***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
You seem to have a very interesting large dataset where a few observations are missing. You could model effects of those missing responses on the topology.
I agree with you that filling in with averages would not be a right approach (at the first glance atleast!) when you have evidence that they didn't communicate. Although you can perform some interesting simulations regarding the robustness of your outcomes (compare your results when you fill in averages or percieved numbers using information from other responses).
You may find this paper interesting :
----- Original Message -----
From: Anabel Quan <[log in to unmask]>
Date: Tuesday, May 6, 2003 8:40 am
Subject: Re: missing data
> ***** To join INSNA, visit http://www.sfu.ca/~insna/ *****
> Dear all,
> I am a phd student at the University of Toronto and I am currently
> working on my thesis
> and analyzing whole-network data.
> The problem that I encountered was not missing nodes, but missing
> responses. There were 15 missing responses in total from 5832 responses
> (8 items X26X26 network size). At this point, I could use one of the
> statistical methods (average, etc.) to replace the missing points. As I
> conducted a case study that included observations and I know the
> respondents and their interactions fairly well, I was wondering what
> approach to take. Would it make sense for me to substitute the missing
> values with an average (which is probably not correct) or with data
> points that I think would be appropriate for the relationship? This is
> especially tempting in cases where I know that two people do not talk to
> each other (which is also reflected in the responses of the other 7
> network items). I am really concerned because taking this route would
> mean that I am making my data up?
> Thanks for pointers to articles or any advice on this point,
> Anabel Quan Haase
> University of Toronto
> 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.
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.