Title: Joint social selection and social influence models for networks: The interplay of ties and attributes.
1Joint social selection and social influence
models for networks The interplay of ties and
attributes.
- Garry Robins
- Michael Johnston
- University of Melbourne,
- Australia
- Symposium on the dynamics of networks and
behavior - Slovenia, May 10-11, 2004
- Thanks to Pip Pattison, Tom Snijders, Henry Wong,
Yuval Kalish, Antonietta Pane
2- A thought experiment
- Most models that purport to explain important
global network properties are homogeneous across
nodes. - Might a simple model of interactions between
node-level and tie-level effects be sufficient to
explain global properties? - Develop a model that incorporates both social
selection and social influence processes. - Which global properties of networks are
important? - Simulate the model to see whether the these
properties can be reproduced in a substantial
proportion of graphs?
3- 1. Develop a model incorporating both social
selection and social influence effects
4Simple random graph models
- For a fixed n nodes, edges are added between
pairs of nodes independently and with fixed
probability p - (Erdös Renyi, 1959)
Bernoulli random graph distribution
X is a set of random binary network variables
Xij Xij 1 when an edge is observed, 0
otherwise x is a graph realization ? is an edge
parameter. an exponential random graph (p) model.
a homogeneous model (node homogeneity) p and ?
are independent of node labels
5A Bernoulli random graph model will not fit this
network well
6In this example, actor attributes are important
to tie formation
Social selection
7In this example, actor attributes are important
to tie formation
Social selection
Yellow Jewish Blue Arab (Kalish, 2003)
Exogenous attributes affect network ties
8Binary variables Xij network ties Zi
actor attributes
Exogenous attributes affect network ties
Zi
Xij
Zj
Robins, Elliott Pattison, 2001
9Effects in the model
Baseline edge effect irrespective of attributes
Propensity for actors with attribute z1 to have
more partners
Propensity for ties to form between actors who
both have attribute z1
Equivalent (blockmodel) parameterization
10Social influence Are actor attributes
influenced by fixed network structure?
Robins, Pattison Elliott, 2001
11Social influence Are actor attributes
influenced by fixed network structure?
Robins, Pattison Elliott, 2001
12Social influence Are actor attributes
influenced by fixed network structure?
A cutpoint
Exogenous network ties affect attributes
13Binary variables Xij network ties Zi
actor attributes
Exogenous network ties affect attributes
Zi
Xij
Zj
14Effects in the model
Baseline effect for number of attributed nodes
(z1)
Propensity for attributed nodes to have more
partners
No effect for an actor being influenced by a
network partner need to introduce dependencies
among attribute variables
15Assume attribute variables are dependent if the
actors are tied partial conditional dependence
(Pattison Robins, 2002)
16Effects in the model
Baseline effect for number of attributed nodes
(z1)
Propensity for attributed nodes to have more
partners
Propensity for attributed nodes to be connected
17Why should attributes or ties be exogenous?
Friendship network for training squad in 12th
week of training (Pane, 2003) Green
detached Yellow team oriented Red positive
18Models for joint social selection/social influence
Xij
Zj
Zi
Xik
Zk
19Effects in the model
Quadratic effect in no. of attributed nodes
Baseline effect for no. of edges
Propensity for attributed nodes to have more
partners
Propensity for attributed nodes to be connected
Equivalent (blockmodel) parameterization
20Change statistics
Conditional log-odds for a tie to be observed
Conditional log-odds for an attribute to be
observed
21- 2. Which global network properties are important
? - Small worlds
- Short average geodesics
- High clustering
- Skewed degree distributions
- Regions of higher density among nodes
- cohesive subsets, community structures
22An example network (without attributes)
Confiding (trust) network (Pane, 2003)
23Observed networks Path lengths
Many observed networks have short average
geodesics small worlds The confiding network
has a median geodesic (G50) of 2 not extreme
compared to a distribution of Bernoulli
graphs The confiding network has a third
quartile geodesic (G75) of 2 also not extreme
compared to a distribution of Bernoulli graphs.
24Observed networks Clustering
Many observed networks have high clustering
small worlds
Global Clustering coefficient 3
(no. of triangles in graph) / (no. of 2-paths in
graph) 3T / S2
The confiding network has a global clustering
coefficient of 0.41 a comparable Bernoulli graph
sample has a mean clustering coefficient of 0.25
(sd0.03)
25Observed networks Clustering
Many observed networks have high clustering
small worlds
Local Clustering coefficient For
each node i, compute density among nodes adjacent
to i. Average across the entire graph.
