The commonly used reputation systems in online markets facilitate establishing trust in a potential transaction partner who is often anonymous and geographically distant. Successful transactions will only take place when sellers and buyers trust each other, and a good reputation from previous experience will induce such mutual trust and lower the risks for both parties. In this dissertation, we used meta-analytic methods to synthesize evidence from over a hundred empirical studies to affirm the general existence of reputation effects and explored various potential moderators to explain the variation in reputation effects. Moreover, we argued that the effectiveness of a reputation system may not necessarily be that reputable sellers always earn a large price premium, but rather due to the system’s ability to attract and maintain a majority of trustworthy and reliable sellers, reputation will be less important.