Hi, everyone. Thanks to Matthew Landauer for pointing out the following
misprint in the problem set, problem 7. The set describes Segal-Cover
(SC) scores as ranging from -1 to 1, as they in fact do in the original SC
article. However, Epstein and Mershon (EM) recode the original SC scores
to range from 0 to 1. We are using the EM coding, so the scores range
from 0 to 1 for the problem set. Matt is right that this coding
difference is substantively important in interpreting your results as
well; see his suggestion below. The note on page 269 of the EM article
explains how EM recoded the SC scores.
Thanks again, Matt, and apologies to everyone for the error.
Ryan
------------------------------------------
Ryan T. Moore ~ Government & Social Policy
Ph.D. Candidate ~ Harvard University
Homepage:
http://www.people.fas.harvard.edu/~rtmoore/
Gov1000:
http://www.courses.fas.harvard.edu/~gov1000/
On Sat, 30 Oct 2004, Matthew Walter Landauer wrote:
Hi Ryan,
it seems to me that there might be a mistake in the text of question 7.
It says that S-C scores range from -1 (perfectly conservative) to 1
(perfectly liberal), with 0 being moderate. However, all the actual S-C scores we
have are greater than or equal to zero -- even Scalia gets a 0.00. I am
guessing that the real scale then is 0 to 1, with 0 being a perfect
conservative.
This becomes important when interpreting the plot -- does the slope
coefficient equal the difference in likelihood of voting liberally for a
moderate and a liberal, or the difference for conservatives and liberals?
thanks,
matt