More on the Revision of NEP National Exit Poll Hispanic Data
On Thursday, I covered the downward revision of the NEP exit poll's national Hispanic support for Bush from 44 percent to 40 percent. That swings their Hispanic vote estimate from 53-44 Kerry to 58-40 Kerry. Quite a change: that doubles Kerry's margin among Hispanics from 9 to 18 points. And personally I believe that 40 percent figure is still a touch high and I certainly believe there are still an abundance of unanswered questions about this year's Hispanic results, both original and revised.
Here are some additional materials about the Hispanic results and revisions that you may find helpful. Mark Blumenthal of Mystery Pollster has a post about the revised national Hispanic figures which goes into some detail on a few questions raised by the revision. And the William C. Velasquez Institute (WCVI), which did their own exit poll of Hispanics that indicated a 65-33 lead for Kerry, has a useful press release on the NEP revisions (national and TX) and summarizing their position on Hispanics and the 2004 election. Here's a quotation from the Institute's president, Antonio Gonzalez, on their position:
There is no doubt that some churning of numbers has occurred, meaning Republicans appear to have made significant gains in Texas and Arizona while Democrats appear to have made significant gains in Colorado and Florida. But the net effect among these respective gains is a canceling out of one another. Latino voter partisanship has remained consistent with roughly a 30 point democratic advantage in 2000 and 2004's presidential elections.
WCVI also provides on their website an analysis of their exit poll data by St Mary's University political scientist, Henry Flores, and an extensive powerpoint presentation on their poll's findings.