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

Using Pajek to dissect my 2-mode network data on Japanese advertising
creatives and the teams that produce winning ads, I make a
 now-conventional move—projecting two 1-mode networks. From each of the
1-mode networks, I extract the giant component. Having extracted the giant
component I have Pajek compute the normalized degree, closeness,
betweenness and eigenvalue centrality (hubs and authorities) vectors. I
then use the Vectors>Info command to examine the Spearman rank correlations
between the four vectors. I discover a curious phenomenon. In the case of
the 1-mode network of individual creators, the highest correlation (.7~.8)
is between normalized degree and betweenness. In the case of the 1-mode
network of teams, however, the highest correlation (again in the .7~.8
range) is between normalized degree and closeness.

Has anyone else encountered this phenomenon? Is this a common and/or
mathematically predictable result? If so, please tell me how it works or
where to look for an explanation.


John McCreery
The Word Works, Ltd., Yokohama, JAPAN
Tel. +81-45-314-9324
[log in to unmask]

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.