New Analysis Shows Gallup's Assumptions About 2010 Latino Vote To Be Very Flawed

This morning Simon wrote a new analysis, based on the work of Professor Alan Abramowitz, on the accuracy of  Gallups polling going into the midterm elections.

In Simon's analysis, which can be read here, he argued that the sample size was flawed and tainted the overall accuracy of Gallups polling. In particular he noted that it is highly unlikely that the Hispanic vote is trending Republican. Simon's quote is below:

I think it makes the case, persuasively, that the very Republican heavy Gallup sample and results are deeply flawed, and need to be discounted this year.  

If nothing else the suggestion that the Hispanic vote could end up being plus Republican this year makes the poll a wild outlier.  The Hispanic vote has gone 2:1 Democratic the last two elections.  In the last two major polls of Hispanics this year the numbers came back 3:1 and 2.5:1 Democratic.   Given how anti-immigrant the national Republican Party has been this year, trumping even their over the top rhetoric and proposals of previous years, the idea that there has been a huge swing of Latino voters to the Rs is not plausible.

Lets look at the numbers from recent polls of Hispanics. A recently released Pew Hispanic Center Poll  actually has the Republican position eroding from already weak 67%/31% Obama/McCain numbers:

Two-thirds (65%) of Latino registered voters say they plan to support the Democratic candidate in their local congressional district, while just 22% support the Republican candidate, according to a nationwide survey of Latinos.

A recent Latino Decisions Poll, shows a very similar story:

With little more than two weeks to go until the November election, the most apparent characteristic of this week’s track is stability. Latinos prefer Democratic candidates to Republican candidates, 59.0% to 22.3%, with 18.7% undecided.

To be clear each of these polls shows the GOP Hispanic number to be at 22%, or 9% less than the GOP performance in 2008.   This of course coming in a year when the GOP has substantially improved its position with the overall American electorate.  Given the virulent anti-immigrant stance of the GOP it is simply implausible to have anticipated any kind of GOP gains with Latinos this year. These two polls clearly capture the eroding position of an already weakened GOP with Latinos.  Though Gallup will argue that it is using a tighter likely voter screen which could result in a different result, it is just implausible given these trends lines and basic common sense to construct a scenerio where Latinos have swung back to the GOP this cycle.  These results of course draw into question the entire Gallup likely voter sample.