*****  To join INSNA, visit  *****

Greetings SOCNET members,

I am a PhD student working with information ego-networks collected within a
place-based community.  I was hoping to elicit some suggestions for how to
go about using p*/EGRM techniques to analyze these ego-networks.  The data
we have derived is only partially observed in that we have ego's ties to
alters and alter's ties to alters, but only withn the boundaries of each
individual ego-network.  We do not have any information regarding the ties
across ego-networks. Also, in that this data was collected in a place-based
community, we do not have clear boundaries for what constitutes the complete
network (i.e., we don't have a master list of all the residents), which
means we do not know precisely what tie we do not have data for, which makes
some of Handcock and Giles work on adjusting for known missing data less
applicable to our situation (I think).

As of now I'm thinking one way to apply ERGM techniques to this dataset is
to simply treat all 218 of our ego-networks as individual complete networks
and analyze them as such.  Then do some kind of meta-analysis across the 218
ego-networks to glean some patterns and trends.  Another option might be to
put all 218 ego-networks into a single matrix, comprised of all our
respondents and all the people sources they named, and control for the
mutual exclusiveness between each ego-network  by giving "0" values to all
the potential relationships outside the confines of each ego-network.  The
problem with this is that in this tight community we have reason to believe
that mutually exclusivity cannot be assumed in that some alters were
coincidentally named by more than one ego.  But this type of data was not
systematically built into our survey design.

Any feedback would be welcome.  I look forward to your responses.

Lindsay Young
Doctoral Student
Northwestern University
School of Communication
Media, Technology and Society (MTS)

SOCNET is a service of INSNA, the professional association for social
network researchers ( To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.