U.S. Academics Use Formulas to Predict Bush Win

Personally, I've long maintained that Bush would have a comfortable win. But I'm less sure of that these days.

From your link:

Thanks for that, PDiddie.

I don't know what I think. If my voodoo works, Bush will make a jaw-droppingly horrible mistake in public -- debates, preferably -- and that that'll be that. But my voodoo doesn't have a spotless record.

"But one glaring error is what the forecasters are perhaps best remembered for: they predicted in 2000 that Democrat Al Gore would win easily, pegging his total at between 53 and 60 percent of the two-party vote."

Umm, well.... What was the actual total?

We shouldn't need voodoo, Sozobe, just more people paying attention.

Okay, sigh... we do need voodoo.

I saw a piece a week or so ago that noted this formula (and what it suggested) but placed it alongside some other 'tried and true' predictors which pointed elsewhere, and it noted the Gore prediction as well....can't recall where I read that, sorry.

No model could have predicted the shenanigans that went on in the 2000 election. Gore won the popular vote, he also would have won Florida by 105 votes according to the New York Times if the vote had been honestly counted . I think the model must be taken seriously and the question asked why is it predicting a win. Kerry has four month to prove it wrong.

Significantly less, Suzy.

Gore got 48.4% of the popular vote (Bush 47.9%).

So the forecast was really way off there.

Nothing to do with the Florida shenanigans, Acquiunk.

I'll buy it, and not just because I'm in the market for this bit of news.

For this reason: Bush has had really tough polls. All his babies were in the toilet until the economy and Iraq began looking better.

Despite Bush's bad news and polling data---Kerry could never turn it into an advantage. There is something very wrong about that.

When you have two guys vying for something--and one is doing poorly--what reason do you possibly cite for the competitor not gaining? I could be wrong, but I really don't think he's an accepted player--at least at this point.

From the article

"The forecasters chalk up the 2000 error to Gore's campaign, which distanced itself from the Clinton record. All the models assume the candidates will run reasonably competent campaigns, said Thomas Holbrook, a professor at the University of Wisconsin at Milwaukee."


Gore did not run a competent campaign, and that was obvious by October 2000. The model also predicted a Gore win and despite his incompetence as a campaigner he did win. When an election is as controversial has the 2000 election was, I am less inclined to blame the model than to ask what did it predict compared to the reasons for what actually happened. The model assumed a competent campaign and an honest election. Neither of which occurred and no predictive model can factor for incompetence and malfeasance accurately.

To return to my original point, what are the elements that cause this model to predict a Bush win? Those must be seriously considered and addressed,

It aint over yet, Sofia.

Hm. I still think the second half of that argument - the "honest election" and "malfeasance" one - is really neither here nor there in this context.

I mean, here you are referring to the x thousand of votes that, in your opinion, were wrongly disallowed from being cast/counted in Florida?

Even if all of the disputed Florida votes you're talking about had been allowed to be cast/counted, that still would have only added zero-point-something percent to Gore's vote.

So as an argument to explain a 5-12% misestimation with, it counts for practically zilch.

As for how competent a campaign was handled ... did Dole run that competent a campaign in '96, when the academics' prediction turned out to be spot on? And Dukakis in '88?

The model assumes that incumbency, a rising economy. and a unified party are all positive for a candidate. In 1988 George Bush Senior had all three, Dukakis did not run a campaign sufficiently distinctive to counter those while Bush practically hugged Reagan, a popular president. In 1992 Bush had a falling economy and a third party insurgency that took nearly 19% of the vote. Still Clinton won with only 47% of the vote. In 1996 Dole had to deal with incumbency a rising economy and a third party insurgency that took over 8% of the vote. Still Clinton won with less than 50 percent of the vote. Neither Dukakis, Bush or Dole ran incompetent campaigns, they simple were not attractive enough to overcome the negative factors they had to deal with. Gore on the other hand was heir to incumbency, a nation internationally equal to none, a spectacular economy and did not have a serious third party challenge. He lost by abandoning the advantages of incumbency, running a disorganized campaign and presenting a muddled message. He still beat George W Bush Jr 48 to 47 percent and lost by allowing the Republicans to steal the Florida vote. Models in the American winner take all elections are less important for their predicted percentages than for their general statement, win or lose. The model predicted Gore would win, and by the numbers, which is ultimately what these models deal with, he did. It cannot predict poor campaigning, a corrupted election process or an ideologically biased Supreme Court.

Formula for predicting president
I just read a different formula for predicting who will win the presidency using a mathematical system based on experience, etc. It correctly predicts every winner since George Washington. See site at "members.bellatlantic.net" - which you can find if you search under "formula for predicting president." according to this formula, Bush should win over Kerry.