Don McBurney wrote:
"One not-so-minor point: The purpose of randomization is not to produce
groups that are approximately equal--matching can do a much better job. 1)
Randomization is the only method that guarantees that there is no confounding
except by chance, and 2) the probability that differences between groups occured
by chance can be assessed by inferential statistics (which assume the random
model)."
I agree but with the caveat that matching can only do a better job on the one or two
variables that are used for matching. In agreement with point #1, random assignment
guarantees no confounding on any variable (including those for which we have no
measurement and even including those we have never imagined) except by chance and, as
with point #2, the probability of chance being the explanation can be estimated with
inferential statistics. Matching only guarantees equivalence on the particular
variables being matched, not on all of the unimagined or unmeasured variables.
Rick
Dr. Rick Froman
Associate Professor of Psychology
John Brown University
Siloam Springs, AR 72761
(479) 524-7295
e-mail: [EMAIL PROTECTED]
web: http://www.jbu.edu/academics/sbs/rfroman.asp
---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]
