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

Hi all,
When sketching on a small SNA-style research project concerned with mail
correspondence on a small, local web-community (aka networking site), a
question/dilemma has arisen.

Background: the project intends to to a role-structural analysis of the
internal mail correspondence (valued directional data) on a
web-community that started in July last year. We have full access to all
such correspondence, both technically and legally/ethically, having data
on sender, receiver, and date of 20000+ internal mails sent so far.

However, as the web-community has grown from zero to 1050 users since
its launch last summer, and as it is constantly growing (approx 50 each
month at the moment), thi non-constant number-of-actors is problematic.
Creating a 1050x1050-matrix and using correspondence data since its
genesis would assume, wrongly, that all 1050 current members have been
members from the start. Possible alternatives, as I see them, would be
to only use data for a fairly narrow time-period, let's say from
January-March this year, either using the actors that were members as of
January, or as of March. As we will primarily look at the
role-structure, a third alternative would be to look at more
"time-slices", shorter periods, for instance how the number of role-sets
and their inter- and intra-positional ties looked for each month.

How have others solved this problem, in theory and practice, with an
expanding set of actors? All suggestions are much welcome!

Yours,
Carl

--
Carl Nordlund, BA, PhD student
carl.nordlund(at)humecol.lu.se
Human Ecology Division, Lund university
www.humecol.lu.se

_____________________________________________________________________
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.