Lecture 4 Flashcards
Two ways of delimiting a graph
- Normalist: determined by criteria of researchers
- Realist: network as recognized by its actors
… data: analysis of variables
Attribute
… data: network analysis
Relational
One mode network
Network comprised of one category of actors, e.g. companies, individuals
Two-mode network
Network comprised of two categories of nodes, e.g. Individuals & events, companies & directors
Data collection sources: survey research
- Row-based: relations going to other actors
- Row and column based: relations coming from and going to other actors
- Consensus method: indicate relations among every pair of actors
Data collection sources: ethnographic research
Observer records all interactions and relations that take place in front of observer
Data collection sources: documentary research
Gather data from written/stored documents, databases
Data collection sources: experimental research
Data gathered under experimental conditions, e.g. computer simulations
Adjacency matrix
Squared, one mode (m • m)
Affiliation/incidence matrix
Rectangular, two-mode (m • n). Can extract information to form adjacency matrix (m • m OR n • n)
Degree centrality
Number of incoming (indegree) and outgoing (outdegree) linkages against rest of nodes. Local centrality measure
Closeness centrality
Centrality based on closeness to all other nodes, against other nodes’ closeness. Global centrality measure
Betweenness centrality
Centrality based on how many geodesic paths cross the node, against other nodes’ betweenness centrality. Global centrality measure
Status
Total weighted number of paths reaching a node, compared to its eigenvector
