Soc
ial N
etw
ork
Ana
lysi
s
12 -
Session 2 -
Social Network AnalysisConcepts and terminologyExamples of application in
Public Health
Manila, PhilippinesManila, Philippines28-30 April 200828-30 April 2008
Steeve Ebener, WHOSteeve Ebener, WHO
Soc
ial N
etw
ork
Ana
lysi
s
22 -
2Content1. What is a network ?2. Network types
• Mode• Ego / Complete network
3. Ego network1.Measures
4. Complete network1. structure2. Measures (centrality,
centralization, cohesion)5. Efficient network form ?6. Examples of application in Public
health
1
3
Soc
ial N
etw
ork
Ana
lysi
s
32 -
What is a Network ?• A set of actors (node, points)
− individuals (e.g. persons)− collectivities (e.g. firms,
nations,...)
• A set of ties (links, lines, edges, arcs) of a given type that connect pairs of actorsSet of ties of a given type constitutes a social relation
Borgatti
Soc
ial N
etw
ork
Ana
lysi
s
42 -Borgatti
What is a Network ?Taxonomy of network ties among
persons
Soc
ial N
etw
ork
Ana
lysi
s
52 -Borgatti
What is a Network ?Directed vs undirected ties
• Undirected relations (ties)– attended meeting with– Communicate daily with
• Directed relations (ties)– Lent money to
Soc
ial N
etw
ork
Ana
lysi
s
62 -Borgatti
What is a Network ?Strength of tie
• We can attach values to ties, representing quantitative attributes– Strength of relationship– Information capacity of tie– Rate of flow or traffic across tie– Distance between node– Probabilities of passing on information– Frequency of interaction
Soc
ial N
etw
ork
Ana
lysi
s
72 -
Network types
Borgatti
Soc
ial N
etw
ork
Ana
lysi
s
82 -
Network composed of two types of social entities (say persons and organizations)
Network types
Network composed of only one types of social entities (e.g. persons or organization).
• 1 mode network
• 2 mode network
offers interesting analytical possibilities for gaining a greater understanding of "macro-micro" relations
Soc
ial N
etw
ork
Ana
lysi
s
92 -Borgatti
• In organization studies
− Micro refers to studies in which the actors are persons− personality -> Status
− Macro refers to studies in which the actors are firms− Firm size -> Profits
• But in network research...
Network typesBeware the Micro/Macro distinction
− Micro means focus on actors− which could be firms...
− Macro means focus on the network in which actors are embedded
Soc
ial N
etw
ork
Ana
lysi
s
102 -
All the actors of a network with all the ties which exists among them.
Network types
A focal actor (therespondent, calledego), together with the actor's contacts (called alters), and often, a limited set of ties among the alters.
• Ego network
• Complete network
Borgatti
Soc
ial N
etw
ork
Ana
lysi
s
112 -
Network types
Borgatti
Soc
ial N
etw
ork
Ana
lysi
s
122 -
Ego Network
• PN Exposure
Hanneman
Just picked one:
Measures
Soc
ial N
etw
ork
Ana
lysi
s
132 - Valente
Ego Network
Degree to which an individual is exposed to an innovation through his/her personal network.
Network exposure provides:1. awareness information2. influence/persuasion 3. detailed information on how to get the
innovation, possible problems, updates, refills, enhancements, novel uses
4. something to talk about5. ...
Personal Network (PN) Exposure
Soc
ial N
etw
ork
Ana
lysi
s
142 -
14
PN Exposure=33%PN Exposure=33% PN Exposure=66%PN Exposure=66% PN Exposure=100%PN Exposure=100%
= Non User= Non User = User= User
Valente
Ego NetworkPersonal Network (PN) Exposure
Soc
ial N
etw
ork
Ana
lysi
s
152 -
Complete NetworkStructure (shape)
Large variety
two examples:• Core/Periphery structure
− Network consists of single group (a core) together with hangers-on (a periphery)− Core connects to all− Periphery connects only to the core
− Short distances, good for transmitting information− Identification with group as whole
• Clique structure− Multiple subgroups of factions− Identity with subgroup− Diversity of norms, belief
Borgatti
Soc
ial N
etw
ork
Ana
lysi
s
162 -
Complete NetworkTwo main measures
• Cohesion– Density– Distance– Transitivity – ...
• Centrality & Centralization– Degree centrality– Eigenvector centrality– Closeness centrality– Betweeness centrality– Centralization– ...
Soc
ial N
etw
ork
Ana
lysi
s
172 -
Complete NetworkCentrality & Centralization• Networks Analysis can be used to
identify positions/role in the network
• One significant position is finding the person or people at the center
• Centrality has been a central preoccupation with network analysis
• There are many ways to identify central persons (centrality) and rate who near the center each node is.
Valente
Soc
ial N
etw
ork
Ana
lysi
s
182 -
Complete NetworkCentrality & CentralizationLocal centrality vs. global
centrality
• Local centrality - when a point has a large number of connections to other points in its immediate environment.
• Global centrality - has a strategic position within the network.
Centralization • overall centrality of the
network
Valente
Soc
ial N
etw
ork
Ana
lysi
s
192 - Borgatti
Complete NetworkCentrality
Soc
ial N
etw
ork
Ana
lysi
s
202 -
Complete NetworkCentrality• Degree: the most active or
popular node in the network• Closeness: the most independent
node and the one in excellent position to monitor the information flow in the network.
