Skip to main content
opinion

U.S. President Barack Obama speaks at a campaign event at Cornell College in Mt. Vernon, Iowa, on Oct. 17, 2012.Carolyn Kaster/The Associated Press

Is it possible that Barack Obama is doing much better in this election than the polls are suggesting? It appears that Americans may be missing a crucial and potentially decisive factor in their forthcoming election, because they conduct most of their polls only in English.

In this U.S. election, more than any before, the result is likely to hinge on the voting decisions of Spanish-speaking Americans. But because most of the major polling companies do not ask questions in Spanish, they are getting inaccurate results.

We know this from experience. In the 2010 Nevada election, Democratic candidate Harry Reid supposedly trailed Republican Sharon Angle by 3 percentage points in the polls. Yet Mr. Reid "surprisingly" won by 5 points. This 8 percentage point difference in projections, it turned out, was close to the difference between results from English-only polling versus polls taken by one of the few companies doing their surveys in both Spanish and English – whose projections correctly called that election.

And we know this from the one major study done of Hispanic polling results – a May 2001 paper in the Journal of Hispanic Behavioral Science looking at polling in a variety southern Florida races. The study concluded: "had the polls been conducted in English only, results would have been profoundly inaccurate and invalid."

This should have been a lesson learned by U.S. pollsters, but looking at the polls and projections for the 2012 race, that doesn't seem to be the case. As such, one has to take current polling results showing a close race – especially in certain Latino-heavy swing states – with a few grains of salt.

Latinos are now an important and potentially decisive factor in winning an increasing number of states. The Latino population has tripled over the past 25 years to 23 million. Every month 50,000 to 60,000 American-born Latinos turn 18, automatically becoming eligible to register to vote. On the state level, Latinos are already a significant portion of the electorate in New Mexico (38 per cent), California and Texas (25 per cent), Florida (17 per cent), New York (13 per cent) and New Jersey (10 per cent).

When you look projections for key swing states from the only company to poll in English and Spanish, the numbers get more interesting.

LatinoDecisions was the company that correctly projected the 2010 Nevada election result – and it did so by giving the voters it called a choice of language. What was missed in the English-only polling was that Latinos were enthused about voting and over 80 per cent of them were voting for Mr. Reid.

Is it possible that similar shifts are being missed in Latino-heavy swing states such as Florida? We know that something similar to the 2010 Nevada phenomenon is taking place this year in Arizona. According to LatinoDecisions polling, the majority of Latinos in that state are extremely enthusiastic in support of Mr. Obama and an even higher number are more enthused about this election then the previous one – in large part because they are angry about Arizona's infamous anti-immigrant state law, passed by its Republican government.

Will Mr. Obama win Arizona? Probably not, but it's going to be a lot closer than most polls indicate. And, looking at the growth of the Latino population, if the Democrat does not win this year in Arizona it is likely that, as with Nevada, the state will still move from being a GOP bastion to become a swing state – and maybe even to a Democratic majority.

All of this makes two bigger points about the impact of Latino voters.

First, we are missing a lot of what is happening because of English-only polling. And second, we are asking the wrong questions. Everyone is looking for a simple story and single, decisive indication of importance. Instead what we are seeing in Nevada and now Arizona appears to be a state-by-state, election-by-election transformation, not a sweeping national movement, as individual races, local issues and candidates galvanize Latinos.

This will be a particular challenge for the Republicans, whose traditional white Protestant working-class base is being chiselled away by the growing Latino voter base and related demographic shifts.

Texas shows the limits of current GOP strategy. The state party has done a remarkable job of keeping Hispanic turnout low – part of a deliberate strategy of not backing any issue that would inflame Latino voters, and also discouraging Latinos from coming out to vote unless the GOP was running a strong Latino candidate.

That strategy has begun to fail in Texas as more extreme Tea Party elements of the party have pushed tougher voter ID measures and cut state spending for education, both moves that have sent Latinos to the ballot box. Texas certainly is not going to go Democrat this election – but the seeds may have been sown for a change in the near future. Because of this large discrepancy in polling, it may be much sooner than we think. And if Texas goes, so goes the GOP as a party capable of winning the White House.

Carlo Dade is a senior fellow at the School of International Development and Global Studies at the University of Ottawa and a former executive director of the Foundation for the Americas.

Interact with The Globe