How did Bush win the 2004 election?

[bup]

@John Mace

And, before someone resurrects the “exit polls prove Kerry should’ve won” meme, here’s one good artilce explaining what happened from the Washington Post. You can easily google up others.

New Woes Surface in Use of Estimates

Two key interrelated problems are highlighted:

1. Prelimary results leaked out too early, and a server crash prevented a midday update from getting out on time (hence what appeared to be, but wasn’t actually, a surprisingly large swing in exit polling from morning to evening)

2. Early morning exit polling is skewed due to demographics of those voting at that time of day. For example, the early morning participation in the exit poll was 59% female.

I’ve actually seen that article before, and don’t find it too impressive. I’m certain that exit poll predictions take things like ‘59% of the people we talked to were female’ into account.

This in particular confuses me:

@Washington Post

That is wrong, countered Joe Lenski of Edison Media Research, which conducted Tuesday’s exit poll with Mitofsky International for the National Election Pool, a consortium of the major television networks and the Associated Press. “No wrong projections [of winners] were made; the projections were spot on,” he said. “The members used this data with sophistication and understanding of what data can and cannot be used for.”

Did he mean without sophistication and understanding…? Or what?

The Mystery Pollster does a good job with it, I think, and comes down on the side of there not being a huge statistical anomaly, but allowing that it’s a complicated mess. Here’s a good example.

[John_Mace]

@bup

Yeah. Hard to read thread. Not a loaded question: what sort of evidence do you (or anybody) think would turn up if there had been actual voter fraud? That’s the frustrating thing for me.

There could be any number of pieces of evidence, but the bottom line is that it would be part of a report that actually concluded there actually was enough fraud to tip the election, not one that says there could have been some problems.

[John_Mace]

@bup

I’ve actually seen that article before, and don’t find it too impressive. I’m certain that exit poll predictions take things like ‘59% of the people we talked to were female’ into account.

Don’t be so sure. Read the Slate article I also cited. Both sides accuse the other of interpetting the data incorrectly. But neither side (the media vs the polling companies) agrees that the early data, as reported by the media, is an indication that the election was fixed.

[bup]

@John Mace

There could be any number of pieces of evidence, but the bottom line is that it would be part of a report that actually concluded there actually was enough fraud to tip the election, not one that says there could have been some problems.

Oh - agreed. Like I said I can’t look at pdf now, but I agree that if the report only says the henhouse is unlocked, that’s not enough to accuse.

Imagine, though, that you believed in the “truthiness” of there being something amiss in Ohio. What evidence would you look for? What telltale things would you think BlackBox or somebody should look for that would constitute compelling evidence of fraud?

[David_God_of_Frogs]

The more useful question is why the Democrats came as close as they did. Bush got 53% of the popular vote, which is almost a comfortable margin.

[Sequent]

@David

The more useful question is why the Democrats came as close as they did. Bush got 53% of the popular vote, which is almost a comfortable margin.

Where are you getting 53%? Wikipedia sez it was 50.7%

[Sequent]

From the above article:

Although Bush received a majority of the popular vote: 50.73% to Kerry’s 48.27%, it was—percentage-wise—the closest popular margin ever for a sitting President; Bush received 2.5% more than Kerry; the closest previous margin won by a sitting President was 3.2% for Woodrow Wilson in 1916. In terms of absolute number of popular votes, his victory margin (approximately 3 million votes) was the smallest of any sitting President since Harry S. Truman in 1948.

Hardly what I’d call “comfortable.”

[David_God_of_Frogs]

never mind, I must be smoking something. I think I was thinking 3% margin, but that wouldn’t make sense either.

[John_Mace]

@Sequent

From the above article: Hardly what I’d call “comfortable.”

Not confortable at all. In fact, I bet the Bush folks were sweating bullets all day.

Personally, I was shocked that he won. It may very well have been the ObL tape that tipped it*. Was there any state that had an ant-SSM initiative on the ballot that was a close state? If not, I don’t see how that would have made much of a difference.

*to fall prey to the very fallacy I talked about in my first post. Just because something happened last, doesn’t mean it was any more a deciding factor than any other issue that came up earlier.

[dalej42]

@John Mace

Did Ohio have a SSM initiative on the ballot?

Yes.

[John_Mace]

@dalej42

Yes.

Hmm. Damn clever, them Republicans.

[Sequent]

http://en.wikipedia.org/wiki/Same-sex_marriage_in_the_United_States

On November 2, 2004 (Election Day), state constitutional amendments prohibiting same-sex marriage were passed in eleven states: Arkansas, Georgia, Kentucky, Michigan, Mississippi, Montana, North Dakota, Oklahoma, Ohio, Oregon, and Utah (see: Utah Constitutional Amendment 3).