• betweenness: The most powerful broker or gatekeeper in the network
• Eigenvector: the most popular ("in the know") node in the network
Soc
ial N
etw
ork
Ana
lysi
s
212 -
Decentralized Centralized
Valente
Complete NetworkCentralization
Soc
ial N
etw
ork
Ana
lysi
s
222 - Valente
Complete NetworkPeripheral players• Most people would view the nodes on the
periphery of a network as not being very important. In fact, since individuals' networks overlap, peripheral nodes are connected to networks that are not currently mapped.
• These nodes might have their own network outside of the company - making them very important resources for fresh information not available inside the company!
Soc
ial N
etw
ork
Ana
lysi
s
232 - Borgatti
Complete NetworkTies as conduits• Certain ties can serve as pipes or
roads that enable flows/traffic (and in their absence, prevent it)
• information, solutions, material aide, resources
• Attitudes, behaviors, practices • Interpersonal models of
diffusion, influence
Soc
ial N
etw
ork
Ana
lysi
s
242 - Borgatti
Complete NetworkFlow processes• Path: can't repeat node
– 1-2-3-4-5-6-7-8– virus: host become immune
or die• Trail: can't repeat line
– 1-2-3-1-7-8– gossip
• Walk: unrestricted– 1-2-3-1-2-7-1-7-1– Dollar bill moving through
economy
Soc
ial N
etw
ork
Ana
lysi
s
252 -
Complete NetworkCutpoints and Bridges
Borgatti
• Cutpoint: a node which, if removed, would increase the number of components
• Bridge: a tie that, if removed, would increase the number of components
• If a tie is a bridge, at least one if its endpoints must be a cutpoint
Soc
ial N
etw
ork
Ana
lysi
s
262 -
Complete NetworkCutpoints and Bridges
Borgatti
• Bridge: a tie that, if removed, would increase the number of components
Soc
ial N
etw
ork
Ana
lysi
s
272 -
Complete NetworkOther node related concepts• Dyad - two individuals (as
husband and wife, mother and son) maintaining a sociologically significant relationship
• adjacent - two points connected by a line
• neighborhood - set of adjacent points
Valente
Soc
ial N
etw
ork
Ana
lysi
s
282 - Borgatti
Complete Network
Length & Distance• Length of a
path/trail/walk is the number of links it has
• Distance between two nodes is the length of the shortest path
Cohesion
They rules: http://www.theyrule.net/
The shortest the average distance the faster the information flows
Soc
ial N
etw
ork
Ana
lysi
s
292 - Borgatti
Complete Network
DensityCohesion
b) is more dense than a)Proportion of pairs of actors that are actually tied
The higher the density of the group, the more at risk (disease spread)
Soc
ial N
etw
ork
Ana
lysi
s
302 - Borgatti
Complete Network
Transitivity
Number of transitive triples divided by the number of potential ones (number of paths of length 2)
Cohesion
The higher the percent of transitive triads the more cohesive the network
Soc
ial N
etw
ork
Ana
lysi
s
312 - Valente
Efficient network form ?• Finding efficient forms may
completely depend on behavior being studied– a dense network facilitate information
flow but also diseases transmission• What is the trade off between
individual satisfaction and network-level performance?
• Can optimal forms be created or are the most optimal ones those that exist?
Soc
ial N
etw
ork
Ana
lysi
s
322 -
Examples of application in PHAnalysis of the DTTB program
"To enhance the social and professional support system of the DTTB participants, with the end view of improving diagnostic capability and facilitate their integration into a new health network"
Objective
Several phases study:•Develop a better understanding of the knowledge flows and gaps of DTTBs•Analyze the connection and speed of transfer of information within the health system. •Determine the impact of geography and connectivity (ICT) on the DTTBs care delivery capacities.
Soc
ial N
etw
ork
Ana
lysi
s
332 -
Examples of application in PHAnalysis of the DTTB program
Process Logbook
SNA software
Analysis
Soc
ial N
etw
ork
Ana
lysi
s
342 -
Examples of application in PHRelations among drug injectors
Borgatti
Which two people should be isolated in order to slow spread of HIV ?Bridges between groups represent key pathways for disease spread.
Soc
ial N
etw
ork
Ana
lysi
s
352 -
SOURCE: James Moody. http://www.soc.sbs.ohio-state.edu/
Examples of application in PHColorado Springs Sexual Contact Network
Soc
ial N
etw
ork
Ana
lysi
s
362 -
Examples of application in PHHealth promotion
• Network data may help improve health promotion programs
• The Messenger is the Message• Identify opinion leaders (over 15 studies)• Identify leaders and match them to
nominees• Identify groups and find leaders within
Snowball/Respondent driven sampling for recruitment (VPS)
Valente
Soc
ial N
etw
ork
Ana
lysi
s
372 -
Examples of application in PHCommunity of practice building
Mapping the extend of a new community of practice and identify knowledge hubs through snow ball survey.
Potentialleaders for thiscommunityMeasuring the success of communitiesof practice
Soc
ial N
etw
ork
Ana
lysi
s
382 -http://www.orgnet.com/experts.html
Examples of application in PHKnowledge continuity analysis
Soc
ial N
etw
ork
Ana
lysi
s
392 -
Examples of application in PHAssessing the current state of collaboration
Soc
ial N
etw
ork
Ana
lysi
s
402 -
Examples of application in PHVisualisation
Soc
ial N
etw
ork
Ana
lysi
s
412 -
• Information and persuasion flow through social networks.
• Inter-organizational collaboration, coordination, and cooperation.
• Analysis of the surveillance network efficiency
• ...
Valente
Examples of application in PHand some more...
Top Related