Emotions in reviews and review biases in online reviews using text sentiment analyses

03/10/2022 14:00
Turkey

The seminar titled "Emotions in reviews and review biases in online reviews using text sentiment analyses" will take place on October 3, 2022, at 14.00 on Zoom. Please click here to attend the seminar.

This research investigates how social influence (i.e., prior average rating) impacts subsequent ratings and to what extent the reviewers' expressed emotions (i.e., joy and anger) and hotel characteristics (i.e., price category and affiliation) can, in turn, affect the relationship between social influence and subsequent ratings. Reviewers' emotions in hotel reviews from TripAdvisor and Expedia have been analyzed with text analytics. Our results demonstrate that the positive relationship between social influence and subsequent ratings becomes less positive when the review expresses joy and more positive when the review is anger-based. The negative effect of budget (vs mid-priced) hotels on subsequent ratings are more negative with increasing prior average ratings. However, the effects of luxury (vs mid-priced) and independent (vs chain) hotels on subsequent ratings do not change with prior average ratings, indicating no social influence effect. Prior average ratings have a stronger effect on ratings of budget hotel consumers and anger-embedded reviews. We recommend not exposing prior average ratings to visitors when they initiate writing a review and if they do, expose to prior average ratings next to hotel class-specific average ratings or de-bias them before exposing them.