Doctor of Philosophy (PhD)


Agricultural Economics

Document Type



The major focus of this study was to identify the linkage between information and consumer behavior and to examine the role that a new media index played in the identifying the tone of news stories. Past studies have relied on using the number of media stories but, this study tests a new approach, i.e., in order to analyze the effect of information by incorporating the sentiment indices. These indices were created based on the tone of the news story. The results from sentiment indices were compared to the results from the information index. Barten’s Synthetic model (BSM) was used to capture the demand interrelationships and a Polynomial Inverse Lag model (PIL) was utilized in order to identify the size and length of change in demand. A time series data set, comprised of household purchases from 2008 to 2010, was created with information obtained from Nielsen HomeScan panel data. This study investigates three cases of food safety incidents, in which each case presents a unique scenario of food safety incidents. The first case is related to the PCA peanut butter recall of 2009, the second case investigates the refrigerated cookie dough recall of 2009 and the third case considers the effect of the Gulf oil spill on the demand for meat products and, in particular, on the demand for frozen seafood. This study was able to confirm significant changes in the demand of affected products in the post-event period. Furthermore, the study brings important contributions to the existing literature. To our knowledge, this is the first study which utilizes a natural word processing algorithm to identify the sentiment of the news story related to a particular food safety incident and that actually assigns a sentiment score to the story. The comparison between the media index, created by using the number of news stories related to the particular event, and the sentiment index reveal that the sentiment index exhibits a stronger effect on the demand for the products tested. The results from sentiment indices show that it can be used as a feasible alternative to the currently used media indices to measure the information effect on demand.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Harrison, Robert W