Joao, I would suggest looking at the work of Warren Sack (UC Berekely) and Marc Smith (Microsoft), two scholars doing work with massive Usenet data sets. david silver http://faculty.washington.edu/dsilver/ On Mon, 28 Jan 2002, Joao Vieira da Cunha wrote:

A few people asked me to share the answers to my question on newsgroup analysis, so here they go (although these are more questions than answers):

1. Several people shared links to Network analysis software, I have used this software for email but I still have to find an useful way to use it on newsgroups because most of the postings are to the group as a whole.

2. Off-the-shelf qualitative data analysis: great but all of the packages from Sage can't work with more than 500 records (unless you are willing to wait 20 min to rename a document). I have 3000 (which is small as far as newsgroups go)

Any suggestions

Joao

_______________________________________________ Air-l mailing list Air-l@aoir.org http://www.aoir.org/mailman/listinfo/air-l

On Monday, January 28, 2002, at 11:40 AM, Joao Vieira da Cunha wrote:

A few people asked me to share the answers to my question on newsgroup analysis, so here they go (although these are more questions than answers):

1. Several people shared links to Network analysis software, I have used this software for email but I still have to find an useful way to use it on newsgroups because most of the postings are to the group as a whole.

You can construct social networks from newsgroup threads by treating a reply by person B to person A's message as a message from B to A. Although it is true that such a message is posted to the group and not sent directly or exclusively to person A, this concept may suffice depending on your goals. Also, it doesn't make as much sense to generate reply-based social networks if few messages get replies or threads tend to be sparsely developed. Andrew