I am a political scientist interested in applying social network theory
to the analysis of alliance formation in international relations. My
familiarity to social network analysis is mostly limited to a textbook
and some articles referenced in that. There are three issues in the
computation of structural equivalence measures that I have not been able
to find sufficient discussion on. I would appreciate any insights in
these questions, including literature references.

1) What is the appropriate way of dealing with different participation
rates? Have alternative solutions to standardizing/using correlations
been proposed? For example, has anyone proposed measures that use
information about the participation rate of actors to assess the
reliability of correlation coefficients?

2) Are there discussions on alternative ways to include indirect links
when relations are valued?

3) In the particular network that I am studying, not all potential
relations have the same “relevance,”  based on geographic and other
reasons that have little to do with the policy choices, which are my
primary theoretical interest. Thus the absence of an alliance may be the
result of a contentious relationship or plain indifference. Have methods
been proposed to take this into account, for example where relevance is
a function of a set of covariates at the dyad level, and works as a
"selection" mechanism?

Thanks,  Erik.

Erik Voeten
Post-Doctoral Fellow
Center for International Security and Cooperation
Stanford University
Stanford, CA 94305
phone : 415-982 6230 (home)
 650-724 6283 (office)