The confiding network has a local clustering
coefficient of 0.58 a comparable Bernoulli graph
sample has a mean clustering coefficient of 0.25
(sd0.04)
26Observed networks Clustering
Many observed networks have high clustering
small worlds The confiding network has a global
clustering coefficient of 0.41 The confiding
network has a local clustering coefficient of 0.58
27Observed networks Degree distribution
Many observed networks have skewed degree
distributions as is the case for the confiding
network
28Observed networks Higher order
clustering k-triangles (Snijders, Pattison,
Robins Handcock, 2004)
1-triangle
2-triangle
3-triangle
Alternating k-triangles
Permits modeling of (semi) cohesive subsets of
nodes (cf community structures)
29Observed networks Higher order clustering
Observed networks often exhibit regions (subsets
of nodes) with higher density In which case, we
will see an alternating k-triangle statistic
higher than for Bernoulli graphs
The k-triangle statistic is not simply equivalent
to global clustering
30Summary Some global features not uncommon in
observed networks
- Short median geodesics (G50)
- Short third quartile geodesics (G75) perhaps?
- High clustering
- High k-triangle statistics
- Skewed degree distributions
- Bernoulli distributions tend to have short median
geodesics, low clustering and low k-triangles - Hence a basis for comparison
31- 3. Simulate the model to see whether global
properties can be reproduced
32Simulation of the model
- Use the Metropolis algorithm
- procedure similar to Robins, Pattison Woolcock
(in press) - Typically 300,000 iterations
- reject initial simulations for burnin
- Sample every 1000th graph
- Inspect degree distributions across sample
- Compare each graph in sample with a Bernoulli
graph distribution with same expected density - Hence can determine if graph
- - has short G50, G75
- - highly clustered high k-triangles
- Define highly clustered and short G50 as SW50
(small world) - Similarly define SW75
33Quadratic effect in no. of attributed nodes
Baseline effect for no. of edges
Propensity for attributed nodes to have more
partners
Propensity for attributed nodes to be connected
First simulation series 30 node graphs
34Change statistics
Conditional log-odds for a tie to be observed
- Expect density to be same among
- non-attributed nodes (zi zj 0)
- attributed nodes (zi zj 1)
35Numbers of edges and attributed nodes
36Assortative and dissasortative mixing
37Acceptance rates
38Clustering
39k- triangles
40Geodesics and clustering
41Small worlds
42Degree distributions
43Graph is SW50 (but not SW75) t-statistic for
k-triangles (relative to Bernoulli) 2.02
44The graph also has a skewed degree
distribution Although unusual for graphs in this
distribution
45Conclusions for this series of simulations
- The parameter estimates results in approximately
equal numbers of attributed and non-attributed
nodes - Density within the two sets of nodes are similar
and high. - As the attribute expansiveness (ß1) parameter
becomes more negative, and the attribute
connection (ß2) parameter more positive - acceptance rate for attributes decreases,
- clustering and community structure increases, 3rd
quartile geodesics decrease, but median geodesic
remain relatively short - Graphs with small world features, but not with
skewed degree distributions, are common within a
medium range of the attribute expansiveness
parameter.
46Quadratic effect in no. of attributed nodes
Baseline effect for no. of edges
Propensity for attributed nodes to have more
partners
Propensity for attributed nodes to be connected
Second simulation series 30 node graphs
47Numbers of edges and attributed nodes
48Assortative and dissasortative mixing
49Acceptance rates
50Clustering
51k- triangles
52Geodesics and clustering
53Small worlds
54Degree distributions
55Degree distributions
56Degree distributions
57Graph is SW50 (but not SW75) t-statistic for
k-triangles (relative to Bernoulli) 3.98
58Graph is SW50 (but not SW75) t-statistic for
k-triangles (relative to Bernoulli) 3.98 And
with skewed degree distribution
59(No Transcript)
60(No Transcript)
61Conclusions for the second series of simulations
- The parameter estimates result in a minority of
attributed nodes with high internal density, and
a majority of non-attributed nodes with lower
density. - As the attribute connection (ß2) parameter
increases, no of edges and attributes increase
somewhat, and acceptance rate for attributes
decreases, - clustering and community structure increases, 3rd
quartile and median geodesic become longer. - Degree distributions become skewed, and then
bimodal - Graphs with small world features, and with
skewed degree distributions, make up a sizeable
proportion of distributions with large attribute
similarity parameter.
62Some final comments
- This thought experiment demonstrates that
several important global features of social
networks may be emergent from attribute-based
processes of mutually interacting social
influence and social selection - Short average paths
- High clustering
- Small world properties
- Community structures
- Skewed degree distribution
- Moreover, the models do not presume fixed
attributes - although the structural properties begin to
emerge as attributes become sticky (changing
more slowly)
63Some final comments
- Network models typically assume homogeneity
across graphs. - This assumption may not be appropriate to the
actual processes that are generating the network. - One way that homogeneity may break down is
through attribute-based processes. - Other possibilities include social settings
geographic proximity - Network studies may require a careful
conceptualisation of process to ensure that
models are properly specified. - Because process is (usually) local, with global
implications, the possibility of node-level
effects should not be excluded.