< re: "illiteracy" of Gore voters based on 10-item vocabulary test >
On 17 Nov 2000 06:50:05 -0800, [EMAIL PROTECTED] (William B. Ware)
wrote:
> Should we not be concerned with some measurement issues before we debate
> the evidence? What were the items on the 10-item test? That is, everyone
> seems to be jumping the gun... doesn't anyone care about validity anymore?
>
And then I suppose there is the matter of the 1800+ persons who were
supposed to adequately represent 7 categories of political
"belongingness"... which do not display an obvious basis of ordinality
of the scaling or samples....
Do we have a-priori types which have been quota- sampled, or
do we have a Procrustean fit of that many individuals ("whatever is
nearest)?
- What's the basis of being "strong" Republi-crat? There's not a
questionnaire, is there? Years of consecutive votes?
Do you think the sampling did a proper job of "representing" (to
borrow a recent Subject ) the proportions that would be present by
age, race, gender, geography? Of course, it is a separate question
of who might be "brighter" after you statistically controlled for
those natural variations.
- One of the strongest correlates (or determinants?) of the Gore/Bush
vote is the urban/rural split. I have long assumed that there are
several intellectually-related things (like good newspapers and
informed readers and attitudes) that result in Democrats winning
the majority of the big-city voters. That seems to me to be mainly
another joke, but it is one where the data are more robust than for
Herman's hypothesis.
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
