Austin Et Al. Respond
Bibliographic Data
We thank Spielman for commenting on our article and giving us an opportunity to continue the dialogue on methodological issues in studying food environments. In response to his point about network versus straight-line distances, we would like to note that in a related study, our research team found network and straight-line distances in Chicago to be highly correlated, and results did not differ when models were tested with network versus straight-line distances. Chicago streets follow a fairly dense grid, which may explain why a difference was not found between the 2 methods. That said, Spielman has identified an aspect of spatial analyses that needs more methodological development. ArcGIS software1 will calculate network distances, but it is only as good as the available network data. For instance, street network data generally available for geocoding addresses do not indicate barriers to pedestrians, such as a divided roadway with limited pedestrian crossings. In response to Spielman’s comment about zoning regulations affecting land-use patterns, we would like to clarify that our analyses and the reported confidence intervals were not based on an assumption of complete spatial randomness of restaurants throughout Chicago. Rather, simulations used to generate confidence intervals were based on a null hypothesis of spatial randomness of restaurants relative to the spatial location of schools. We too were interested in examining the potential importance of zoning regulations, so we repeated our spatial analyses within strata defined by degree of commercialization. As reported in our article, we found evidence of clustering within areas of the city characterized by high and moderate levels of commercialization. Taking a step back from the details of the statistical methods, an observation we would add to the discussion is that rather than overestimating the availability of unhealthful food in school neighborhoods, our study is very likely to have underestimated it. Our fast-food database did not include convenience stores, corner stores, gas station food markets, and vending machines that offer children access to an abundance of inexpensive, calorie-dense food of low nutritional value. Future work that includes these additional food outlets in characterizing food environments in school neighborhoods may in fact paint a more dismal nutritional picture than our study did
Cartography · Confidence interval · Econometrics · Geocoding · Geography · Randomness · Spatial analysis · Spatial ecology · Statistics · Street network · Transport engineering · Zoning · Computer Science · Ecology · Engineering · Mathematics · Organic Food and Agriculture · Urban Agriculture and Sustainability
| Unique citing works | 1 |
|---|---|
| Citations per year | 0,07 |
| Citation span | 2011 - 2011 (1) |
| Citation velocity | historical |
| Highly cited | No |
| Citation types | Neutral: 1 |