November 29, 2016

Hillary Clinton won Latino vote but fell below 2012 support for Obama

Hillary Clinton won 66% of Latino voters on Election Day, according to updated National Election Pool exit poll data, a level of Democratic support similar to 2008, when 67% of Hispanics backed Barack Obama. However, Clinton’s share of the Latino vote was lower than in 2012, when 71% of Latinos voted to re-elect Obama.

While Clinton underperformed among Latinos compared with 2012, Republican Donald Trump won 28% of the Latino vote, a similar share to 2012, when Mitt Romney won 27%, and to 2008, when John McCain won 31%, according to exit polls. (It is important to note that the national exit poll is a survey with an overall margin of error of plus or minus 3 percentage points for the national result.)

On immigration issues, 68% of Hispanic voters opposed building a wall along the entire U.S. border with Mexico, compared with 46% of whites and 82% of blacks, according to NBC News exit polls. When asked about unauthorized immigrants, 78% of Hispanic voters said they should be offered a chance to apply for legal status, compared with 67% of whites and 82% of blacks. Overall, 46% of Hispanics cited the economy as the most important issue facing the country, followed by terrorism (20%), immigration (19%) and foreign policy (11%).

(For more analysis of the 2016 exit polls, see “Behind Trump’s victory: Divisions by race, gender, education” and “How the faithful voted: A preliminary 2016 analysis.” For an explanation of how exit polls are conducted, see “Just how does the general election exit poll work, anyway?“)

Some pre-election polls found levels of Latino support for Clinton that were similar to the ones that the national exit poll showed. However, most pre-election polls found lower Latino support for Trump than the exit polls did. For example, 19% of Latino registered voters backed Trump in Pew Research Center’s fall National Survey of Latinos.

In the days before Election Day, there was evidence of a possible historic surge in Latino voter turnout nationwide. Reports from Florida, Nevada and elsewhere showed strong early-voter turnout among Latinos. And the national exit poll suggests that Latinos did make up a larger share of voters in 2016 than previously: 11% this year, up from 10% in 2012 and 9% in 2008. Preliminary estimates show that slightly more votes were cast nationwide compared with 2012, leaving it unclear how many Latinos actually voted in 2016. (This year’s Latino voter turnout, which has historically trailed other groups, won’t be known until sometime in 2017 when the U.S. Census Bureau publishes its report on U.S. voting.)

Turnout aside, a record 27.3 million of Latinos were eligible to vote in 2016, up 4 million from four years ago – the largest increase of any racial or ethnic group. And the Latino electorate grew in many states since 2012, including the battlegrounds of Arizona, Florida and Nevada.

Note: This analysis was originally published Nov. 9 and has been updated. It reflects data for 2016 as published by and/or as of 12 p.m. on Nov. 29.

If data are subsequently re-weighted by the National Election Pool (NEP), the consortium of news organizations that conducts the exit polls, the numbers reported here may differ slightly from figures accessible through the websites of NEP member organizations.

Topics: U.S. Political Parties, Voting Issues, Voter Participation, Hispanic/Latino Vote, Voter Demographics, Hispanic/Latino Demographics, 2016 Election

  1. Photo of Jens Manuel Krogstad

    is a writer/editor focusing on Hispanics, immigration and demographics at Pew Research Center.

  2. Photo of Mark Hugo Lopez

    is director of Hispanic research at Pew Research Center.


