***** To join INSNA, visit http://www.insna.org *****
I am revising a paper on imputation of skewed variables. This is not
a network paper, but an example using network data would be very appropriate.
In particular, I'm seeking data with (at least) two repeated measures
of popularity. A friendship network, say. At time one, each
respondent in a group lists his or her friends, perhaps using a list,
perhaps distinguishing among close friends, friends, and
acquaintances. The people who get named the most are the most popular
(highest indegree, perhaps with greater weight given to stronger
ties). At time 2, the process is repeated. The correlation between
time 1 and time 2 popularity scores is an index of stability.
What I like about this example is that popularity is (likely) very
skewed, and the correlation between waves 1 and 2 is (likely) very
strong. Also, there are likely to be missing values if (say) some
people leave or join the group between waves, or if people are
inadvertently left off the list of possible nominations. So it's
exactly what I need for my paper: high skew, high correlation, and
missing values that need imputing. (I realize there may be unique
issues in imputing network data, but those are really beyond the
scope of my paper.)
The Add Health data follows this format, but it's a bit of trouble to
access and transform into the necessary form. Since I just need a
simple example for a statistics paper, I'm seeking data sets that are
smaller and ready to go. If one of you has something to share, I'd be
grateful to hear about it.
Many thanks and best wishes --
Paul von Hippel
Paul von Hippel
Department of Sociology / Initiative in Population Research
Ohio State University
300 Bricker Hall
190 N. Oval Mall
Columbus OH 43210
Office hours TThF 3-5pm
I read email every weekday at 3.
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.