Socio-Linguistic Differences Between Reviewing People and Reviewing Goods: A Relational Message Production Logic of Crowdsourced Reviews on Airbnb

This paper distinguishes between two ideal types of crowdsourced review activities––reviewing goods and reviewing people. Although many online reviews mix the two activities, we argue the theoretical distinction sheds light on the reviews’ message production logic regarding the relational, prosocial orientation of social interaction. We test this logic with a sample of mutual reviews (N= 191,526) between Airbnb hosts and guests. We contrast the reviews’ socio-linguistic features, including prosocial language and positive emotion, against different relational opportunities induced by lodging options (with different extents of face-to-face interaction) and a policy change (which suppressed review reciprocation). We found when guests review hosts, which more closely resembles reviewing goods, the message features are biased by relational opportunities. However, when hosts review guests, which more closely resembles reviewing people, these opportunities have different if not opposite impacts. The findings suggest the importance of relational dynamics underlying the message production of reviews.