Lecture Summary 0HM220
For the exam:
Articles are relevant for the arguments, not the details. What comes out of the test.
Articles: Theory arguments and tests of the theory
Meeting 1
Landmark studies
The sociogram: Jacob Moreno
Sociogram -> depiction of a social network.
Social network -> links between people
Hawthorne studies
Bank wiring room study:
Focus on a few people.
Cliques in which people create norms.
Super connectors
AID virus
Illegal networks
Trace potential problems, figure out who to target.
You want to reduce the network.
Contagion
Physical or psychological, information or imitation -> transmission from one actor to another.
Structural balance theory
Sentimental relationships:
(A1) A friend of a friend is a friend,
(A2) A friend of an enemy is an enemy,
(A3) An enemy of a friend is an enemy,
(A4) An enemy of an enemy is a friend -> incorrect -> An enemy of an enemy is not a
friend but more likely an enemy as well
Six degrees: millgram study
Ask to connect -> 6 steps were needed for successful chains.
Small worlds
,By adding random links between caves a small world arises.
Network Jargon
Networks are made up of nodes and relationships between those actors.
Vertex/node Smallest unit in a network
Link Link between two vertices in a network
Loop connects a vertex to itself
Line Undirected
Directed -> arcs
Degree
Degree of a vertex represents the number of links it has to other vertices.
Degree in directed networks
Indegree incoming arcs (people that love you)
Outdegree outgoing arcs (people you love)
X influences Y -> outdegree measures power, while indegree measures dependency
X likes Y -> indegree measures popularity, outdegree measures agreeableness
Average degree
Indication of how connected the network is
Degree distribution
Probability that a randomly selected vertex in a network has degree k.
Adjacency matrices
Undirected network-> matrix is always symmetric
Friendship Paradox
First number is the number of friends
Second number is the mean number of friends of her friends
Assume they base their decisions on the majority.
Network position
Importance of position in the network
The Italian elite
Trade, marriage etc, but the people are not connected their self.
, Centrality measures
Indicators of centrality:
Degree centrality
How central and how powerful -> number of connections you have
Medici has the most connections.
Betweenness centrality
How much power do you have? -> people have to go through you to reach others
How often are you between others.
1. For every pair (except A) of nodes calculate the number of shortest paths (geodesics)
between the pair
2. For every pair of nodes calculate the number of times a shortest path goes through A
3. For each pair divide the number obtained in 2 by the number obtained in 1
4. Calculate the sum of all 3’s
Normalized betweenness centrality
Division by (n-1)/(n-2)/2
Eigenvector centrality
How well can I reach others?
Eigenvector centrality-> centrality measure that includes how well-connected friend (and their
friends) are.
Meeting 2 The basic Network
Arguments
1. Why do
2. Which networks characteristics
3. Have what effect?
Beneficial network characteristic (of overall network): Network closure beneficial for collaboration
between companies. High network closure leads to more trust -> safeguard against opportunism.
Less likely to cheat.
Network diversity (of a single actor): High degree of network diversity beneficial because it provides
access to more diverse resources, innovative ideas.
For the exam:
Articles are relevant for the arguments, not the details. What comes out of the test.
Articles: Theory arguments and tests of the theory
Meeting 1
Landmark studies
The sociogram: Jacob Moreno
Sociogram -> depiction of a social network.
Social network -> links between people
Hawthorne studies
Bank wiring room study:
Focus on a few people.
Cliques in which people create norms.
Super connectors
AID virus
Illegal networks
Trace potential problems, figure out who to target.
You want to reduce the network.
Contagion
Physical or psychological, information or imitation -> transmission from one actor to another.
Structural balance theory
Sentimental relationships:
(A1) A friend of a friend is a friend,
(A2) A friend of an enemy is an enemy,
(A3) An enemy of a friend is an enemy,
(A4) An enemy of an enemy is a friend -> incorrect -> An enemy of an enemy is not a
friend but more likely an enemy as well
Six degrees: millgram study
Ask to connect -> 6 steps were needed for successful chains.
Small worlds
,By adding random links between caves a small world arises.
Network Jargon
Networks are made up of nodes and relationships between those actors.
Vertex/node Smallest unit in a network
Link Link between two vertices in a network
Loop connects a vertex to itself
Line Undirected
Directed -> arcs
Degree
Degree of a vertex represents the number of links it has to other vertices.
Degree in directed networks
Indegree incoming arcs (people that love you)
Outdegree outgoing arcs (people you love)
X influences Y -> outdegree measures power, while indegree measures dependency
X likes Y -> indegree measures popularity, outdegree measures agreeableness
Average degree
Indication of how connected the network is
Degree distribution
Probability that a randomly selected vertex in a network has degree k.
Adjacency matrices
Undirected network-> matrix is always symmetric
Friendship Paradox
First number is the number of friends
Second number is the mean number of friends of her friends
Assume they base their decisions on the majority.
Network position
Importance of position in the network
The Italian elite
Trade, marriage etc, but the people are not connected their self.
, Centrality measures
Indicators of centrality:
Degree centrality
How central and how powerful -> number of connections you have
Medici has the most connections.
Betweenness centrality
How much power do you have? -> people have to go through you to reach others
How often are you between others.
1. For every pair (except A) of nodes calculate the number of shortest paths (geodesics)
between the pair
2. For every pair of nodes calculate the number of times a shortest path goes through A
3. For each pair divide the number obtained in 2 by the number obtained in 1
4. Calculate the sum of all 3’s
Normalized betweenness centrality
Division by (n-1)/(n-2)/2
Eigenvector centrality
How well can I reach others?
Eigenvector centrality-> centrality measure that includes how well-connected friend (and their
friends) are.
Meeting 2 The basic Network
Arguments
1. Why do
2. Which networks characteristics
3. Have what effect?
Beneficial network characteristic (of overall network): Network closure beneficial for collaboration
between companies. High network closure leads to more trust -> safeguard against opportunism.
Less likely to cheat.
Network diversity (of a single actor): High degree of network diversity beneficial because it provides
access to more diverse resources, innovative ideas.
