Texas??? Hilary???

[joseftu]

Polls, again, and the very little significance they really have--all that granted.

But TEXAS (http://blogs.usatoday.com/onpolitics/2007/06/is_texas_turnin.html)???

McCain and Clinton essentially tied (36% McCain, 35% Clinton).

How will the Democrats find a way to lose this one? That's the big question. ;)

[Biker]

Highly doubt the poll was a wide section of the population and was conducted mainly in major cities. You could poll a thousand people here in Austin and Clinton would be ahead by a huge margin.

[joseftu]

The 2007 Texas Lyceum Poll is a random-digit dial telephone survey of Texas
adults. Telephone coverage within the state of Texas is approximately 97%.
Randomized selection procedures were assiduously followed throughout the
process, even at the level of selecting individuals within the household. The final
sample size is 1,002 adult Texans. The instrument itself relies on questions that
have been used previously in national polls, and have been shown to be both
valid (correlating with plausible independent and dependent variables) and
reliable (robust to question order and interviewer effects). A Spanish version of
the instrument was developed and respondents were given a choice of
participating in English or Spanish. Bilingual interviewers were utilized, and
approximately 60 interviews (6% of the sample) were completed in Spanish. The
overall response rate (completed interviews / contacts) is 40%. This rate is
partially the result of an extended time in the field, which facilitated call-backs.
The overall margin of error for the sample is +/- 3.0 percentage points at the 95%
confidence level. The data used to generate top-lines and tables are weighted by
U.S. Census Bureau estimates with respect to age, gender, and race. For example,
Census data indicate the proportion of 18-29 year old Hispanic females in Texas,
and we use these estimates to weight the survey data. As expected, the most
significant weights are applied to young, male, minority respondents (who are
under-represented here, as they are in almost all polls in the U.S.).
I would imagine that a random-digit telephone poll would represent a random selection of the population--including major cities and other areas. Maybe the majority of the voting population of Texas does live in major cities?

I don't know--I'm no statistician. I generally think that these polls are about as valuable as the astrology column in the local paper--and should be labeled "for entertainment only." That's about all they're good for.

It is entertaining to think about Hilary winning Texas, you know? :)

[Coot]

Entertaining indeed, of course they did produce Lyndon Johnson. :P