Red versus Blue
February 9, 2004 7:24 AM   Subscribe

The Election Projection Website. A semi-scientific website that attempts to forecast the 2004 presidential election. Via Newmark's Door.
posted by trharlan (28 comments total)
 
The two months between September's Republican National Convention and Election Day will be a great time to be Republican. I can't wait!

...

Can we trust a privileged businessman who has served but 6 years in elected office to handle the affairs of the most powerful nation on earth? In 2000, voters put their faith in an untested George W Bush. Four years later, his courageous, principled, and steadfast leadership have led this country through some of its most trying times. Even those who dislike and disagree with President Bush would be hard-pressed to deny the resolve of his leadership. He provided and continues to provide a steady hand when we need it most. Voters will feel eminently more confident to put their trust in him again this year.

...

I had to add this one. Bush has made a career out of having his opponents "misunderestimate" him. They show no signs of realizing that they really aren't dealing with a moron. How many more times will the Democrats ponder, "How did he do that?"

Gotcha, pal.
posted by magullo at 7:34 AM on February 9, 2004


Yeah, I was about to put down the same quote too. By "semi-scientific" you mean "parroting all the GOP-supported talking points to say why Bush is going to win."

Using Schwarzenegger and Dean as reasons for a Democratic Party loss isn't scientific, it's factless partisan opinion. John Zogby is an openly conservative Republican, and he gets a relatively accurate job done. This guy's no Zogby.
posted by XQUZYPHYR at 7:50 AM on February 9, 2004


The map from the site is laughable:


posted by the fire you left me at 8:02 AM on February 9, 2004


Yeah, it's completely biased, esp. if you read the long list of reasons he thinks Bush will win. I am curious why you think the map is laughable? Really, when it comes to electoral votes, any sort of configuration could be possible in a close race -- just matters who wins the state. It really doesn't show how close a race is. I'm surprised to see a Republican concede California, Pennsylvania and Michigan.

After watching the Russert interview yesterday, I now think there's a slight chance Bush can be beaten. It's certainly an uphill fight, and I hope that this web site is widely disseminated so that his forces get cocky.
posted by Slagman at 8:40 AM on February 9, 2004




Most interesting is the poll results he has collected right on the front page:

Job Approval: Approve: 49.0% - Disapprove: 45.8%; Margin down 7.8 to 3.2%

Bush vs. Dem: Bush: 45.4% - Kerry: 48.8%; Margin down 8.0 to -3.4%

Right Track / Re-elect: Pos: 46.2% - Neg: 48.0%; Margin down 0.2 to -1.8%


And most importantly:

National Margin
+0.39% (down from +6.46%)

Bush loses Minnesota
Bush loses Maine (district 1)
Bush loses Michigan
Bush loses Pennsylvania
Bush loses Oregon
Bush loses Wisconsin


Looks like he might need to update his "projection."
posted by monju_bosatsu at 8:56 AM on February 9, 2004


It's not "laughable" but it's unlikely. Basically the map is 2000 giving Iowa over to Bush.

Though the election is clearly going to be decided by a few swing state differences from 2000, Iowa is likely not one of them. Chuck Grassley isn't going to have hardship facing re-election, as such Bush has no secondary need to waste time campaigning there.

Bush is running a defensive campaign, despite the offense both parties will take. The GOP needs to merely maintain the states they won in 2000; it's the Democrats that need to make gains.

It's laughable in the longer run because this guy is making his "calculations" based on rock-solid analysis like "#7- the Democrats are all weak." (Except for the ones outpolling Bush right now. Right.) In other words, he's laughable not because of the map, but because he's taking the map- which is not really a masterpiece of unique calculation- and using it to justify his baseless personal analysis, which is significantly irrelevant.
posted by XQUZYPHYR at 8:59 AM on February 9, 2004


John Zogby is an openly conservative Republican, and he gets a relatively accurate job done.

JZ: "I am personally a Democrat"
posted by donth at 9:10 AM on February 9, 2004


And he doesn't even take the Scalia factor into account.
posted by goethean at 9:49 AM on February 9, 2004


I think it's a well-balanced site, given the open Republican nature of the owner.

As for information, he has all the latest poll figures, with changes from the last polls. The map is based on polling figures, not his opinion, so it seems rather accurate.

The 'swing States' section is especially interesting, given the listen of states not up for contention, and the ones that will in the end decide the race.

The map is always somewhat misleading, since the West and NorthEast states have higher populations, comparitively, to the mid-states.. so even if someone carries the majority of mid-west states (over 50% of the states in the Union, say), but not the biggies.. well, they'll still lose.

The popular opinion poll numbers are good, as well.

