Re: graphic of exit polls vs results in DRE vs paper states

[Ralph Shnelvar <ralphs@xxxxxxxxx>]

Dear Paul:

On Sat, 6 Nov 2004 00:46:27 -0700, you wrote:

>Frightening, but not very scientific. One might want to know what exit polls
>have looked like in the past in comparison to actual votes. I personally
>would want to look at exit poll comparisons before 2000.

Sorry, it is "scientific".  It may not be accurate or a representation of
the Truth, but it is definitely scientific.

The statistics presented are called "cross sectional". You're looking for
"longitudinal data".  Both are subject to abuse.

I understand your very valid point.  You just used the wrong word.



To follow up on Paul's point:

The statistics presented have some structural problems.  I will list a few
of them.

(1) Who took the polling data?

(2) What were the sample sizes of the polling data?

(3) Where was the polling data extracted?  You could easily get these
results if polling was done in heavily Democratic precincts in the paperless
states.

(4) Were there other irregularities being reported?

(5) There seems to be an average of about a 5% spread (advantage to Bush)
between the polling data and the official results.  For small sample sizes
this is well within the margin of error.  A simple Kolmogorov-Smirnov test
would shed some light on this.

(6) Do Republicans in paperless states respond differently to invasions of
privacy than Republicans in paper states?

Ralph Shnelvar


>
>-----Original Message-----
>From: Evan Daniel Ravitz [mailto:evan@xxxxxxxx]
>Sent: Friday, November 05, 2004 9:02 PM
>To: cvv-discuss@xxxxxxxxxxxxxxxxx
>Subject: graphic of exit polls vs results in DRE vs paper states
>
>
>http://www.whatreallyhappened.com/IMAGES/exit_poll.gif
>