Print

Print


*****  To join INSNA, visit http://www.insna.org  *****

A colleague replies to my previous post.

> 
> I think what you're talking about is called ethnography.  For a number of years, I've been stressing its importance in network analysis, usually to deaf ears.  Let me give a couple of examples:  In teaching about networks, I place I place 13 nodes seemingly at random on the screen, and the connections among them, and ask people what this is.  No one gets it.  Then I draw in the baseball diamond.  The message is that you can know all about attachments (edges) but have to talk to people to find out what they are doing.  A more meaningful example from the literature was the observation that injection drug users who cleaned their syringes and didn't share needles were still transmitting HIV.  At 3am one night, an ethnographer hanging out with some IDUs saw that they were "backfilling," sharing the cleaning water.  These are, of course, "anecdata," but I think they illustrate what you're getting out.  It has always seemed to me that formal questionnaires only confirm what you knew enough to ask about, and don't tell you anything new.  Talk to the molecules.
> 
> 
I reply.

Thanks for getting back to me. But I don't think the answer is ethnography. I have the ethnography in spades. I am working with data from an industry that I have worked in and around for more than three decades, with an active trade press that has been publishing hundreds of pages of new material month after month for more than a half century, and because I once worked for one of Japan's two largest agencies, I have connections that have secured interviews with the most central figures in my networks. I am drowning in ethnographic and historical data. The problem for me just now is the gap between the social physics, the random graph models that predict very neatly the main features of the networks I am studying, giant components, skewed (possibly power law) distributions in centrality measures, that sort of thing, and the qualitative material. Community-detection algorithms generate all sorts of results with no clear relationship to known key features of the industry, an oligopoly dominated by two large agencies, and a media market transformed first by TV and now by the Internet,  a transformation that has altered the distribution of different types of teams and the different skill sets required to produce print advertising, TV commercials, and now Web-based interactive campaigns. One approach is to accept a split between theory and reality like that A.N. Whitehead labels "the bifurcation of the world" in Science and the Modern World, the split between the "primary qualities" (the stuff that Newtown's mechanics could handle) and the "secondary qualities" (everything else, which is most of experienced reality). The other is to ask what sorts of questions a better theory would have to answer and see if someone can come up with one. It's a crazy idea; but just imagine.

Additional thoughts and suggestions would be most welcome.

John
> 

_____________________________________________________________________
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.