Bush Landslide (in Theory)! Questions for Ray C. Fair

The New York Times ^ | August 15, 2004 | Deborah Solomon

[Conservative_boy_in_Bangkok]

"the entire staff of the New York Times threatened to drink cyanide if Bush wins" If oly we could trust the New York Times

[WOSG]

NYT headlines I'd love to see:

Economy Continues to Grow Strongly; Lamestream Media Frets

Is 4.8% growth bad? Polls show Bush not getting credit he deserves on economy

Productivity surges ahead for 4th straight year

Despite July numbers, 1.5 million jobs created in past year, since 2003 tax cuts

[WOSG]

Your hypothetical has 0% chance of occuring, so why ask.
Kerry/Edwards is the ticket.

[The G Man]

Albert Wang rocks, dude!

[rebel_yell2]

How do you "know" that he is "data mining?" It is standard procedure to make one's predictions "out-of-sample" (what you call "prospective.") One simply fits the econometric model without the event to be forecast, e.g. fit the model with all past elections except one. Fair may well have done this to "predict" each election since 1916. Or he may have not. Unless you know the facts, you should not be dismissing his research.

[Conservative_boy_in_Bangkok]

Excuse me. do you even read the other posts? Some one else originally posed the question, "what if Kerry drops" I just took it step further, and even if no one else posed the question. There is nothing wrong with asking it SO BACK OFF.

[rebel_yell2]

More important than cell phones is the "gate-keeper" phenomenom. People now use answering machines, voicemail and caller ID to screen their calls so that the pollsters never reach them. This has drastically reduced response rates to the point where "random" samples lose their randomness, and, therefore, their validity. Surveys should have a response rate of at least 50 percent and preferably 80%. Most now come in at under 20%. So they simply call more and more people and end up with a badly skewed set of responses.

[rebel_yell2]

It saddens me that you teach this to students at Yale, who could be thinking about society in complex and meaningful ways.

A reporter actually said this?!!!

This isn't an interview; it's an editorial. Implicitly, she wants social issues to be "important" in the equation. The data don't support her liberal world-view, so she is "saddened" that he is teaching this fact to his students.

[My2Cents]

Popular vote: Bush - 56%; Kerry - 42%; Others - 2%; to go along with your 45 state blowout.

[My2Cents]

it will be a hard nail in the coffin of the professional pollsters

That will be a blessing. There's one area of the economy that I hope is hit hard with layoffs after November: polling organizations.

[My2Cents]

My theory on the current polls is that their samples are based upon past voter turnout, which are inherently flawed in giving a picture what is happening throughout the country. This year, I believe, as the first presidential election after 9-11, will result in a unforeseen turnout favorable to Pres. Bush. There will be major crow-eating among media pollsters and pundits on the evening of Nov. 2.

[McGavin999]

It is vitally important, for the wellbeing of this country, that this be true. First of all, it will shut the euros up! Second, it will bring the dems to their senses. They have gone so far out into left field that it is not healthy for this country. Someone needs to put a sock in the mouths of those who are damaging the country and the only ones who can do that is the democrats themselves.

[labard1]

"you should not be dismissing his research."

I'm not "dismissing" it. I'm just trying to explain why there may be less here than appears at first blush. I think what he's doing is fun, and I hope he's right. And he's an engaging guy-- not a stuffed shirt. However, I've already posted why I have doubts about his methodology.

The professor says all factors other than economics are only worth about 2.5%. Most of the time that may be true. Sometimes it's VERY wrong. So the formula is right until it isn't. Just like "no Republican has ever won without carrying (fill in your favorite jurisdiction)" is right until it isn't. There are compelling reasons why economics is important in an election. But sometimes (perhaps only rarely) other things are more important. Then the formula can be disastrously wrong. And the deviation can be a lot more than 2.5% then.

[labard1]

You're dead on about the "the "gate-keeper" phenomenon." When I did this stuff almost 40 years ago, we often got response rates above 90%. Those days are long gone.

[nyconse]

Make that November 3rd. First they must try to keep the polls open in Dem areas...cry, refuse to give GW states that he won (holding back on even Georgia for hours in 2000). On November 3rd, the media will announce that Bush's attack dogs swayed the election at the last minute and of course the evil (also stupid president) is not legitimate. They will entreat the Dems to hold the line on judges (last chance for Dems since they can't get elected it seems). The kool-aide drinkers will lose more elections (slow learners) before they admit maybe it's the Democratic message...the American people have rejected.