[Apos]

It’s worth noting that most of those states already had passed anti-same sex marriage laws, adn that getting them on the ballot was something my counterparts in the Ohio Republican party spent a fair amount of time on early on. I note that Virginia is preparing an anti-SSM vote that could end up bumped over into 2008 (the bill is currently the particularly nasty kind that seeks to prevent gays from even gaining the legal incidents of marriage.)…

I have a question: before the election, changes in the terror alert level came pretty commonly, along with warnings of potential threats. After the election, it seemed like all this stuff stopped. Am I right in this perception? Has the raising of the terror alert level happened much since 2004 and was it publicized as heavily?

[John_Mace]

@Apos

I have a question: before the election, changes in the terror alert level came pretty commonly, along with warnings of potential threats. After the election, it seemed like all this stuff stopped. Am I right in this perception? Has the raising of the terror alert level happened much since 2004 and was it publicized as heavily?

I don’t think the election is a relevant demarkation point. The last time the Homeland Security Advisory System was raised nationwide was late 2003 thru early 2004:

On a blanket, nationwide level, it has been raised to orange five times:

* September 10 - September 24, 2002, around the first anniversary of the September 11, 2001.
* February 7 - February 27, 2003, near the end of the Muslim religious holiday Hajj. Intelligence reports suggested the possiblity of terrorist attacks against "apartment buildings, hotels, and other soft or lightly secured targets."
* March 17 - April 16, 2003, around the beginning of U.S. and Coalition military action in Iraq.
* May 20 - May 30, 2003, "The U.S. Intelligence Community believes that Al Qaida has entered an operational period worldwide, and this may include terrorist attacks in the United States." - Tom Ridge
* December 21, 2003 - January 9, 2004, citing intelligence information suggesting large-scale attacks around the holiday season.

[SuaSponte]

I read an article in The New Yorker a while back that reported on a conference held after the election that included the top pollsters and strategists from the Bush and Kerry camps (can’t find the article on TNY’s web site - I think it was before February 2005).

The analysis of the election results agreed to by the participants was that for a small but significant percentage of actual voters, the only issue was security, and Bush got the votes of some 90% of that group. And they were the margin of victory, including in Ohio.

The conclusion was only someone like Wesley Clark had a shot to beat Bush.

Sua

[SentientMeat]

@Sua

for a small but significant percentage of actual voters, the only issue was security, and Bush got the votes of some 90% of that group.

And one must ask on what basis? (PDF)

[SuaSponte]

@SentientMeat

And one must ask on what basis? (PDF)

No, one musn’t. The OP presents for debate the issue of how Bush won. It does not present for debate whether Bush should have won on those grounds.

Sua

[SentientMeat]

@SuaSponte

No, one musn’t. The OP presents for debate the issue of how Bush won. It does not present for debate whether Bush should have won on those grounds.

Very well. In answer to the OP I propose that part of the reason Bush won was because his voters believed demonstrably factually incorrect statements such as “Iraq was directly involved in the 9/11 attacks, or gave Al Qaeda substantial support”, “the US found the Iraqi WMD’s after the war” or “the majority of world opinion favoured the US invasion of Iraq” to a far greater extent than Kerry supporters.

[spifflog]

@John Mace

Did Ohio have a SSM initiative on the ballot?

But you’re falling for a common fallacy-- it’s like when you watch a basketball game and think that the last few minutes are what determined the winner. Every minute counts just as much as the final minutes. And every battleground state could be considered the key. Florida was close last time (actually would have gone to Gore if all the ballots had been recounted), but clearly went to Bus this time. Maybe Florida was the key, not Ohio. Also, you’re basically assuming that Kerry shoudl have won, but something went haywire. Why not look at some states that Kerry narrowly won and ask why they didn’t tip over the Bush?

The thing that turned the election was the security issue. In the end, people felt safer with Bush, rightly or wrongly. The ObL tape a few days before the election may have scared enough people…

All the ballots were recounted. After the fact, the USA Today, Washington Post and others did a recount of all the ballots. Surprisingly enough, when the recoount still went to Bush, the story became page 10 news.

[Digital_Stimulus]

@John Mace

I don’t think the election is a relevant demarkation point. The last time the Homeland Security Advisory System was raised nationwide was late 2003 thru early 2004:

That can’t be right; I’m not totally devoid of memory. Unless the difference is between “nationwide” and, say, NYC, levels.

Certainly, there were numerous occasions throughout the summer leading up to the election where the level was raised and consequently brought the spectre of attacks back into the national spotlight.