Ohio Does Its Magic: Hillary’s 3% Exit Poll Margin Morphs to 10% in the Vote ( TIA ) - x

[tiptoe]

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=4954291&mesg_id=4954291

[southern_belle]

1. kick

[mod mom]

2. Brunner is slowly cleaning the dirt between the cracks.

[EFerrari]

3. Is there something funny in Ohio?

It seems that way to me but I always start from intuition and work back to facts.

[selador]

4. how many data points

do we need to recognize that exit polls are inherently unreliable.

im not saying there was or wasn't chicanery. i am just saying that anomalies between exit polls and vote results suggest in general that exit polling is a poor methodology for collecting accurate results as to how people actually voted.

there are numerous theories on this, but it just seems pretty clear that exit polling shows mass suckitude

[OnTheOtherHand]

5. "mass suckitude"

That's pretty much right, but around here some take exit polls as an article of faith. It makes for some pretty weird 'conversations.'

If someone has real evidence about Ohio, that would be good to know.

[Bonn1997]

6. Do you have evidence that exit polls were unreliable before

electronic machines with no paper trail were used?

[OnTheOtherHand]

7. will 1988 suffice?

That year, in the CBS/New York Times exit poll, the average weight on white Bush voters was about 11% bigger than the average weight on white Dukakis voters, which implies about a 5-point error on the margin. The model error may have been smaller, or larger. In the raw data, there were more Dukakis supporters than Bush supporters -- but the researchers may have deliberately oversampled African-American precincts, so I limited the comparison to whites.

In 1984, there was about a 7% swing on weights (among whites) in the other direction. In 1980 there was about a 3% swing overstating Carter's vote share -- that was pretty good, but probably still biased. Carter voters actually outnumbered Reagan voters by about 2.4 points in the whole sample, although again, I'm guessing that the researchers deliberately oversampled African-American precincts.

I think the error in 1992 is well known by now.

The question isn't whether these were "before electronic machines with no paper trail were used," since such machines have never been used everywhere in the nation. But I daresay that none of these elections was decided on DREs.

[Bonn1997]

8. So huge exit-actual poll discrepancies happened

about once a decade previously and they happen about once a week now?

[OnTheOtherHand]

11. huh?

You've presented zero evidence that U.S. exit polls were ever reliable. I've presented evidence that the general-election exit polls were off in 1984, 1988, and 1992 -- which, if I'm not mistaken, is more than "about once a decade." Just how far are you going to move the goalposts?

If there are primaries every week, that gives the exit polls the opportunity to be wrong "about once a week." That isn't usually the case -- and the primaries often aren't as competitive for as long as these have been. Moreover, the phenomenon of having exit poll tabulations posted on the Internet when the polls close is pretty new, so we wouldn't necessarily notice whether the primary exits were accurate or not.

The Ohio exits could have been right, or they could have been wrong. But there's no point in starting from the pretense that exit polls are known to be accurate, now or before DREs were widely used.

[Bonn1997]

12. The difference is

that in those seem to be just random errors, which will happen occasionally (but rarely) if you look at a large sample of elections. Now, what we see is one party (the Republican party) and one candidate in each party (McCain and Clinton) systematically (or consistently) benefiting in the actual vote count relative to the pre-polling and exit polling. A basic statistical concept (I teach stats and research methods at a liberal arts college) is that an unreliable measure won't systematically relate to any outcome. If a measure is simply unreliable, it should not consistently favor one candidate or one party. In other words, we'd see exit polls giving an advantage to Obama just as often as they do to Hillary and to the Democratic candidates just as often as they do to the Republican candidates. But that obviously isn't the case. Instead, there has to be some other process that explains this systematic favoring. My view is that it's fraudulent vote counting. What's your view as to what variables explain the reliable advantage the Republican party in many recent elections and Hillary and McCain in these primaries have had?

[OnTheOtherHand]

13. if you teach stats, why not look at the data?

On what basis do you venture that "those seem to be just random errors"? And what "large sample of elections" have you considered? Would you like to provide a baseline of elections in which you consider U.S. presidential exit polls to have been accurate?

