Table 2. Odds of Purchasing Foods on Sale Versus Not on Sale Among Shoppers Who Purchased These Foods, Philadelphia, Pennsylvania, April through August 2010 and December 2010 through October 2012.
| Food Category | ORa (95% CIb) | P value |
|---|---|---|
| Low-calorie foods | 1.3 (1.0–1.7) | .08 |
| Fruit | 1.1 (0.7–1.7) | .61 |
| Vegetables | 1.3 (0.9–1.8) | .15 |
| Low-fat dairy | 4.7 (0.9–24.9) | .07 |
| High-calorie foods | 2.4 (2.0–3.0) | <.001 |
| Sweet snacks | 5.9 (3.5–10.0) | <.001 |
| Savory snacks | 1.1 (0.6–2.0) | .77 |
| Sugar-sweetened beverages | 2.6 (1.9–3.7) | <.001 |
| Grain-based snacks | 6.6 (3.6–12.0) | <.001 |
Abbreviations: OR, odds ratio; CI, confidence interval.
a
Fixed effects logistic regression models predict that food was purchased (“1”) compared with not purchased (“0”) in weeks that food was on sale (“1”) compared with weeks food was sold at full price (“0”). Estimates are based on 79,087 observations from 81 households that had purchase data on more than 1 day. Models adjusted for household exposure time in the study.
b
95% CIs constructed from robust standard errors.