Abstract:

The rating system is at the core of many digital platforms. However, the quality of the personal interaction between buyers and sellers may bias to a large extent the ratings left by reviewers, and consequently the performance of the provided service. We perform text analysis on reviews left on the Airbnb platform in Barcelona in 2019, using a text analytics algorithm to infer hosts' attitude – i.e., “kindness” - towards their guests. We find that a kinder attitude by the host is associated to higher ratings. this is true also for some specific measures of rating, such as the location score rating, that should be independent from the quality of the interpersonal relationship. This suggests that the host’s behavior can induce a psychological bias on the rating left by the guest, who may feel indebted to reciprocate the host’s kindness with a generous score. Our third result is that, after controlling for the rating, a kinder attitude by the host is associated with significantly higher occupancy rates. This finding suggests that text reviews contain (possibly hide) information that the rating scores do not fully capture, and that this information has an impact of its own.

To join the meeting click here.