THE PROBABILITY IS ZERO: THANK YOU, BILL GATES, FOR GIVING US EXCEL.

I did some minor editing of the text for display on my site, but the table is a pain. Do you have it in tab-delimited format? PM me.

I was going to parse it for you, but when I copied it and pasted it to notepad and then opened that in OpenOffice spreadsheet, it appears to be tab delimited already.

So cut and paste to notepad and save and you'll have it tab-delimited.

Because he switched to a monospace font and used spaces. Thanks for thinking of me, though. I opened the text in BBEdit and did some quick search and replace, and it works fine now!

do not heve the power to do anything about it.

1. Historically, exit polls have been much more accurate than
standard polls."

Just the opposite is true. Almost every exit poll of the last 20 years has been far more inaccuate than the pre-election polls. The reason for that is probably the non-random nature of the interception of voters.

As a matter of fact the pre-election polls were very accurate this year.

His curt, know-it-all manner, seemingly so self-assured, masks profound ignorance and/or cognitive dissonance on the subject.

Fact. Since 1952, TV networks have used exit polls to call elections.
Fact. They were 99.999% accurate in their calls.
Fact. Exit polls intially called Florida for Gore in 2000.
Fact. They were correct. They just did not realize that 175,000 punched cards would be spoiled, the vast majority in minority democratic precincts.
Fact. Exit polls are used all over the world to check for fraud. They are accurate to within 1%.
Fact. You can do random sampling without a telephone.

Questions for you:
What constitutes a random sample?
Do you know how to calculate the Standard Deviation?
Do you know how to calculate the Margin of Error?
Are those polled ever unsure who they voted for?
Can you document a single case where one has lied about who they voted for in an exit poll?

and the poster makes a point- exit polling isn't randomized and there are many ways for bias to enter the sample.

That doeasn't necessarily render them all innaccurate (and there's plenty of room for bias in pre-election polling, too).

It's just that with bias- the statistical measures are rendered less meaningful- and the probabilities that result are of suspect validity.

That, unfortunately, has been a problem with your analyses (which I have appreciated, btw) all along.