Tuesday, September 28, 2004

Frank Gallup is Turning Over in His Grave

Frank was proud of the quality of his polling organization. In fact, we all were, and Gallup became the generic name for all scientific polling. Alas, things are not the same today:

...according to http://pollingreport.com, August surveys of likely voters showed Kerry up by one to seven points in nine national polls and Bush up by three to four points in three national polls. Curiously, all three of the outlier Bush-favoring polls were conducted by Gallup. The other nine were conducted by nine different polling organizations using nine different polling methods. One has to conclude that Gallup's methods are seriously flawed.
Recently, The Wall Street Journal, Pew, Harris, Investor's Business Daily and Democracy Now all released polls showing the race was neck and neck again with neither candidate ahead by more than two points. On Sept. 17, Gallup released a poll claiming that Bush is up by 13 points nationally.

Of course, it could be that the Gallup polls are correct and all the others wrong. But the fact that Gallup was wrong in the 2000 elections, too, makes me suspect that something is wrong in their polling methods.

The most recent Gallup polls suggest that Bush has an eight percent lead among likely voters. The reasons for this sizable lead might be in the way the Gallup methodology defines 'likely voters'. For some reason the filter they use lets more Republicans through than Democrats. Steve Soto points out the consequences of this. The latest Gallup poll had the following sample proportions:

Likely Voter Sample Party IDs – Poll of September 24-26
Reflected Bush Winning by 52%-44%
Total Sample: 758
GOP: 328 (43%)
Dem: 236 (31%)
Ind: 189 (25%)

No wonder that the poll predicts a strong Bush lead; after all, the Republicans are the largest group of those polled. This wouldn't be a problem if Republicans in fact voted in these very same proportions compared to Democrats and Independents. But this is not what happened in the 2000 elections. If it is not what happens this year, Gallup will, once again, fail in its predictions.

Other pollsters appear not to use the same filter that gives a concentration of Republicans in the Gallup polls. It would be interesting to know why various pollsters pick the filters they do, but it would also be interesting to know if Gallup's outlier status has something to do with this:

...James Clifton who bought the Gallup organization is a big Republican donor. He gave thousands to Right Wing Republican Georgia Senate candidate Herman Cain. Cain ran as a huge backer of cutting taxes for the wealthiest Americans. This is essentially the same tax position supported with vigor by the Bush-Cheney ticket.

Link to Steve Soto's blog The Left Coaster via Kos.