***** To join INSNA, visit http://www.insna.org ***** The first and most important point in this message is to offer grateful thanks to Ian McCulloch and Andrej Mrvar, whose advice has led to what is, in retrospect, an entirely predictable explanation of my problem. My bust.

The tale begins when I realized that in setting up the networks for my project I had set up the edge list in ads (rows=events)-creators (columns=actors) order, reversing the standard convention by which rows=actors and columns=events. Thinking to minimize confusion, I decided to recreate the networks in creator-ad order. This, in itself, is an almost trivial task. Generating a new network by exporting data from Filemaker and using Text2Pajek to convert the edge list into  Pajek.net file takes only minutes. But as the proverb says, "Haste is the devil's work." I failed to check to be sure that the vertices in the edge list were sorted in the proper order to match my attribute partitions, sailing blindly ahead and creating six networks, all with the same flaw. 

Thanks, however, to Ian and Andrej, I have not only found the mistake, I have also learned some useful new things about debugging a problem like this and discovered a way to extract more information from my data.

As my Japanese colleagues would say, Mendo wo kakechatta, sumimasen (Sorry for making all this fuss), arigato gozaimasu (Thank you very much).


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 (http://www.insna.org). To unsubscribe, send an email message to [log in to unmask] containing the line UNSUBSCRIBE SOCNET in the body of the message.