping behavior, which may vary based on the attributes of food shoppers and their activity space—places where they live, work, access resources, and socialize. Building on transportation research about accessibility, we reframe the issue of food access and equity from one about ‘what is nearby?’ to ‘where do people shop?’. We use a conditional logit model to analyze disaggregate data from a door-to-door survey of food shopping choice and food store surveys conducted in a predominantly non-Hispanic Black and middle-income and low-income section of Philadelphia. Our results highlight the importance of distance from home to food stores, overall, but they also emphasize the influence on food store choice of the race and sex of food shoppers, travel mode, and where they spend time other than at home, as well as food prices and the availability of healthful foods. This approach to understanding food access holds promise for future research that can link store choice to specific food purchases and health outcomes as well as for refining place-based strategies for improving access to healthful foods.

Where households across income levels shop for food is of central concern within a growing body of research focused on where people live relative to where they shop, what they purchase and eat, and how those choices influence the risk of obesity and chronic disease. We analyzed data from the National Household Food Acquisition and Purchase Survey (FoodAPS) using a conditional logit model to determine where participants shop for food to be prepared and eaten at home and how individual and household characteristics of food shoppers interact with store characteristics and distance from home in determining store choice. Store size, whether or not it was a full-service supermarket, and the driving distance from home to the store constituted the three significant main effects on store choice. Overall, participants were more likely to choose larger stores, conventional supermarkets rather than super-centers and other types of stores, and stores closer to home. Interaction effects show that participants receiving Supplemental Nutrition Assistance Program (SNAP) were even more likely to choose larger stores. Hispanic participants were more likely than non-Hispanics to choose full-service supermarkets while White participants were more likely to travel further than non-Whites. This study demonstrates the value of explicitly spatial discrete choice models and provides evidence of national trends consistent with previous smaller, local studies.

An empirical regularity designated as the Number-Average Size (NAS) Rule was first identified for the case of Japan by Mori, Nishikimi, and Smith , and subsequently extended to the United States by Hsu . This rule asserts a negative log-linear relation between the number and average population size of cities where a given industry is present. In this paper, we utilize the cluster-identification methodology developed by Mori and Smith to sharpen this notion of “industrial presence” by focusing only on cities that constitute at least part of a significant spatial agglomeration for the given industry. Our key result is to show that the NAS rule continues to hold (even more strongly) under this sharper definition.