Social Networks: An Introduction

1
Social Networks An Introduction

• Jeroen Bruggeman, University of Amsterdam, Fall
  2008
• Chapter 2, network representation and
  conceptualization

2

• Some animals are social, in particular monkeys
  and apes they learn some of what they know by
  imitation cooperate mostly with kin have simple
  culture
• Humans learn most of what they know from others
  collaborate (e.g. exchange) also extensively with
  non-kin have complex culture, in evolving
  figurations (Elias) of social relations
• How can we comprehend complexity of social life?

3
Zoom in on dyad

• Two people in a social relationship have
  memories and expectations, usually based on their
  interactions (but sometimes on fantasy),
  associated with emotions, and molded by norms,
  gossip and other information from their social
  environment (except for newborns)

4

• There exists a large variety of social
  relationships, in kind and strength.
• Strength of positive relations varies with (1)
  emotional intensity, (2) trust expectation that
  other wont do harm, (3) time spent, and (4)
  reciprocity (Granovetter)
• There are negative relations, too
• How can we sensibly deal with this clutter of
  social relations?

5
Represent them as network!

• Actors (humans, or corporations) as vertices
• Dyadic relationship as tie (line connecting a
  vertex with another one)
• Strength of relationship as numerical value of
  tie
• Asymmetry of relationship by arc edge for
  (approximately) symmetric rel.
• Kind of relationship by tie color (or other
  marker)
• Network (graph) is set of vertices and ties.

6
12
Sexual relations at Jefferson High School
(Bearman cum suis)
7
Advantage network representation

• Systematic overview and clarity far beyond
  unaided comprehension of participants and
  researchers alike

8

• Two people connected by tie are neighbors, based
  on their social, not physical, proximity.
• At Jefferson 288 children are in a connected
  sub-network where each child can reach every
  other one through a path (concatenation of ties)
  they are in a component, a maximal set of nodes
  that is connected no connected node is excluded
  from it, and delimits where something, e.g. a
  sexual disease, can spread.
• A component is not a clique a fully connected
  network of 3 or more vertices

9

• If, along with sexual relations, also other kinds
  of relations (e.g. friendship) are represented in
  a network, its a multi(plex) graph.
• Sum of multiple ties as overall tie strength
• A network of a single person (ego) and his/her
  social contacts (alters) is called ego network
• Total number of egos contact persons is called
  degree (for edges), irrespective of tie values
• Asymmetric relations (e.g. phone calls) indegree
  for incoming arcs outdegree for outgoing arcs.
  Degree indegree outdegree

10
Network data file

• John Mary 1
• Mary John 1
• John Peter 2
• .
• Each social relation (e.g. friendship) as a set
  of ordered triples the order matters!
• Symmetric tie as two opposite arcs, or as edge
• Relation establishes topology of network
  different graphical representations with same
  topology of the same network

Peter
John
2
11
Network data file

• Matrix as alternative but equivalent form
• Diagonal (usually) filled with zeros

12
Data and observations

• Memories, emotions, and expectations of subjects
  mostly unobserved by social researcher
• Ties as interpretations of (possibly ambiguous)
  social behavior see Ch 8
• Choice of threshold for duration/intensity of
  behavior to distinguish ties from ephemeral
  encounters depends on research goal, e.g. close
  friends versus soccer match

13
Exercise 2.1 ego network

• Choose about 20 of your social contacts,
  preferably from several different social circles,
  e.g. family, sport, work, etc.
• Think about a threshold for ties, and about tie
  values and direction
• Depict your ego network with igraph see Ch 2
  handout

14
Conceptualization

• Networks look cute, but do not by themselves
  provide understanding of social mechanisms
• Social theory is required!
• Challenge of conceptualization thoughtful
  explication of theoretical concepts in terms of
  networks (or network models)

15
Example social cohesion

• Social ties that hold group together more
  extensive treatment in Ch5
• Intuition more ties (m), more cohesion
• Threshold low, symmetry irrelevant more or less
  stable ties through which gossip norms can
  pass.
• Custom to compare with maximum possible, for
  given group with n people ½ n(n-1)
• Density 2m/n(n-1)

16
Example (Moody White)
17

• Problem groups networks tend to be clustered
  into smaller groups, and density within clusters
  (not a precisely defined concept) is higher than
  between, while density is blind for inhomogeneity
  as well as for some other things to be discussed
  in Ch5
• Density is a very useful notion, but an awkward
  measure of cohesion.
• In general, we need both precision and
  sociological imagination!