Background on the outlier Gallup poll (now 5% Bush lead):
Tuesday :: Oct 19, 2004
Yesterday's Gallup Poll Assumed Reduced Minority Vote From 2000, Greater Number Of Upper-End Conservative Voters
Now that the Gallup Poll from yesterday has been blasted far and wide as proof that Bush has supposedly surged to an 8% lead over Kerry despite the debates, it is warranted to look a little closer at the composition of that likely voter sample that was used by Gallup this week to project that if the election was held this week, Bush would defeat Kerry 52%-44%. Remember that Gallup and other pollsters who refuse to weight their samples for party identification say that weighting samples to reflect demographic and census data is more reliable than using party ID as a factor. Yet after looking closely at the demographic breakdown of the sample that Gallup used to reach their conclusions, it becomes clear that Gallup has become the in-house pollster for Karl Rove and the GOP’s view of how the American voting populace should look. Why?
Because according to Gallup’s poll this week, they expect the electorate to be 85% white, 41% conservative while only 19% liberal, and a third to make over $75,000 per year.
Well, of course an electorate of that composition would vote for Bush by 8% over Kerry. But is that the electorate we’ll see this year? And does Gallup really believe that only 15% of those who go to the polls this year will be nonwhites, or are they privy to some inside information from the GOP about minority voter suppression that the rest of us don’t have?
Gallup has been very forthcoming in responding to my requests for information, and they obviously stand behind their methodology and assumptions. But take a look at the demographic breakdown of this week's sample, and ask yourself how likely is this to replicate itself on Election Day, given the increases in registration this year?
Total Weighted Sample: 557 Likely Voters
(2000 exit poll actual results in parentheses)
By Political Ideology:
Conservative: 41% (29%)
Moderate: 41% (50%)
Liberal: 18% (20%)
Party ID:
GOP: 39% (35%)
Dem: 35% (39%)
Ind: 25% (27%)
Income:
Over $75,000: 32% (28%)
$50-75,000: 16% (25%)
$30-50,000: 26% (24%)
$20-30,000: 11%
Under $20K: 9%
And if this wasn’t bad enough,
Race:
White: 85% (81%)
NonWhite: 15% (19%)
Black: (a subset of NonWhite) 8% (10%)
I think those numbers speak for themselves. Gallup is using a sample that assumes 1) depressed minority participation this year from 2000, 2) assumes a drastically higher participation of conservatives as compared to 2000, and 3) predicts that Bush would win by 8% with that electorate. It should also be noted that Gallup's LV sample contains only 11% in the 18-29 year old age grouping, compared to 17% in the 2000 exit polls. Recent polls show that Kerry/Edwards is doing well with younger voters.
So what you have is a LV sample being used by Gallup that disenfranchises minorities and younger voters. And to Gallup this is entirely credible.
But in a gross example of political disinformation that is ideal for the White House and the GOP, these results were broadcast far and wide this week by a lazy and gullible media as representative of likely voters to show that John Kerry is losing support. |
