摘要：Understanding the role of health information in food and nutrient demand has become an important issue over the last decade. Endogeneity and measurement error are two empirical problems that are inherent in this type of analysis. While some type of instrumental variables estimation would appear the obvious solution, this paper provides several theoretical and empirical reasons why this is not the case in cross-sectional analysis. An alternative estimation strategy is pursued, an empirical example given, and the implications discussed.