I've bookmarked it, thank you. It's a good source of raw data, under the blatant Republican rhetoric.
posted by rich at 9:55 AM on February 9, 2004


I can save'em alot of time. Bush is gonna dance a kazatzky on Kerry's chest. Gives me no great pleasure to say that, either.
posted by RavinDave at 10:29 AM on February 9, 2004


donth: my mistake, I was thinking of Frank Luntz, the guy who does the polling stuff on MSNBC. My bad.
posted by XQUZYPHYR at 10:36 AM on February 9, 2004


There is actually a large academic literature in political science on forecasting presidential elections based on economic conditions and other quantifiable variables. Here are some of the best-known books on the subject.
posted by notme at 10:48 AM on February 9, 2004


Heh, and like most other fair and balanced, semi-scientific analysis of the subject, it ends with:

But don't stop donating and volunteering.  That will play a most critical part in making this view become a reality.  He is certainly not assured of re-election, but, with our continued support and hard work, all signs point to a second term for George W. Bush, the 43rd President of the United States.

Clearly, this guy is just presenting the unbiased, unvarnished truth!
posted by spilon at 11:36 AM on February 9, 2004


A semi-scientific website

If, by "semi-", you mean "wholly non-", then yes...
posted by mkultra at 12:01 PM on February 9, 2004


I stand by a "slight chance." Everyone thought he was a buffoon in 2000 too. Don't take the little dip for granted.
posted by Slagman at 2:09 PM on February 9, 2004


Oh yeah...Zogby's not a Republican. I think NYPost has used his polls, but that's just business. He hasn't endorsed anyone so far but it's most likely Zogby will be working for the Democratic nominee.
posted by Slagman at 2:11 PM on February 9, 2004


OK, here is another one. It's from Ray Fair, professor at Yale. Wall Street Journal had an article [subscription required] about it, and one could argue that was pretty biased (just for the fact that only one model was evaluated).

Here are several paragraphs from WSJ:

[M]r. Fair has been tweaking his formula since about 1972 and it doesn't do a bad job of predicting the vote percentage of the incumbent. In fact, the standard error is a pretty decent: 2.4 percentage points ...

[T]o start with, Mr. Fair doesn't appear to be on the GOP payroll; he told me recently that the latest results aren't to his personal political tastes ...


Here is some real stuff: unlike the model above, the formula is given by regression:

[B]y tinkering and tweaking with a variety of economic data, Mr. Fair has settled on three economic variables. The first is the economic growth rate in the three quarters immediately prior to the election (voters have short memories). The second is the inflation rate during the entire presidential term (they don't forget inflation). The third is what Mr. Fair calls Good News, which is the number of quarters during a presidential term that growth exceeds 3.2% (they have some memory for headlines). ...

The result:

[S]o here's what you get when you plug in, roughly, the consensus economic forecast: President Bush runs away with it, in a laugher, [no comment] with 58% of the vote ...

[B]ut you angry Democrats can save yourselves the trouble. I've tried to create a Democratic victory with the formula and it's nearly impossible. Only a complete collapse of the economy even gets you within the 2.4 percentage point margin of error. You'd have to have -2% growth over the next three quarters and a 4% inflation rate to get there.


You can play with the formula here.

The reason for this has much to do with the specific presidential constant for George W. Bush. While I can't argue with the statistical reasons for it, this is my biggest gripe with the formula. President Bush is handed 55.57% of the vote before the economy is even factored in.

Looking back at past elections, Mr. Fair has found that incumbency is a very lofty perch. He's also found a slight edge for Republicans. And, it's significant that President Bush is running for re-election after the opposition party (the Democrats) had most recently occupied the White House. You old timers and presidential history buffs are now nodding. ...

[F]inally, the formula has been badly wrong only once: that was in 1992, when it predicted the incumbent would win with 51.7% of the vote. It was off by 5.1 percentage points. That, of course, was the last time a man named Bush ran for re-election.

posted by MzB at 2:37 PM on February 9, 2004


How do you factor in that the incumbent is an idiot?
posted by Slagman at 4:16 PM on February 9, 2004


[I tend to ignore this kind of remarks, but ... ]
"Use a dummy variable!"
posted by MzB at 4:34 PM on February 9, 2004


Wow, MzB. Very well done.
posted by Civil_Disobedient at 8:52 PM on February 9, 2004


In the meantime, here is a Reader's Digest version of a possible first debate between Kerry and Bush:

