Are embedded online communities of practice more successful than exclusively virtual ones?

Is a deliberate shaping of the mixture of virtual and 'real-life' interaction in online communities a crucial condition for diminishing typical problems of knowledge sharing? Typical problems that increase the risk of failure are: 1. a lack of trust between members, 2. free rider behavior, and 3. a lack of stable membership. Social network researchers claim that relations offline would affect relations online. However, there is only limited knowledge about what characteristics of the online interaction are affected by the existence of offline relations.The paper attempts to answer two questions: a) how often do the different problems emerge in online communities? and b) does a mixture of virtual and 'real-life' interaction of at least some members -in contrast to 'purely virtual' interaction- reduce the prevalence of the problems that the whole community faces? Since the mixture of virtual and 'real-life' interaction of at least some members, that is the degree of embeddedness of an online community into offline networks, can be influenced, knowledge about the impact of the degree of embeddedness (question b) is an important input to the design and management of online communities. Knowledge about the prevalence of the different problems (question a) is crucial for focusing the knowledge management efforts in different types of online communities on their most prevalent problems. The analysis of the emergence of problems uses data of surveys conducted in 33 online communities of a virtual organization that hosts communities for teachers in higher education (response rate: 37%). The descriptive results show that the problem of membership fluctuation plays a role only in a small number of online communities. Problems of trust and especially problems of free riding are much more significant problems. The results of the tests of the three hypotheses, that is the hypotheses concerning the impact of embeddedness on the three problems of interaction, provide evidence for the first and second hypothesis, but not for the third hypothesis about membership fluctuation. The embedded online communities tend to have less problems of trust and less problems of free riding than the virtual ones. It is concluded that offline networks have clear beneficial effects on online knowledge sharing. The findings contribute to our knowledge about how offline networks influence online interaction.