Chapter 5: Appendix B: Predicting feedback
We collected three datasets of Yelp reviews from mid-2011 through mid-2012: (1) the feedback sample described in Appendix A, which for this study was reused to develop content analysis codes (n 5 640 review texts); (2) a set of filtered reviews (n 5 522); and (3) a new test set of reviews (n 5 865). For the filtered dataset we sampled up to 80 filtered reviews from each of the eight restaurants in the developmental sample. For the test dataset, we added four new restaurants, yielding a sample of six expensive and six inexpensive restaurants that had the highest review counts in their respective categories. We then gathered reviews from the most recent eight months, capping the number of reviews per restaurant at 100. With our two MBA research assistants, we read widely in the sample of reviews, to identify text properties that seemed to be associated with better reviews. For example, some reviews described specific menu items, and those reviews seemed to be of better quality. The coded properties that survived a series of pretests are given in Table B.1. For validation, the table shows that the codes that, in our opinion, marked out better reviews, were also significantly more common in reviews written by Yelp Elite. These properties also distinguish filtered from regular from Elite reviews; see Table B.2. Better-quality reviews were less common among the filtered reviews, and more common in reviews from Elite, honored reviewers.
You are not authenticated to view the full text of this chapter or article.
Elgaronline requires a subscription or purchase to access the full text of books or journals. Please login through your library system or with your personal username and password on the homepage.
Non-subscribers can freely search the site, view abstracts/ extracts and download selected front matter and introductory chapters for personal use.
Your library may not have purchased all subject areas. If you are authenticated and think you should have access to this title, please contact your librarian.