A virtual community ("VC") is a group of people who interact primarily via IT, rather than face-to-face, for informational, social, professional, educational or other purposes. Given the growing popularity and special features of VCs (user-generated contents, virtual social interaction and voluntary participation), this dissertation intends to understand two user behaviors: users' continuance intention to participate and consumer purchase decision. The dissertation consists of three studies. Study 1 focuses on how the quality of user-generated contents and system performance impact users' continuance intention to participate in information-exchange VCs. In particular, this study investigates VC quality from a multi-dimensional perspective. As an extension of Study 1, Study 2 examines the role of users' past behavior on future participation intention. The objective of Study 3 is to understand the role of user-generated contents, namely, user reviews, and system performance in consumer decision-making outcomes in the context of transaction-based VCs. This study differentiates types of social influence developed from user reviews and investigates their sources and impacts on consumer purchase decision and system evaluation. The model also incorporates two individual characteristics to understand consumer differences in the formation of social influence. This dissertation intends to contribute to the IS literature on user behaviors in VCs and the value of VCs. It provides meaningful insights for VC design and management.