tags: , , , , Blog

Latino Decision's Matt Barreto Explains How Polls Underestimate Latino Voters

Share This:

Over the weekend, The New York Time’s polling guru, Nate Silver, asked whether polls are underestimating the effect of the Hispanic vote.  The answer appears to be yes.  Silver spoke with Matt Barreto of Latino Decisions, who looked at some of the polling that has been conducted this year and explained how this underestimation happens.  As Silver wrote, it “may be because many polling firms that conduct interviews only in English miss some Hispanic voters who are more comfortable speaking Spanish.”  And because “primarily Spanish-speaking Hispanic voters are more likely to vote Democratic than those who have more English fluency,” polls that under-count Hispanic voters are more likely to mistakenly project in favor of Republicans.

Today, Professor Barreto was at George Mason University in Arlington, Northern Virginia, to discuss new polling of Latino voters in that state. America’s Voice had a conversation with him about Latinos and polling in 2012. Matt explained that “pollsters have not been keeping up with the growth of the Latino market, especially the immigrant and Spanish-speaking market.” He pointed out that many pollsters don’t do any Spanish-language polling, or introduce flaws into their methodology by not being ready to conduct Spanish-language polling.  That means up to 40% of the Latino community is excluded from traditional polling–one reason why pollsters were so wrong about the 2010 Nevada Senate race, when many of them mistakenly believed Sharron Angle would win because they weren’t counting enough Latino voters.

Barreto also mentioned a couple states which he thinks might be “this year’s Nevada,” including Senate races in Nevada, Arizona, Massachusetts, Connecticut and Virginia.

Ultimately, Barreto warns his fellow pollsters that they “need to update their methods, otherwise they run the risk of missing the Latino vote.”

If you’re interested in the 2012 elections, it’s worth a watch: