Robert Dawson writes:
>It's an interesting data set, an object lesson in design of
>questionnaires, and the whole episode has been genuinely educational.
Amen to that. I used a web site on the ballot design:
http://fury.com/galleries/palmbeach/index.php
<http://fury.com/galleries/palmbeach/index.php>
to discuss the issue of questionnaire design in a class I taught yesterday.
An important issue is that no one ran a pilot on this ballot. So even though
the ballot may look reasonable at first glance, even to a democratic
commissioner, there were problems that manifested themselves on the election
day. How much better would it have been to get those problems out in the
open before election day. It would have only taken a few dozen people in the
pilot to identify problems with that ballot.
Also interesting are the comments on usability testing by Dan Bricklin that
someone was nice enough to post earlier:
http://www.bricklin.com/log/ballotusability.htm
<http://www.bricklin.com/log/ballotusability.htm>
A fascinating comment is "there is research that says shorter people don't
vote on ballot questions as frequently as taller people because the
questions are displayed way above their eye level on voting machines."
The various regression analyses suggested by Greg Adams:
http://madison.hss.cmu.edu/ <http://madison.hss.cmu.edu/>
and John Marden
http://cuwu.editthispage.com/2000/11/08
<http://cuwu.editthispage.com/2000/11/08>
are also wonderful teaching examples (thanks again to the folks on EDSTAT-L
who posted these sites). I never had a good way of explaining the
interpretation and use of predicted values and residuals in a regression
model until this county by county polling data set came along. I used the
Bush-Buchanan regression model in a class I taught last Friday, and I had
the group (mostly doctors) fascinated with Statistics.
Steve Simon, [EMAIL PROTECTED], Standard Disclaimer.
STATS: STeve's Attempt to Teach Statistics. http://www.cmh.edu/stats
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
