Separating what we eat from where: Measuring the effect of food away from home on diet quality

Many argue that food away from home (FAFH) is a contributing factor to the obesity epidemic, showing that body mass index and consumption of FAFH are positively correlated. However, correlation analyses using a simple regression approach, such as the Ordinary Least Squares (OLS), do not prove that FAFH causes weight gain. We use a first-difference estimator to establish a causal relationship between FAFH and dietary intakes. Using dietary recall data from the 2003-2004 National Health and Nutrition Examination Survey and the 1994-1996 Continuing Survey of Food Intakes by Individuals, we find that FAFH does indeed increase caloric intake and reduce diet quality, but that the effect is smaller than if estimated using OLS. Thus, models based on associations are likely biased upward, as much as 25% by our estimates.