A bad restaurant experience could leave diners suffering from a mild form of trauma, say researchers at Stanford University. According to the Telegraph, Dan Jurafsky — a professor of linguistics and computer science and author of The Language of Food — presented his findings over the weekend at the American Association for the Advancement of Science's annual conference. SBS writes that Jurafsky and his team combed through 887,658 Yelp reviews from 6,548 restaurants in six American cities — New York, Los Angeles, San Francisco, Chicago, Philadelphia, and Boston — posted between 2011 and 2013.
"This idea of getting through suffering collectively is there in these one-star reviews. These are minor traumas."
The researchers found that language used by Yelpers when writing about "bad" experiences is "similar to that used by people recalling terrorist attacks or horrific accidents." Jurafsky notes that the language of these one-star reviews does not focus on the food and they are "in the past tense" not the present, which trauma victims apparently use to "distance themselves from the situation." There are also a lot of "pronouns and mentions of other people," many negative words like "terrible and awful," and a heavy use of group words like "we" and "our."
Jurafsky notes that people writing "after 9/11" and students writing in the school paper after a "campus tragedy" also used "this exact language." They especially deployed the use of "first person plural" to say "this bad thing happened to us as a group and we're going to get through it together." He adds, "This idea of getting through suffering collectively is there in these one-star reviews. These are minor traumas."
Happy Yelpers, on the other hand, have been found to write about sex and drugs.
This is isn't the first time researchers have turned to Yelp for data: New York City health officials sifted through the restaurant review site to find unreported cases of foodborne illnesses. They searched for specific terms like "sick" and "vomit" in nearly 300,000 reviews and they managed to identify three unreported outbreaks with the system.