2004 November 08 Monday
Steve Sailer: Exit Poll Estimates For Hispanic Vote Not Credible
The claim that George W. Bush nationally won 44% of the Hispanic vote is based on an absolutely huge and unbelievable 59% of Hispanics voting for Bush. Blue/Red voting results by county for 2000 and 2004 show few shifts in Texas. For the Hispanic vote for Bush in Texas to have gone up so much the white vote for Bush would had to have gone down.
The exit poll claims Bush's share of Texas Hispanics leapt from 43 percent to a staggering 59 percent. (My recollection is that this Texas figure was originally something like 52 percent, but in the rejiggering, it was inflated to an unlikely 59 percent in the final numbers.) Texas is what's driving this 44 percent national figure.
This is particularly odd because you would think such a shocking improvement with an important bloc in Texas would lead to a much better overall performance by Bush in his home state. Yet, Bush's growth in his share of the total vote in Texas was only 1.9 percentage points, below his national average of 3.1. The exit poll tries to explain this by claiming that—while Bush's share of the white vote grew by four points nationally—in Texas it shrank by 1 point, which seems odd, to say the least.
If 59% of Texas Hispanics supported Bush, then Bush should have carried just about every county in the state. But most of the heavily Hispanic
Rio Grande Valley remained firmly in Kerry's grasp. Of the 15 Texas counties lost by Bush, 13 had Hispanic populations of 75.0% to 94.3%. The other two were Travis County (Austin), a college and government town, and Jefferson County in the East, which is 32% black.
The Hispanic vote in Texas also would have to be almost twice the 32% of California Hispanics that voted for Bush according to exit polling in California. Why would that be the case? How could national polls before the election have Kerry having about a 2-to-1 advantage with Hispanics and then in a short period of time have half that advantage disappear?
The Rove/Bush strategy did not make any remarkable gains with Hispanics in other states. In the rest of the country Bush may not have surpassed Reagan's 1984 share of the Hispanic vote (reported in some places to have been 35%). The Hispanics are going to remain overwhelmingly Democrats and white people are going to continue to shift from the Democratic Party to the Republican Party. Eventually America will become politically much more divided based on race. A drastic reduction in the current immigration rate could reduce the depth of that growing split.
Update: Michelle Malkin covers the Hispanic vote question and links to a Mike Tolson piece in the Houston Chronicle on measuring Hispanic voting behavior.
The institute, essentially a wing of the San Antonio-based Southwest Voter Registration Education Project, found itself at odds with the numbers put out by the main national exit poll in the previous two national elections. The problem, Gonzalez said, lies in the way the poll goes about collecting them.
"Network and media surveys are not designed to measure Latinos," Gonzalez said. "They are designed to measure the general market. The Latinos are not suburban. We're the most urban electorate in America. There are not lots of rural or suburban Latinos anywhere. What you get when you have a general market survey is one that shows more Latinos who are Republican."
What may have happened was that minority activists and their allies obtained influence within the exit-poll organization. The objective would have been to exaggerate the minority vote, and to try to push the election to Kerry. When that didn't work, perhaps their machinations were discovered by higher-ups; who proceeded to adjust the minority vote way down. Dick Morris reported the unadjusted exit-poll as saying that "hispanics accounted for 12"% (searchable), the adjusted exit-poll reduced that to 8%. Somehow it appears that ~100%(!) of these latino respondents were Kerry supporters. The bulk of the adjustment to bring the exit-poll into line with the actual voting seems to have come from the deletion of latino Kerry voters, mainly in FL and TX. To approximate reality, they now need to eliminate 2 points of latino Bush votes, and 2 points of black votes. Another sign of the fictionalization of the exit poll, is the 36% of Nader's votes which supposedly came from latinos. A clear indication of monkeying also shows in the reduction of the non-hispanic white proportion of the total vote from 81 to 77%; it is really four points higher.
