The large majority of reputation systems use features such as star ratings and reviews to give users a reputation in online peer-to-peer markets. Both have been shown to be effective for signaling trustworthiness. However, the exact extent to which these features can change perceptions of users' trustworthiness remains an open question. Using data from an online experiment conducted on Airbnb users, we investigate which of the two types of reputation information-average star rating or the number of reviews-is more important for signaling a user's trustworthiness. We find that the relative effectiveness of ratings and reviews differ depending on whether reputation has a strong or a weak differentiation power. Our findings show that reputation effects are contingent on and susceptible to the context created by the alternative choices presented to users, highlighting how reputation information is displayed can drastically alter their efficacy for engendering trust.