When can it be said, 'You are what you know'? A multilevel analysis of expertise, identity, and knowledge sharing in teams

Individually held knowledge is one of an organization's most valuable assets. The extent to which an organization can leverage that asset depends on its members' not only applying knowledge in their work, but also exchanging and transferring knowledge with others in the organization. We still know very little, however, about why some knowledge workers are more or less willing to share their specialized knowledge with others. I argue that a robust explanation can be found in the risks or opportunities that knowledge sharing poses to personal identity. Specifically, knowledge workers' willingness to share knowledge with others can be explained by the importance they place on that component of personal identity associated with expertise (i.e., their expertise identity). I systematically explore contingency factors that might influence the effect of knowledge workers' expertise identity on their willingness to share knowledge, including other aspects of the self, dyadic social relationships, team identification, and the organizational environment. Finally, I argue that the effects of people's knowledge sharing will be evidenced in the learning outcomes realized by those around them.I conducted a cross-sectional survey study at a national engineering firm. The final sample included 221 members of 40 continuing teams (55% response rate). In addition to self-report data, surveys captured respondents' round-robin peer ratings of fellow team members on multiple constructs, including a measure of individuals' willingness to share their specialized expertise with others in terms of sharing the full range of personal techniques, reasoning, and experience that form the basis of their own mastery. I conducted analyses using multilevel modeling and social relations modeling techniques. Results supported 4 of 6 hypotheses. An individual's willingness to share knowledge with others was higher when expertise identity was high and dyadic trust, receiver expertise, and team identification were also high. Further, people with high expertise identity were less willing to share knowledge than people with low expertise identity when dyadic trust, receiver expertise, and team identification were low. Implications of these results, limitations of the study, and directions for future research are discussed.