Skip to main content
. Author manuscript; available in PMC: 2011 Jul 1.
Published in final edited form as: Prev Med. 2010 Apr 18;51(1):63–67. doi: 10.1016/j.ypmed.2010.04.009

Table 4.

Regression results on the differences in aggregate shelf space of foods between African-American and mixed racial census tracts, New Orleans, LA, 2004-2005

Food Shelf Space a β b SE p-value Mean difference c
Fruits
 Fresh -1.34 0.61 0.03 -19.7
 Canned -0.09 0.35 0.79 -1.1
 Frozen -0.38 0.14 0.01 -1.0
 Total -1.17 0.65 0.08 -22.4
Vegetables
 Fresh -1.47 0.67 0.03 -27.3
 Canned -0.10 0.54 0.85 -1.8
 Frozen -1.23 0.38 0.00 -12.4
 Total -1.54 0.89 0.09 -42.3
Energy-dense Snacks
 Salty Snacks -0.29 1.10 0.79 -12.1
 Cookies/Crackers/Pastries -0.84 1.13 0.46 -33.5
 Candy -0.25 1.00 0.80 -9.4
 Carbonated Beverages 0.04 1.31 0.98 1.6
a

Each row represents a separate OLS square root transformed model in which the dependent variable is the amount of shelf space of a given food. Reference category is mixed racial tracts, i.e. less than 80% African-American residents. All models are adjusted for tract population density.

b

Beta-coefficients from the regression models are not readily interpretable since the dependent variables were square root transformed. A negative sign on the coefficient indicates a lower amount of aggregate shelf space for African-American tracts.

c

Mean difference in aggregate shelf space (meters) of specific foods between African-American and mixed racial tracts. Differences are calculated after back-transforming raw regression results to the original units, i.e. meters of shelf space. A mean tract population density of 3,820 pop/km2 was used in these calculations.