Political junkies, pollsters and both campaigns live and die by poll numbers, especially exit polls. With the election only five days away, Latino political scientist Gary Segura, from Stanford University, issued a warning: beware of exit polls, when it comes to Latino voters.
“We again want to caution the press and the public,” writes Segura in a Latino Decisions post. While the National Exit Polls are good at calling overall, two-party state results, particularly about the voting patterns of white, non-Hispanics, Segura says “they do far less well in accurately capturing the vote of Latinos,” he states. This is due to a number of reasons, Segura says. One is that the precincts measured in these exit polls are not enough to accurately represent Hispanic voter numbers.
There is another important factor in why exit polls do not accurately survey Latino voters, and that is language. Over 30 percent of Hispanic voters prefer to be interviewed in Spanish, a not insignificant number, and in fact, only 22 percent of Latino voters in a recent Latino Decisions poll never watch Spanish-language television. Yet Segura explains that very few exit polls do; in 2004, less than 5 percent of the national exit poll data conducted interviews in Spanish. “Exit polls are under-surveying Spanish-dominant Latino citizens by a huge factor, introducing a systematic bias toward more assimilated, higher income, better educated voters,” states Segura.
Latino and other minority voter are also misrepresented when it comes to socio-economic and other characteristics. National exit polls have 50 percent more minorities with college degrees than the reality, and they show minority voters making 5 percent higher income than they do, representing an error rate of 11 percent. “In both cases, income and education levels are shifted toward higher status, less Democratic, minority voters,” says Segura. States’ exit polls can show even greater discrepancy. Colorado, which has a growing number of Latino voters, has a 30-point gap between the share of exit poll respondents with a college degree and the actual numbers, based on the state’s Census figures.
Putting all this together, Segura estimates the National Exit Poll (NEP) data can be anywhere between 11 and 24 points away from the actual votes cast by Latinos for the 2012 election.
What about early voting, how can exit polls capture that? In half of the battleground states – North Carolina, Colorado, Florida, Nevada and possibly Iowa, half the voters are voting before Election Day. To factor in early voting, pollsters are supplementing exit polling with phone surveys. When it comes to Latino voters, the issue is that many national phone surveys do not always use a truly representative sample of Hispanic voters. In the case of Latino Decisions, who does survey a representative cross-section of U.S. Hispanics, their latest tracking poll showed that as of last week, 8 percent of Latinos had participated in early voting.
In this close election and as the Latino vote grows in size and therefore influence, the issue than is to ensure that the polling, whether it be in person or over the phone, gets the Hispanic vote right.