Bush: War on terror
Kerry: Vietnam
Bush: Evil doers
Kerry: Band of Brothers
Bush: Grave and growing
Kerry: Imminent?
Bush: Let me step back a moment
Kerry: (sensing weakness) Mekong Delta
Bush: (regaining composure) Massachusetts Liberal
Kerry: I don't like to talk about it, but I'm a war hero.
Bush: Bring it On
Kerry: No, I have three words for you. Bring ... It ... On
Bush: Let's Roll
Kerry: I'm coming, you're going and...
Bush: I haven't heard you talk like that since our Skull and Bones initiation week.
Kerry: ... don't let the door hit you on the way out.
Bush: (pounding a fist) The Almighty, Faith-based, Mel Gibson, God Bless
Kerry: Ketchup fortune, good hair, tall, I can still drink
Bush: Fake Vietnam stories
Kerry: That was Bob Kerry
Bush: Dukakis
Kerry: Ashcroft
Bush: Lobbyists
Kerry: Halliburton
Bush: The South!
Kerry: Smart people
Bush: You're rich!
Kerry: You love the rich
Bush: Gay marriage, Sodomy, the L word, Carson Kressley
Kerry: Ralph Reed?
Bush: Shadows, caves, murderers
Kerry: Program related activities?
Bush: (with the classic accusatory smirk) ... Senator!
Kerry: (taken aback) ... (long pause) ... Flightsuit!
Bush: Military spending dove
Kerry: Flightsuit
Bush: Votes to cut funding for intelligence agencies
Kerry: Flightsuit
Bush: Bring it On
Kerry: Bring it On

posted by fold_and_mutilate at 9:26 PM on February 9, 2004


That Presidential Vote Equation looks about as useful as the "taller candidate always wins" theory, which has almost as good a record.

I mean, it's only dealing with a couple dozen elections, and you get to do things like include "WAR = 1 for the elections of 1920, 1944, and 1948 and 0 otherwise." It doesn't seem like it should be all that difficult to come up with a set of factors that work most of the time if you can just pick anything related to the economy and throw it in.

I bet if I had a few weeks to spare, I could come up with an equally convincing formula predicting the outcome based on fashion trends. "ZIPPER_INVENTED=1 for elections > 1924, 0 otherwise."
posted by sfenders at 9:40 PM on February 9, 2004


Bush is going to be defeated in a landslide. No matter who he runs against he doesn't stand a chance in hell.
posted by filchyboy at 9:46 PM on February 9, 2004


The map is identical to 2000 with only Iowa and New Mexico swinging the other way. Iowa will still be tough for Bush to win, now with Kerry as the presumption nominee, considering Kerry got his oomph there, and Iowa hasn't gone Republican since Reagan in 1984. New Mexico, I know less about, politically.

But Kerry's winning *all* of Gore's states still leaves him 10 electoral votes short, due to reapportionment since the last election. Bush would win 278 to 260.

So what does Kerry have to do? Besides winning Gore's states, he maybe needs a vice presidential nominee who can deliver one f-ing state of 10+ votes to his column that Gore didn't win. Examples would be

1. Florida. Anybody who thinks they know which way Florida will go is fantasizing.
2. Ohio (trended more Republican in last election, but there's been a lot of lost manufacturing jobs there)
3. Tennessee
4. Missouri
5. ***North Carolina***

or put together a couple suprises, even, say Colorado+Nevada, or Arizona+New Hampshire. Looking at how New Hampshire broke in the primary, I look like Bush could be beaten there. New Hampshire voters hate being taken for granted.

That's how you beat Bush.
posted by jamsterdam at 10:21 PM on February 9, 2004


It's way too early for a projection like this to be worth anything but amusement. In fact, astrology would be more dependable. A lot can happen in ten months, for better and for worse on all sides. This projection disregards third party intrusions. This projection can't predict world events or changes in Iraq or new disclosures about Bush's administration which will come to light between now and then. We don't even know who the democrat nominee's gonna be. We assume Kerry but that's not cut and dried yet.

It's still anybody's ballgame.
posted by ZachsMind at 10:33 PM on February 9, 2004


Also noteworthy (well, sort of) are the predictions from the Annals of Improbable Research. Who Should Win the 2004 U.S. Presidential Election? Bush beats any Democratic nominee except for Clark or Dean, unless he authorizes the combat use of nuclear weapons, in which case Bush wins outright.
posted by yarmond at 11:26 PM on February 9, 2004


MzB: your formula making idol should learn about over fitting. There are literally infinite (and uncountable even) formulas that will fit all available data and still provide no useful predicative power, if all you do is look at variables without coming up with a theory. There just isn't enough data to come up with a good formula based on the economy now. It also dosn't figure in job losses.

---

Also, Kerry totaly sucks ass. Even edwards would be better. We could continue the trend of retard presidents that way.
posted by delmoi at 12:15 AM on February 10, 2004


« Older vr tour   |   Airline Faith Newer »


This thread has been archived and is closed to new comments