From: "Robert G Lewis" 


"Richard B."  wrote in message
news:kmtho017kedob376a3ic3p5ftdosef23jp{at}4ax.com...
> On Wed, 03 Nov 2004 09:00:01 -0500, Mike N. 
> wrote:
>
>>How could they be so far off?
>
> I think it was the bad exit polling data, which had Kerry up in many
> places.
>
> - Richard

http://www.washingtonmonthly.com/archives/individual/2004_11/005075.php

XIT POLLS....The CNN guys were blathering on a few minutes ago about how
the exit polls were screwy once again this year, and wondering whether they
could ever be trusted again. And it's true that some of the early exit
polls were off by several percentage points.

But those were early exit polls, and everyone knows they have a fairly high
margin of error. So I just made a cursory check of the final exit polls and
compared them to the final vote - and guess what? They were pretty close:
within one point for the national vote (which had an MOE of about 1%) and
within a couple of points for the state polls (which had an MOE of 2-4%
depending on the state). In other words, they were all within the
statistical margin of error.

So what's the problem? Sampling error is irreducible, so there's no way the
exit polls could have been any better than they were. What's all the
griping about?



I think the early results were wrong but I'd glad that the final exit polls
were close. If there were a large difference there would be far more
questions about the vote.

Bob Lewis

--- BBBS/NT v4.01 Flag-5