Consumer Research with Big Data: Applications from the Food Demand Survey (FooDS)
基于三年多每月进行的食品需求调查数据，研究了消费者偏好的异质性，包括不同群体对在家和在外就餐的需求弹性差异、时间对选择实验的影响，以及用传统logit和机器学习方法刻画素食者特征。
In three separate studies based on data from the Food Demand Survey (FooDS), which has been conducted monthly for over three years, this paper explores heterogeneity in preference across consumers in traditional demand systems, heterogeneity in preferences over time in choice experiments, and the tail of the distribution for a particular food consumption pattern—vegetarianism. Results show that elasticities of demand for food at home and food away from home vary widely across different groups of consumers defined by a priori cluster analysis based on demographic and attitudinal variables. Results from a choice experiment are found to depend on when the experiment was conducted and on the market prices prevailing at the time of the survey. Given the large sample of consumers observed over time, there is sufficient data to demographically characterize a small portion of the population—vegetarians—using traditional logit models and a machine learning method ‐ a classifications tree.