The experts on Hispanic voting in Texas, the nonpartisan Willie Velasquez Institute in San Antonio, reported this breakdown for Hispanics in the governor’s race:

Bell 39.6%
Strayhorn 28.6%
Friedman 14.3%
Perry 13.9%
Other 3.6%

The Velasquez Institute based its findings on exit polls of 440 respondents conducted in 32 selected precincts across the state.

As much as I respect the Velasquez Institute’s work over the years, I don’t believe these numbers. Perry campaigned hard in the Valley. He had the support of a number of (presumably) Democratic mayors–mayor being a nonpartisan office in this state. Democratic sheriffs appeared on his TV commercials. One of the few major pre-election polls that released its cross-tabs showed Perry with 42% of the Hispanic vote.

At the same time that Perry was getting, supposedly, 13.9% of the Hispanic vote, the Velasquez Institute exit polls produced these results in the Senate race:

Hutchison 53.7%
Radnofsky 35.6%
Others 10.7%

So Hutchison ran FORTY points better than Perry among Hispanics? I don’t believe it.

Mike Baselice, Perry’s pollster, sent me the CNN/AP exit poll data for Texas. It was based on 2,090 respondents statewide. Here is what it showed:

Bell 41%
Perry 31%
Strayhorn 18%
Friedman 9%

This looks much more like it.

Is there any way to explain the discrepancy between WVI and CNN/AP? The only thing I can think of is that the precinct selection by WVI is somehow off the mark. If WVI is only looking at barrio precincts–a criticism that was raised in 04 when WVI had to raise its original lowball estimate of Hispanic support for Bush–then the Institute is missing the movement of more-affluent Hispanics to the suburbs. And those voters are more likely to vote for Perry than residents of urban neighborhoods.