EGONET BROKERAGE MEASURES

1)      Ego-centered versus socio-centered research

a)      What it means

b)      Ego-alter terminology

c)      Data collection

i)        Name generator

ii)       Name interpreter

iii)     Alter attributes

iv)     Alter ties

d)      Kinds of analyses done with ego network research

i)        Selection (e.g., homophily) & diffusion

ii)       Heterogeneity

iii)     Quality

iv)     Shape

2)      Structural holes

a)      Concept: lack of ties among ego’s alters

b)      Benefits according to Burt:

i)        Autonomy

(1)   Example: the guy at the bar

ii)       Control

(1)   Example: Carter presidency

iii)     Information

(1)   Example: Burt’s managers

c)      Measures:

i)        Non-redundancy: 1 minus density of the Egonet (ignoring ego), at least for non-valued, undirected data

ii)       Constraint: extent to which ego’s direct and indirect ties lead to same alters

iii)     Example: symcampnet and campnet datasets

3)      Approaches to social capital

a)      Burt’s structural holes is purely topological

b)      Coleman’s approach of closed networks is the opposite – and also purely topological

c)      Nan Lin’s approach of social resources – who you know, not just how you know, what kind of people

d)      Fundamental divide in mechanisms:  topological versus flows

e)      Next: an approach of combining topology with node attributes

4)      Gould & Fernandez

a)      Alters are categorized by node attribute

b)      Egos with same level of brokerage can play different roles

i)        Coordinator, representative, gatekeeper, consultant, liaison

c)      Information benefits increase as you move from left to right

d)      Consider connection between brokering between groups and personal innovativeness

i)        Understanding how to work with other groups better – understanding their language

ii)       Transferring ideas directly

iii)     Seeing the value of their problems & solutions for another group: analogizing

iv)     Synthesizing

e)      Something to be careful of: ties are not the same as flows. Structure provides potential for behavior, but the agent makes choice of whether to execute the behavior

i)        So the coordinator role may not in fact pass on information

f)        Roads and traffic. Relations and interactions.

g)      Example: campnet and campattr2 col 4

i)        Correspondence analysis

ii)       Opening brokerage file as 2-mode dataset in netdraw

5)      E-I Index

a)      What is it?

i)        Counting the number of ties to outsiders (External ties) relative to the number of ties to insiders (Internal ties)

b)      What for?

i)        Measure of global cohesion. Ties across boundaries.

c)      E-I index is (E-I)/(E+I)

i)        E = between group ties

ii)       I = within group ties

d)      Research on the E-I index (Krackhardt and Stern)

e)      Run example using campnet and campattr2 datasets

f)        Caveats

i)        EI index can run between -1 and +1. But certain distributions of group sizes can make either end impossible to attain. So we can rescale between the possible minimum and maximum points.

ii)       The expected value if people choose partners without regard for group membership is not necessarily 0

6)      Density tables

a)      The density of ties within and between classes of nodes

b)      Patterns of global cohesion