57 / 372
臺大管理論叢
第
26
卷第
2
期
57
conduct may generate certain risk for them. The findings offer four significant empirical
contributions. First, cognitive social capital is the most influential factor in determining
Facebook users’ perceived psychological safety and knowledge sharing self-efficacy, as well
as their knowledge sharing behavior. Engaging in a meaningful exchange of knowledge
requires at least some level of shared language and vocabulary (Nahapiet and Ghoshal,
1998). Although Facebook users use photos, audio, and video content to communicate with
one another, written language still seems to be the major communication vehicle. Our
findings indicate that in social networks, cognitive social capital represents the shared
language and narratives (Nahapiet and Ghoshal, 1998), which ultimately stimulate users’
knowledge sharing behavior by strengthening their sense of psychological safety and
knowledge sharing self-efficacy.
Second, relational social capital significantly influenced psychological safety, but its
impact on knowledge sharing self-efficacy and knowledge sharing behavior was negligible.
One reason behind this observation is that relational social capital (i.e., interpersonal levels
of trust, norms, obligations, and identification) may not develop easily in SNSs (e.g.,
Facebook) because of the possible lack of highly interdependent and frequent interactions
and co-presence (Nahapiet and Ghoshal, 1998). In addition, because most interactions on
SNSs are acquaintance based, users occasionally meet one another in offline settings; thus,
online knowledge sharing self-efficacy and behavior may become relatively less critical.
Moreover, our results show that psychological safety does not exhibit a significant influence
on knowledge sharing behavior. Many users may believe that Facebook is not a safe place to
share knowledge because of unpredictable risks, such as when users’ self-disclosures cause a
misalignment of their actual self-images with ideal ones in terms of online impression
management.
Third, our findings indicate that structural social capital has no direct effect on either
knowledge sharing behavior or psychological safety. This result differs from that in research
in the fields of organizational behavior (Bourdieu, 1986) and virtual communities (Wasko
and Faraj, 2005). A plausible explanation is that structural social capital as manifested in
organizational behavior/virtual community studies develops from a “closed environment”
context. That is, the participating members are well-defined and identifiable, and the purpose
or objective for knowledge sharing (which may be work related, interest driven, need based,
or goal oriented) is also specific. Those who occupy pivotal or central positions in social
networks may possess more influential power or share more knowledge. However, Facebook
is not a closed environment, but an open and loose system with no finite boundaries and