As you know, it's perfectly possible for a measure to be both unreliable and biased.

If your "view" is that fraudulent vote counting accounts for why the exits showed Obama close in Massachusetts, Arizona, and New Jersey when neither the vote counts nor pre-election polls did, you might want to consider whether you can bring any supporting evidence to bear.

Any number of things could explain exit poll bias, and I hardly have the means to disambiguate them.

[Bonn1997]

14. Not really

"As you know, it's perfectly possible for a measure to be both unreliable and biased."
That's not correct. The central feature of reliability is consistency of measurement, and bias involves some consistent (and inaccurate or invalid) favoring. Any measure would have to be reliable (but not valid) to have a bias. Maybe you mean validity when you're talking about reliability?

[OnTheOtherHand]

16. no, I mean reliability

Reliability is a function of the variability in the measurement. But a measure can be highly variable and still be biased.

Now, I suppose we could debate whether I've mischaracterized the reliability of the polls. We're told that in the 1988 (CBS/NYT) through 2004 general elections, the mean Within Precinct Errors were -2.2, -5.0, -2.2, -1.8, and -6.5. I'd say that looks unreliable and biased. YMMV. Bear in mind, however, that given the nature of U.S. elections, each of those is averaged across dozens of state elections.

[Bonn1997]

15. Looking at the data

As far as the looking at the data is concerned, I would say this book is the most comprehensive, compelling examination of the data
http://www.amazon.com/Was-2004-Presidential-Election-Stolen/dp/1583226877/ref=sr_1_1?ie=UTF8&s=books&qid=1205200316&sr=8-1

[OnTheOtherHand]

17. but Freeman didn't look at the data

I see no evidence in the book that he made any use of the publicly available datasets. He seems to have done very little that wasn't derived from the Edison/Mitofsky evaluation report. It's a very thinly argued and reasoned book, all in all.

Did you notice how weak his argument for exit poll accuracy was? Not bad if this were Germany....

[tiptoe]

19. You should look at this analysis which confirms Freeman:

Election Fraud Polling Analysis: Confirmation of a Kerry Landslide

[OnTheOtherHand]

20. what, are you joking?

It's not clear that you even read this stuff. It damn well doesn't confirm Freeman. For one thing, still no one has provided evidence that the U.S. exit polls ever consistently matched election returns, nor that discrepancies are correlated with voting equipment, or with any other indicator of vote miscount. There is no serious response to this. Just a lot of hand-waving.

You're entitled to your own reality, but nothing says that the rest of us have to live in it.

[Tarheel_Dem]

9. k&r

:kick:

[WillYourVoteBCounted]

10. huge amount of republican cross over, huge n/t

[kster]

18. All guns are draw on each other, take it to em TIA

DON'T GIVE UP, we could put our guns away, but, we have four crooks preventing us from doing that, 1) The Government, who think they can prevent us from Hand Counting ALL the Ballots, 2) The vendors who think they can prevent us from viewing the inner workings of their secret vote counting machines, 3) Mitofsky who thinks he can prevent us from viewing the real exit numbers, and 4) The media who think they can prevent us from getting the word out to the American people en masse, bad news for the crooks

We are getting the word out, about this totally corrupt carnival like operation that up until now has been called an American election, thanks to the likes of TIA, Its not over till its over.

DECLARE WAR ON THE SECRET VOTE COUNTING MACHINES!!!

Bring it!!!

[nedder08]

21. Is this how our leaders fix nations most imp issues ?

What does everyone think about this video http://www.youtube.com/watch?v=E7150u9sfCo
Dont you think these are grave issues which needs serious thought?

[ZaDomSpremni]

22. Keep in mind rush limbaugh drove thousands of republicans to vote

Rush urged republicans to "pull hillary out of the fire" in order to keep the democrats spending money and attacking each other.

Basically his stance is "Well I can't stand mccain in getting people to vote for him, but here is what I can do to help".

Hillary would have lost both if it weren't for Republican voters voting in the democrat primary, IMHO THOUSANDS of unexpected republicans turned out to vote in ohio and texas and went 3 to 1 for hillary.