Network Numerology: Demystifying Numbers in Social Network Analysis ‘Simply because your data links people and you can visualize that, it does not mean you have performed network analysis. This is akin to displaying a line plot of some stock's price over a quarter and claiming you have performed statistical analysis – all you have done is report data! As with all other statistical processes, network analysis is meant to draw meaning and inference from the structure, which requires an understanding of these methodologies, their strengths and limitations’. Drew Conway, Political Scientist, 2009.
33
Embed
Network Numerology: Demystifying Numbers in Social Network ... · Network Numerology: Demystifying Numbers in Social Network Analysis ‘Simply because your data links people and
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Network Numerology:Demystifying Numbers in Social Network Analysis
‘Simply because your data links people and you can visualize that, it does not mean you have performed network analysis. This is akin to displaying a line plot of some stock's price over a quarter and claiming you have performed statistical analysis – all you have done is report data! As with all other statistical processes, network analysis is meant to draw meaning and inference from the structure, which requires an understanding of these methodologies, their strengths and limitations’.
Which is why most Knowledge Management projects fail, many organisations adopt the Slothfulness and Mendacious strategy, and HyperEdge specialises in Networks and Visualization!
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z
‘Each of us is part of a large cluster, the worldwide social net, from which no one is left out. We do not know everyone on this globe, but it is guaranteed that there is a path between any two of us in this web of people’.
Dunbar’s Numbers are an indicator of meaningful relationships and the maximum effective number of people in a network. The usually accepted number is 153. There is an mega-band number of around 700, and an upper limit of about 1,500 – see Dunbar, R 2010, How many friends does one person need? Dunbar's number and other evolutionary quirks., Faber and Faber, London.
‘Whatever a central management imposes, informal networks develop in ways that shape how an organisation works. These multiple networks involve information-flow, knowledge transfer, work cooperation, support, friendship and antagonisms. They are crucial to organisational functioning’.
Professor Garry Robins, Network Scientist, Melbourne University, 2006
Communication in practice
A 2011 study of 2,500 participants by the Massachusetts Institute of Technology found that the most important predictor of team success is in its communication patterns.
Of note the study found that:– communication patterns are as significant as all other factors, including intelligence,
personality, and talent combined;
– researchers could foretell which teams would out-perform the others simply by looking at the data on their communication patterns, even without meeting the team members;
– connectivity, activity, and energy were the key communication dynamics that enabled or effected performance;
– mapping communication behaviours over time, and making small adjustments to move it closer to the ideal, dramatically improves team performance.
Pentland A, ‘The New Science of Building Great Teams’, Harvard Business Review, April 2012
‘In all businesses there are two organisations: one that is shown on the formal organisation chart and another that exists in reality. The latter is made up of not job titles or formal lines of authority, but rather influencers and other individuals.’
Doctor Neil Farmer, Network Scholar and Author, 2008
Reveals how much activity is going on and who are the most active members by counting the number of direct links each person has to others in the network.
Does not necessarily describe power or influence.
People at the centre of the network:• are the connector or hub of
the network,• may be in an advantaged
position in the network.• are usually less dependent
on other individuals.• are often a deal maker or
broker.
Providers and Seekersdegree centrality
1
n
i ijj
k A
Where ki is the degree of node i; n is the number of nodes; Aij is an adjacency matrix; and ijdenotes a tie between nodes iand j.
1
nini ij
j
k A
In-degree is the number of ties directed towards the node.
1
noutj ij
i
k A
Out-degree is the number of outgoing ties from the node.
Highlights people with the shortest paths to other people, thus allowing them to directly pass on and receive communications quicker than others in the organisation.
Is strongly correlated with organisational influence if the individual is a skilled communicator.
These individuals are often network brokers. They are often the ‘pulse-takers’ of the organisation.
Transmitters and Receiverscloseness centrality
1i ij
j
l dn
Where li is the mean distance; n is the total number of nodes; and dij is the length of the shortest path between nodes iand j in a matrix.
