Understanding relational antecedents to ratings inflation in online labor markets

Existing literature and theory suggest that efficient transactions in online markets are undergirded by reliable reputational signals. This assumption is especially true for online markets such as Amazon, Uber, and Upwork, which function at a global scale and facilitate trust between millions of people with little more than public ratings as a representation for explicitly codified reputation. These ratings are assumed to provide transparency and accountability into past behavior. However, recent studies are beginning to show rampant ratings inflation in online markets, threatening the foundation of these markets. Current explanations have almost exclusively looked at ratings inflation after ratings have already been given. In this paper, however, I examine what occurs before ratings are entered to better understand the relational antecedents that lead to ratings inflation. Through an inductive analysis of client-contractor interactions in one of the largest online labor markets, my findings show that ratings can be inflated when the relationship between clients and contractors is going well and also when the relationship and project is struggling.