Compute the mean and median of the interaction scores and use one as the cutoff
(probably median), then to dichotomize pull down the Transform menu and select
Dichotomize using that value. - Tom V.
Carl Nordlund wrote:
> I am sorry for spamming this mailing lists with my newbie questions but I
> have no other place to resort to! Of what I know, I am the only one at our
> uni fiddling around with social networks so I have no other to ask!
> My question concerns finding cohesive subgroups (cliques) in valued
> undirectional data sets. I am currently experimenting with Freeman's EIES
> data concerning messages sent/received between the 32 actors and I would
> like to get a dendrogram or similar where I can see groups of people who
> interact more with eachother than other. (I have symmetrized the data so
> that it tells the total number of messages exchanges between dyads). As I
> understand it, I am looking for cliques/n-cliques but as the data isn't
> dichtonomous I can't get the cliques-tool to work for this data set. Do I
> first have to do a cutoff of the data set at a specific threshold value? How
> do I do that in UCINET?
> In the EIES dataset, values range from 852 to zero. What cutoff-value should
> I apply? Should I make several clique analyses with different cutoff-values?
> Plenty of thanks in advance!
> Carl Nordlund, BA, PhD student
> Human Ecology Division
To learn more about my evaluation book go to: My personal webpage:
Thomas W. Valente, PhD
Director, Master of Public Health Program
Department of Preventive Medicine
School of Medicine
University of Southern California
1000 Fremont Ave.
Building A Room 5133
Alhambra CA 91803
phone: (626) 457-6678
fax: (626) 457-6699
email: [log in to unmask]