Hi everyone,
I am doing a meso-level network analysis and trying to identify the subgroups in the network.?After reading the chapter 11 of Analyzing Social Networks (second edition 2018), I want to combine the Clique analysis and HiClu so as to remove the overlap. Here is my procedure: (1) go to Network|Subgroups|Cliques to get the Actor-by-Actor Clique Co-Membership Matrix; (2) Given that the output of Clique analysis- HIERARCHICAL CLUSTERING OF OVERLAP MATRIX doesn¡¯t report the Q value (see the zache example in the word attachment), I run the HiClu analysis (Tools|Cluster Analysis|Hierarchical) based on the Co-Membership Matrix. And the participation that maximizes the Q value is used to identify the subgroups in my study. Am I right?
Besides, I noticed that the hierarchical clustering results of these two methods are different (please see the two figures in the word attachment). Does anyone have a clue about the differences?
Any information would be appreciated. Thank you!
Best,
Longxia Huo
Ph.D. Student??
?
