Here in three parts is the article review/presentation I prepared for INFS 834, Knowledge Management. Following the videos is the text of the original review text.

Part 1:

Part 2:

Part 3:


McLure Wasko, M., and Faraj, S. (2005). Why should I share? Examining social capital and knowledge contribution in electronic networks of practice. MIS Quarterly (29:1), 35-57.

Problem/Relevance

The problem  Wasko and Faraj address is not so much a problem as a riddle: why would complete strangers contribute their knowledge in an electronic network of practice? With no apparent immediate benefit to the contributor and the awareness that free-riders will profit from knowledge that contributors may have worked hard to obtain, why would contributors engage in what looks like a no-gain scenario?

slide 2 Understanding what motivates contributors in a network of practice is important for various reasons. A single organization cannot find all the information it needs within its boundaries (p. 36); networks of practice are thus vital sources of external connections and knowledge. Bringing new knowledge into an organization is critical to maintaining competitive advantage, yet surrendering that knowledge in a network of practice has the potential to reduce competitive advantage. If contributors perceive this potential loss yet continue to contribute, then they either perceive some outweighing benefit to contributing or they don’t care about competitive advantage, which would certainly upset our understanding of market forces. Whatever the case, knowledge managers stand to benefit from understanding in the most detail possible the motivations behind what is apparently a successful and organic knowledge-sharing mechanism.
Prior Work

Wasko and Faraj build on Brown and Duguid’s (2000) description of networks of practice and their distinction from communities of practice. Here communities of practice are viewed as primarily as internal associations of colleagues within organizational walls with regular interaction, while networks of practice are viewed as more external associations with fewer and looser points of commonality. ¬†Wasko and Faraj still maintain some level of formal organization in their definition of electronic network of practice, requiring that the network still be “sponsored by a specific organization or professional association” (p. 37). However, their definition retains the key distinction that participation is open, voluntary, and thus to a significant extent self-organizing.

Wasko and Faraj point to several factors shown by prior research to promote knowledge sharing but which are clearly absent from electronic networks of practice: strong social ties, co-location, demographic similarity, status similarity, and a history of prior relationship. The alternative factor the authors explore, social capital, also faces challenges in the electronic network of practice arena, as the authors point to research showing social capital is enhanced by “shared history, high interdependence, frequent interaction, and closed structures” (p. 38), features largely absent from electronic networks of practice. Such networks should also be suboptimal media for knowledge sharing given the difficulty demonstrated in previous research of transmitting tacit, embedded knowledge through technology.
Contributions

The authors’ key contribution is the extension of Nahapiet and Ghoshal’s (1998) model of social capital from organizations to electronic networks of practice. They seek to understand the role of each of the four areas of that model in knowledge contributions:

  1. Individual motivation is represented by desires to enhance professional reputation and to help others (p. 40).
  2. Structural capital is represented by network centrality, the increased regularity of contact with other members that supports collective action (p. 41).
  3. Cognitive capital is represented by expertise and tenure in the field (p. 42).
  4. Relational capital is represented by commitment to the network and the extent to which a participant is “guided by a norm of reciprocity” or a sense of “mutual indebtedness” toward fellow network members (p. 43).

The authors seek to test whether each of these seven features contribute to more and more helpful contributions.

Methodology

The authors collected data from a 7000-member legal professional association that maintains a Web-based electronic network of practice (message boards). They examined an exhaustive four-month sample of message postings to obtain data on contributions. The authors then combined this data with demographic data provided by the sponsoring association and survey data provided by willing members. All survey measures come from previous studies. The authors found satisfactory representativeness among the survey respondents (173, or 29% of the total) in participation rates, gender, age, and employment, while noting that respondents general had a higher level of experience than the general membership (p. 46).

The authors measured the dependent variable of knowledge contribution through manual coding and rating of all messages in the sample. They tested an initial subset of 100 messages for intercoder reliability with a domain expert from the sponsoring association; satisfied with the reliability obtained, one author proceeded to handle coding the remaining bulk of the messages. From the resulting ratings, the authors calculated average helpfulness ratings for each network participant. The authors also counted the volume of contributions from each participant.

Conclusions

Numerically, the results are underwhelming. Among some of the notably failed hypotheses:

  • The correlation between how much users enjoy helping and the volume of their contributions, a meager 0.13, only “approached significance” (p. 49)-in clearer terms, was insignificant. Evidently, the pleasure users take in helping others does not motivate significantly greater contribution volume.
  • Even though expertise and tenure correlate significantly with volume, neither factor correlates significantly with helpfulness. If expertise and tenure reflect more and more valuable knowledge, perhaps the holders are less willing to release helpful parcels of that knowledge that might decrease their competitive advantage and thus do not provide much more help than younger, less experienced network members.
  • Expertise and tenure do correlate significantly and negatively with commitment and reciprocity, suggesting that perhaps the absence of noticeably greater helpfulness from more experienced experts stems from relatively lower shared commitment to the values of the network. However, commitment and reciprocity produce either insignificant or weakly negative correlations with helpfulness and volume. We therefore cannot conclude that a loyal participant who feels a shared obligation with network members would be any more helpful or prolific than a member who feels no obligation to other participants and would hardly notice if the network disappeared tomorrow.

Overall, even the significant correlations might be overwhelmed by a strong wind. Centrality shows the strongest correlations, 0.46 with volume and 0.33 with helpfulness. None of the other significant correlations break 0.30. These correlations may be better than nothing, but they do not offer much for designers of electronic networks of practice to hang their hat on with any expectation of strong return on investment.

One might offer practical advice to network sponsors to foster an active corps of participants who build centrality and form the oft-mentioned “critical mass” of users who drive knowledge sharing in the network. One might also emphasize the reputational benefits of participation. However, on such weak correlations, network sponsors must always be attentive to other confounding factors.

Strengths/Weaknesses

The authors seem in places to overstate the case for support of their model. One researcher’s “qualified support for most of our hypothesized relationships” (p. 50) is another’s failure to reject the null hypothesis. “Most” also exaggerates the case: if we break the hypotheses down by helpfulness and volume, Table 3 (p. 50) indicates significant relationships at p < 0.05 or better for exactly half of the fourteen tested relationships, and two of those (commitment-helpfulness and reciprocity-volume) are negative, the opposite of what the hypotheses proposed. Nearly all of the significant correlations are minor.

On the positive side, the authors do offer reasonable discussion of the results that run counter to expectations and prior research. For example, they suggest possible the lack of significant expertise-helpfulness correlation may come from the differences in how expertise is defined from study to study (p. 51). They also speculate that the absence of a clear connection between reciprocity and helpfulness, which contradicts prior research in face-to-face settings, may stem from the lack of face-to-face interaction and any direct, personal reciprocity in electronic networks (p. 51). They also rightly note the weakness of even their strongest finding, the significance of centrality. They acknowledge that centrality may well be a dependent variable, something that results from rather than predicts knowledge contribution (p. 53).

Overall, the authors offer a well-structured and rigorous research project. The disappointing conclusions reflect not a failing of the researchers but the complicated and still not fully understood nature of how self-organized networks manage to promote knowledge sharing even without the typical motivations of knowledge management and the market.

Advertisements