• Closeness centrality begins with the assumption that having short paths to other nodes increases the influence in the network of that node.
• It measures the average distance a node is from all other nodes in a network, and therefore is a proximity measure.
• Unconnected nodes by definition have an infinite distance between them, which means scores cannot be computed for isolated nodes.
• Closeness centrality requires the network, or at least the component under examination, to be complete.
Reveals individuals who:• connect disparate groups
within the network.• hold a favoured or
powerful position in the network.
• have great influence over what is communicated through the network.
• act as intermediaries
Identifies the bridges within the network. They may act as the true gatekeeper deciding what does or does not get passed through the network, or as the “third who benefits” by passing information to others to secure advantage..
Bridges(brokers and gatekeepers) - betweenness centrality
ististst
nxg
Where xi is the betweenness of node i; is the number of paths from node s to node t that pass through node i; and gst is the number of paths from node s to node t.
• Betweenness centrality measures the extent that a node lays on the path of other nodes.
• Betweenness centrality is unlike other centrality measures because it does not measure how well the node in question is connected, but rather how it connects components of the network.
• It is a proxy for understanding strategic position within the network.
• It can be applied to both directed and undirected networks.
Measures how well connected a person is and how much direct influence they may have over the most active people in the network
Measures how close a person is to other highly connected people in terms of the global or overall makeup of the network
Is a reasonable measure of “network positional advantage” and/or perceived power.
Influencerseigenvector centrality
11i ij j
j
x k A x Where xi is the centrality of each node i; k is the eigenvalue, with 1 being the largest and -1 the smallest; Aij is an adjacency matrix; and ij denotes a tie between nodes i and j.
• Eigenvector centrality begins with the assumption that having connections with other central nodes increases the relative importance of that node.
• A high eigenvector centrality score means the node is important because either it is connected to many nodes, or is connected to a few very highly connected nodes
• Eigenvector centrality has the limitation that it works best on undirected networks.
Gatekeeper - a person who transmits information and other resources to the same group or team from sources external to that group or team.
Representative - a person who transmits information and other resources from their group or team to an external group or team.
Liaison - a person who transmits information and other resources from one group or team to another group or team, whilst themselves belonging to a different group or team.
Coordinator - a person who brokers connections within the same group or team.
A CBConsultant - a person who intermittently takes the central lead by connecting others in the same group or team, but who belongs to another group or team.
If an individual only sends messages and receives none then their contribution index is +1.000If an individual only receives messages and sends none then their contribution index is -1.000If the communication behaviour is balanced then the contribution index is 0.000
ContributionFrequency
ContributionIndex
Sender +1
Receiver +1
Expert
Envoi
Escort
Expediter
Gloor, P 2006, Swarm creativity: Competitive advantage through collaborative innovation networks, Oxford University Press, Oxford.
messages sent – messages received
messages sent + messages receivedContribution Index =
1. The links inside the “circles” are posts between like roles. Note there are no posts between Experts.2. The thicker curves linking groups are consolidated exchanges between groups. They do not show frequency, or links from one
individual to another.3. Note the relative density in the Escort and Expediter groups.
‘A good deal of the corporate planning I have observed is like a ritual rain dance; it has no effect on the weather that follows, but those who engage in it think it does. Moreover, it seems to me that much of the advice and instruction related to corporate planning is directed at improving the dancing, not the weather’
Emeritus Professor James Brian Quinn, Tuck School of Business, Dartmouth College, 1980.
Summary.
Social network analysis, done properly, provides:
– a powerful quantitative, qualitative, and visual diagnostic,– empirical information on the “real or shadow” structures and relationships
in an organisation,– a means to reach shared understanding and common meaning,– a baseline for organisational and personal improvement.
The key is “done properly”! You cannot escape the mathematics!
Use the right tool and presentation for the job, and remember visualisation is not analysis.
Whatever your approach ensure you have multiple lines of evidence. For example, narrative provide additional granularity and allow for data triangulation and validation.
Above all else you must understand your organisation, the data, the resultant network and visualisations, and the assumptions you are making.