Hello, 

I have a question about which estimator is most appropriate for my dependent 
variable. I have time-series cross-sectional data from 1980-2000 for countries' 
abortion policies. The data are ordinal (5 point scale from prohibit abortion in 
all cases, allow it to save the life of the mother, permit it on medical
grounds, etc.) but could be collapsed into a binary measure (on demand/not on 
demand).

Ideally I would like to examine what elicits change in state policies between
1980 and 2000. I thought that the easiest way to do this may be to measure the
change between 1980 and 2000. Since these policies are pretty static, I have a
few countries who change and a lot who remain the same. When I simply take
the difference between 1980 and 2000, I am left with a distribution that looks
like a reverse grade distribution.....a few Fs, a lot of Ds, some Cs and a few Bs
and As.  

I would prefer not to collapse it into a binary dv because of the information that 
is lost.  First, the change can occur in both directions (some countries get more 
restrictive while most countries get less restrictive) and second, not all 
differences are the same as simply calculating differences obscures the 
difference between changes from 2 to 3 and 4 to 5).

I would greatly appreciate suggestions about which estimator to use.

Feryal Cherif
Department of Political Science
University of California, Riverside
2230 Watkins Hall
Riverside, CA 92521
(951) 827-5509

**********************************************************
             Political Methodology E-Mail List
   Editors: Melanie Goodrich, <[log in to unmask]>
            Delia Bailey, <[log in to unmask]>
**********************************************************
        Send messages to [log in to unmask]
  To join the list, cancel your subscription, or modify
           your subscription settings visit:

          http://polmeth.wustl.edu/polmeth.php

**********************************************************