Title: The application of Social Network Analysis to Knowledge Management
1The application of Social Network Analysis to
Knowledge Management
- Peter Busch Debbie RichardsDepartment of
ComputingMacquarie UniversityAustralia
2Goals
- Knowledge Management
- Test for articulable Tacit Knowledge (aTK) in
individuals - articulable implicit managerial IT knowledge
- Map intra-organisational diffusion of aTK among
IS personnel - Social Network Analysis
- Knowledge flows/bottlenecking
3Research questions
- 1. Are there observable tacit knowledge
differences between how experts handle the
tacit knowledge issues in the organisation from
those of novices? In other words how do experts
differ in their approaches to those of novices? - 2. Can we identify other tacit knowledge rich
personnel based on the similarity of their
answers with that of the expert group? - 3. Are there certain biographical parameters
(i.e. age, gender, ethnicity, years of IT
experience, ACS level, highest formal
qualification) that differentiate IS individuals
who have accumulated more tacit knowledge from
those with significantly less tacit knowledge? - 4. Do people clique with one another based on
biographical factors such as ethnicity? If so,
does it affect tacit knowledge transfer? - 5. Is there evidence of tacit knowledge
bottlenecking taking place? - 6. Are there observable differences in knowledge
diffusion patterns between IS personnel depending
upon the character of the organisation?
4Highest occurrence (9)
5Tacit knowledge and Groups
6Tacit knowledge and Competition
7Tacit knowledge and Social Environment
8The knowledge hierarchy(Busch and Dampney 2000)
9Testing for tacit knowledge
- Tests
- explication based
- psychological domain
- individualistic
- positivist
- not student populations
10Approaches to testing
- Larkin (1980)
- Physics problem solving
- Benner (1984)
- US Air Force
- Scott (1990 1992)
- Nursing mothers
- Reber (1993)
- Anagram puzzles
- Reed, Hock and Lockhead (1993)
- length of images
- Noh (et.al. 2000)
- AI/cognitive maps
- Case Based Reasoning (CBR)
- Herbig, Büssing and Ewert (2001)
- nursing
- Delphi
- simulation technique (Frederiksen 1966)
- Sternberg (et.al. 1985-)
- critical incident technique (interviews)
- simulation approach (observation)
11Tacit knowledge inventory
- 3 components
- Electronic questionnaire
- Javascript
- Cgi
- Perl
- Html
- Java
- 1. Biographical -gt
- 2. Social Network Analysis (SNA)
- 3. The inventory itself
12Result comparisons ..
0
- Experts vs. others
- colleagues asked to choose
- concept of proficiency
- how did experts answer?
- Likert scale data
- experts answering differently?
- which particular scenarios and answer options
different? - which scenarios greatest degree of variation
ethical/realistic answers - Identification of expert non-experts
13Organisation X
14Org. X Language other than English
Concept lattice illustrating language other than
English
15Organisations Y and Z
16Tacit knowledge sample scenario
Scenario 8, answer 4 of the IS articulable tacit
knowledge inventory
17General differences
0
- Experts
- greater awareness of status related issues
- problems related to avoiding responsibility for
issues - non-experts seem to be a little happier to pass
the buck - going over a superiors head less of a good idea
- taking issues outside of the organisation is less
of a good idea - see less wrong with showing superior better way
of doing things - better idea of practical logistics, what may be
good in theory, not good in practice - Little more guilt if system implemented, not
quite what it could have been
18Ethical differences
0
- Experts ..
- seem more comfortable with telling superiors that
their plate is full enough - if they are to be given extra work
- then it is best not to grumble but tackle the
task at hand - less enthusiastic about passing responsibility
for tasks onto others
- would prefer to be noticed for working harder
- they do seem less comfortable with being a yes
person - seem to feel less at ease with covert means of
gaining advantage - more responsible for a projects success
19Realistic differences ..
0
- Experts
- seem far happier in practice to say they are
already overcommitted - seem far less comfortable in practice questioning
the decisions of a superior - agreeing with superiors if the task at hand needs
to be done, seems to rate more highly with
experts in practice - seem more content in practice with offering
management alternatives if they know of them - appear less inclined practically to want to
achieve means with ulterior motives
- less inclined to pay lip service
- reduced interest in passing the buck
- being leaders from a practical point of view,
rather than followers - Less interested in unpaid overtime
- less interested in practice in asking
subordinates what extra help they may require - reluctant to commit itself to exercises where
outcomes are likely to be less clear
20General differences
0
210
220
23Significant ethical differences between the two
groups
0
240
250
26Significant realistic differences between the two
groups
0
270
280
29SNA questionnaire component
30Participant observation
- Sliding scale (Leedy 1997)
- direct observation lt-gt direct participation
- Physical presence in the workplace
- Unsuitable occupation type?
- Modification of behaviour
31Social Network Analysis(underpinnings)
- Actors or participants in the system viewed as
interdependent upon one another - rather than independent
- Relations among actors
- considered as channels or thoroughfares of
resources - Interaction among actors is directly constrained
or aided - by the structure of the relationships themselves
- Relations that take place between the actors
determine all - economic, political and social structures
32Social Network Analysis
A sociomatrix
- Whole of network/egocentric
- Sociomatrices
- binary vs. valued
- symmetrising
- Degree
- in- and out-
- Prestige/Prominence vs. Centrality
- in- and out- degree
- betweenness of actors
- (those who are pathways to others)
- Density and inclusiveness
The in- and out-degree of 4 actors
33Social Network Analysis
- Sociograms
- with/without multidimensional scaling
- Cliques
A sociogram without multidimensional scaling
Cantonese clique
A sociogram with multidimensional scaling
34SNA - Cliques
Staff who meet hourly
351. Size Size of ego network 2. Ties Number of
directed ties 3. Pairs Number of ordered
pairs 4. Density Ties divided by Pairs 5.
AvgDis Average geodesic distance 6. Diameter
Longest distance in egonet 7. EgoBetween
Betweenness within the egonet 8. UnReach of
ordered pairs with infinite distance
36Organisation X
Staff who meet each other weekly
37Organisation X (senior personnel)
1046 has 50 subordinates1246E has 40-44
subordinates1719E has 25 to 29 subordinates1534
and 12E have 10-14 subordinates
Now including people with 5 to 9 subordinates
38Organisation X
Front-office staff
Back-office staff
39Organisation X
With contractors
Without contractors
40Avoiding one another
41Can get by without seeing
42Organisation Y
43Organisation Y
44Organisation Z
45Organisation Z
46Conclusion
- Participant observation has limitations
- SNA represents a viable alternative
- can help identify flows/bottlenecks
- Reliance on IT
- can be detrimental for tacit knowledge transfer
- Size of the organisation
- generally small is preferable