Print

Print


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

I am analyzing data at the partnership level and have a question
about how to handle correlation in my data.  I am uncertain as to
how my clustering variable should be labeled when using generalized
estimating equations.
My data look like this:
I have partnerships in my database, but individuals can appear in
more than one partnership. My data are configured as such (where
partnership 1,2 and 3 are correlated because of person A, but
partnership 3 is also correlated with partnership 4 by person D):
  partnership 1: person A and person B
  partnership 2: person A and person C
  partnership 3: person A and person D
  partnership 4: person D and person E
How do I construct a clustering variable for this scenario? Simple
SAS code is below-- the question is, if I were to read in four
lines of data corresponding to the four partnerships, what should
be the values for a, b, c, and d?
{Where Drug indicates whether or not the partners use drugs
together, AgeCat indicates if the partners are in the same age
bracket, and Serocon indicates if the partners have the same
serostatus}

  data dyads;
    input cluster Drug AgeCat Serocon;
    datalines;
  a 1 0 1
  b 0 1 1
  c 1 0 0
  d 1 1 1
  ;
  run;

  proc genmod data=dyads;
    class cluster Drug AgeCat;
    model  Serocon = Drug AgeCat / dist=bin type3;
    repeated  subject=cluster / type=exch covb corrw;
    estimate "Drug" Drug 1/exp;
  run;

Thanks for any help, Caroline Korves

_____________________________________________________________________
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.