*****  To join INSNA, visit  *****

Hello everyone,


I am a PhD Candidate working on my thesis in the Faculty of Information &
Media Studies at the University of Western Ontario who has lurked on this
list for a while now. Now I have a question I hope you can help me with.


I am working with social network data for a project focused on networks for
information and help related to HIV/AIDS in rural Canada. We have data for
three regions of Canada. Network participants include people living with
HIV/AIDS, their friends/family members, their health care providers and
service providers. We sampled this network using recruitment through health
and social service agency partners, followed by snowball methods based on
nominations. In so doing, we relied on a sampling approach set out by
Goodman (1961).


We conducted semi-structured interviews with participants in each of these
categories in order to elicit information about their networks and how these
networks work. We chose to conduct interviews only with people who lived in
our rural regions of interest, or who were located in the closest urban
centres but worked with a number of our participants from these rural
regions. However, we documented out-of-region nodes without pursuing
interviews with them. From this, a list of networks for HIV/AIDS-related
information/help/support were created for each participant, and we have
worked to construct whole network data from these interviews by identifying
interconnections between participants based on these multiple interviews.


A problem has arisen, though, in that we initially hoped to be able to speak
with all key participants in these networks, but fewer people participated
than we'd hoped - and in one area, we are missing some key players mentioned
by multiple participants as members of their networks. We have 117
participants from the three regions who participated in semi-structured
interviews, but this represents a small proportion of the people mentioned
in interviews. 


Can anyone suggest an approach to dealing with this missing data so I can
extrapolate/model the whole network? I very much want to use whole networks
as my unit of analysis and whole network metrics (such as density and
fragmentation) to compare these networks as I think this is a particularly
rich area of my data. In the Wasserman & Faust text, I have found citations
to Frank (1971)'s work on this problem, but I'm wondering if any more recent
approaches for this have been developed and whether any SNA software
programs (such as UCINET, which I am currently using) can accommodate these


Thank you very much for your help in advance.




Tiffany Veinot, BA, MLS

PhD Candidate

Rural HIV/AIDS Information Networks Study Project Coordinator 

Faculty of Information and Media Studies

The University of Western Ontario

E-mail: [log in to unmask]

Voice: 416-698-7743

Visit my Academic Profile at:

SOCNET is a service of INSNA, the professional association for social
network researchers ( To unsubscribe, send
an email message to [log in to unmask] containing the line
UNSUBSCRIBE SOCNET in the body of the message.