Do the Math

[wcepler]

Do the Math



**********************************************************************

Mathematics has a subject called Hypothesis Testing which works like this. If something you claim is true, then a critical experiment would have an acceptably probable result and if you don't get that result then the claim is rejected. With a slight adjustment, this is basically the form of the scientific method (remember all those science fair projects for our kids?).



Something like, if you're in Albuquerque, then you're in New Mexico. And if you aren't in New Mexico, then you're clearly not in Albuquerque.



Thus, a fairly recent Ukrainian election was judged to be fraudulent because IF the election was fair, the tallied results would have had to be within margins of error of the exit polls. However, since the discrepancy was "statistically significant", the scientific consensus was that the election was fraudulent.



Ironically, EXACTLY the same thing happened in our country in the 2004 Kerry/Bush election! The exit polls unequivocally gave John Kerry the victory, but e-voting machine smoke and mirrors gave George Bush the election. Of course, one way to account for this near statistical impossibility is to explain away the exit polls. Another interpretation is country wide voter fraud.



But if we're going to swallow the whopper that the exit poll, international standard for measuring election fraud somehow "mysteriously" didn't work in just our country, in just that election, and in just those states without paper trails, then we shouldn't be surprised if we're asked to buy the Brooklyn Bridge again after November 7.

Better yet, let's keep the bridge!

************************************************************************

wcproteus

[Hav]

1. .

Well, I think the differences between many "official" results and the exit polls in 2004 were most likely not outside the margin of error.
Yet, and another DUer put some great work into it, it was strange that almost all official results compared to the exit polls turned out to be better for the Repubs, there was always this shift towards the Repubs. Sometimes a considerable shift. And this DUer calculated how really low the probability was for such a thing that this shift occured so often. And pretty much all these shifts appeared in states with electronic voting...

[wcepler]

2. thoughts

I agree to the "official" margin of error point.

The thing is, I think the exit poll business may the the straw that breaks the camels back this time. We already KNOW we should be winning certain key elections (from the polls , etc.) and the international exit poll standard may just be the match that lights a national fire of rage if these elections are stolen. I don't know what this means, but somehow I don't think the country will ALLOW four elections in a row to be stolen, beginning with year 2000 which Gore clearly won. Wishful thinking?

[Igel]

3. That's not all.

Not only do you need correlation, but you also need some sort of causal mechanism. You can get by in vastly complex systems with a correlation, since causality's a nightmare.

In the case of Ukraine, no causality needed to be deduced. Which isn't to say that there was no causality. Election fraud was observed in more than a few places, both by Orange supporters and by non-Ukrainian observers. Observation beats deduction every day (well, mostly every day ... witnesses can be fooled).

I was a frequent reader/non-participant at Maidan, and sometime volunteer translator for a Canadian prof who was assembling press releases/summaries and the like. It was where Yushchenko supporters coordinated and reported. Maidan's not quite structured like DU, and the politics don't map onto American politics very well; plus it was (still is, I assume) in Russ./Ukrainian. http://maidanua.org/ (I just looked at the site ... much less Russian there than before.)

[applegrove]

4. I for one think the exit polls were off. Just as the reluctant responder theory

said. Because exit polls were even off in places that had paper ballots. Will be interesting to see what happens next week.