***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
Hi. I'm a graduate student in clinical psychology, working on a study of
self- and peer-perceptions of dysfunctional personality. My sample consists
of 2000 military recruits, divided into training groups of about 40 or 50.
We also have self-report ratings of how well each participant knows each
other participant in his or her training group. I'm trying to find social
network characteristics that moderate peer-peer consensus (how two peers
rate the personality of a third actor) and/or self-peer correspondence of
I am particularly interested in finding sub-groups within the networks, on
the hypothesis that correspondence of personality judgments within the
sub-group will be higher than that across sub-groups. I don't have any
formal training in social network analysis, but have been reading as much
of the literature as possible. Now, however, I've reached a bit of an
impasse, and hoped to get some feedback from this group.
Based on what I've read, I think I understand the differences between the
different types of social cohesion measures (cliques, k-plexes, lambda
sets, etc.). However, I'm not sure how to choose a particular one to use.
Does my question of interest (finding groups that improve correspondence
among raters) lend itself more to one or another of the techniques?
A related question: I am using UCINet V, and am interested in the Factions
procedure of that program, which uses a Tabu algorithm to partition the
network. However, this procedure forces all actors into one of the groups,
whereas the social cohesion measures do not require every actor to have a
group membership. Are there any benefits or limitations that I should be
aware of when forcing group membership? Is this sort of procedure used in
conjunction with or instead of social cohesion techniques?
I'd appreciate any suggestions or advice that you may have, either via the
list or backchannel.
Thanks in advance,
Allan Clifton, M.A.
Department of Psychology
University of Virginia
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.