More information highly relevant to this issue is in Tolson's Houston Chronicle article entitled "exit poll math", and Michelle Malkin's site( 11-7-04 ).
Thanks, John, most interesting. Can you provide the URL for that Tolson article? Thanks.
The Chronicle article is at: http////www.chron.com/CS/CDA/ssistory.mpl/politics/2886704 and is also accessible through michellemalkin.com 11-7-04 'whither the Hispanic vote'. The exit poll is in the above article stated to have given the latino vote as 10 million when only 7 million are registered. Not all registered voters of any such class will vote; the exit poll needed to get rid of 4 points of Kerry voters from their estimate in a hurry. It can be deduced that this is where they made their deletions of excess Kerry votes.
The update has the links to Malkin and Tolson; I redundantly referenced them. Robert Gebeloff reports in 'exit poll data proves to be wrong'..., that in NJ, the exit poll oversampled minorities wildly, all the way up to 30%, from 20% in 2000. This will be hard to get the exit poll managers to admit to; that they just removed several million Latino Kerry supporters from their oversampling of minorities, because they had to eliminate Kerry votes somewhere.
Steve, I added an update to this post that includes the links Mike Tolson's article and Michelle Malkin's post. I see that John has also now included the links here as well (though I haven't checked his links I know mine are correct).
CNN's exit poll reports that Hispanics voted 64% for Bush in the southern states, this proportion was up 29 points from 2000. This is a major fiction to cover up some not-so-artful discarding of exit-poll latino-Kerry support, for convenience' sake. The South is a very easy region for them to do this; since most of the states there don't even have any separate latino vote breakdown. TX and OK were also wonderfully convenient places for the exit-poll overseers to ditch Latino Kerry support from their poll with stunning abandon. They must believe that no one cares about the truth; that the right and the left will maintain an omerta on this manipulation, because now Republicans can say they're not racist, they can get latinos to vote for them.
From CNN's exit poll reports, it can be inferred that the Hispanic voters in the south, living outside TX and FL, would've had to vote more than 2-to-1 for Bush, if the exit polls numbers are not falsified. TX is included in their numbers for the south, the totality of which supposedly had latino Bush support of 64%; and TX and FL together are ~40% of the population of this group of southern states, presumably. TX and FL were reported at ~58% Hispanic Bush voting; therefore the remainder of the south must have had something like ten points higher Latino Bush support, if the exit poll didn't just make up numbers to suit their needs.
Right, I dug up the Southern numbers last night. That's the smoking gun. Who knows, Bush might have carried over 100% of Hispanic vote in the smaller Southern states!
What's the motivation here?
The motivation for the exit-poll subterfuge, which inflates the Hispanic Bush vote in the southern states, would be that they had such a high oversampling of latino votes nationally, that it had to be covered-up somewhere. This is a major falsification, to cover up a prior fraud, of a few hours earlier, probably. The exit-poll may have oversampled latinos as much as 2-to-1, countrywide. Whether this would have been deliberate or accidental, is another inquiry. Faced with a large oversampling, and the simultaneous need to delete Kerry support somewhere, they would have picked the latino Kerry vote in the southern states. Yet if you delete almost exclusively, the Kerry-supporting respondents, this causes the Bush percentage to rise to unprecedented, unbelievable and obviously falsified levels. That is, in that group; and contradicting all the other polls. Thus ~30% turns into ~45%, by corrupt legerdemain. Such an occurrence would signify also absolute contempt for democracy and for truth, to say the least. Another motivation to do it in this exact way, might be anti-caucasianism; such and such a media monopoly can easily be suborned to an attitude like that. They would consider themselves the 'conscience of society' for cheating to minimize the importance of the majority's vote. A more pressing motivation would be the desire to hide their errors, or the fraud by which the oversampling occurred in the first instance. Another effect is the delegitimization of the election; if many were to take the unadjusted exit-poll results at face value, and say that the later adjustments were subservient to the putative stealing of the election.