***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
Danielle M. Vogenbeck wrote:
> ***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
> I'm very new to social network analysis and in an attempt learn the method, we
> are performing some very basic research in my graduate department. We would
> like to gather data in two matrices, each measuring something different, and
> then output one network map of the entire data gathered. For example, in one
> matrix, we will measure intensity or depth of a tie, and in the second matrix,
> we will measure "instances of agreement" that the tie exists. For example, one
> person might think that there is a strong relationship between A and B, but no
> one else agrees (a low instance of agreement).
> Is it possible to gather these two measures and then use a program like UCInet
> to produce one network map (as opposed to two separate images)?
I may not understand the problem, but it seems to me that you may only
need one matrix. If, by measuring "instances of agreement", you mean
reciprocity (A names B and B names A), then aren't the instances of agreement
implicit in the first matrix (holding intensity of ties)? (by inspecting symmetrical
elements in the matrix?)
My second point of confusion is the use of the words "map" and "images".
If, by this, you mean some kind of visualization of the network... a picture...
Then one solution may be to use a visualization program like Pajek.
For instance, the lines could have thicknesses representing tie intensity
(and can even show this independently for both directions), and color
could be used to represent reciprocity (or agreement by others, if
I got your problem statemenet wrong).
I'll stop now, in case I've misinterpreted the problem -- but if your
network is large, and you know SAS, I do have a routine
to convert directly to Pajek files, and something could be whupped
up in short order to detect reciprocity/agreement - even if it
has to be stored in another matrix. If your problem is small,
the Pajek files are ASCII and can be manually edited.
I'd try a test network of three nodes for starters...
Others may know better ways to get Pajek or Netdraw to do your
bidding, especially as Steve Borgatti outdid himself to get Pajek,
Netdraw, Mage and UCINet all talking to each other.
My major point is that you may only need one matrix to represent
the problem, and may be able to get the thicknesses/colors to
represent what you want directly from this matrix, using transformations
within UCInet and/or Pajek. I couldn't easily figure out how, so I'd
program it in SAS and convert straight to Pajek... but that's me.
Stephen Q. Muth -- [log in to unmask]
voice -- (719) 471-7025 | FAX -- (719) 473-5474 | cell -- (719)210-3785
SOCNET is a service of INSNA, the professional association for social
network researchers (http://www.sfu.ca/~insna/). To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.