  1. Martin Vega9 months ago

    I think it is problematic that a Latino research organization such as PEW Hispanic would validate Edison Research’s exit poll data in in comparing the levels of Latino support for Clinton vs. Obama, for the following reasons:
    • Edison Research (ER) is not fully transparent about how it derives the Hispanic data in the final survey results it reports to the National Election Pool (NEP):
    o It says it conducts a randomly stratified sample, and the methodology they use is based on a “recent” past election to identify precincts in order to select them for survey inclusion
    o What are some of the key issues in ER’s exit poll?
    o Does their strategy involve any stratification that includes high Hispanic density geographies?
    o As you (Mark) have indicated previously, since 2012 the new Latino voter is being driven largely by the demographic growth Latino Millennials. As such, a majority, if not all the individuals in this cohort might not show up in a “recent” past election; thus, Edison’s methodology would fail to capture this type of voter, who is largely Democratic
    o Edison says that it samples absentee/early voters shortly before the election and combines the results of these voters with Election Day interviews in its final data results. What they insufficiently elucidate is the following: How did they sample these early/absentee Hispanic voters; to what extent did they represent the Hispanic sample in the total sample; and, did they proportionally weigh these voters vis-a-vis their Election Day interviews to account for their growing importance. One important factor to note is that early and absentee voters comprised more than one-quarter of the total vote count in 2012 (26%). If, as has been rightly noted in the 2016 Presidential election, people of color disproportionately voted early, did ER factor any of this in their final tally?
    o If they did not make any requisite adjustments as outlined above, then, one can strongly assume that they significantly over-stated the percentage of the Latino vote that favored Trump
    o Lastly, most of the authoritative research conducted by organizations with solid Hispanic polling and survey methodologies showed that Latino support for Trump consistently ranged from 14% to 24% percent. Even Pew Hispanic’s own survey conducted a month before the election on October 12th, showed that only 19% of registered Hispanic voters favored Trump. Seen in this context, Edison’s Exit Polling numbers appear faulty, and PEW Hispanic does not do justice to itself validating such numbers, particularly, when its own internal data contradict these findings

  2. Anonymous9 months ago

    Latino Decisions is making the case that exit pool data is not the correct measurement.…

  3. Anonymous9 months ago

    Why is it that when minorities concentrate only on their own group it’s considered to be normal, but if a white man concentrates only on white males it’s considered to be racist and sexist? The answer might tell us one reason Trump was elected.

    1. Anonymous9 months ago

      Ouch, that’s downright hurt. But observe from another angle: Since Whites are relatively dominant, they’re supposed to think less about themselves.

  4. Anonymous9 months ago

    Why would anyone assume exit polling is accurate?
    The failures of almost all other polling and MSM pundits should automatically disregard it.

    1. Frank Barter9 months ago

      Good point. Were they all wrong? Or all right and something sinister is happening? Not to be an alarmist but really?

  5. Anonymous9 months ago

    The presentation with percents is interesting, but I think that total numbers of voters should be included as well. Percents give some comparative information, but that will not tell how many overall voters were contained in the different categories, which is also important. Maybe a percent went up or down between 2012 and 2016, but if the voter turnout was also different, then that percent change could be more or less impactful between the two elections.

  6. Deon Nungaray9 months ago

    Univision’s pathetic guidance on politics for the Latino population has been horrific. They need to get out of politics and are not doing Latinos any favors. That being said, they should have fully supported Romney in the prior election and Sanders in this election.

    1. Anonymous9 months ago

      You are 100% right

  7. Jim Kitchen9 months ago

    In reading this article, the election statistics seem quite different than the narrative we see in post-election headlines. The media is running with the narrative that this is a racially DIVISIVE election. However, this article and others (see Pew Research: Behind Trump’s victory: Divisions by race, gender, education) about the exit polls show us that compared to the election in 2012 minority votes were less fixated on one political party.

    If 7% more African-Americans voted for non-democrats and 6% more Latinos voted for non-democrats, wouldn’t that be a sign that race is less of a factor in 2016 than it was in 2012?

  8. Anonymous9 months ago

    I was not surprised by this. Most of these polls include Cuban-Americans, and there was a very good reason in their eyes why they voted for Donald Trump. They felt abandoned by the Democrats after President Obama moved to normalize relationships with Cuba, many feelings they had lost their place as special citizens who consider themselves political refugees from a horrible dictator. They no longer felt that one day they could return home to collect the wealth that was stolen from them. Most of us didn’t really think about this while doing our polling, but I would imagine, at least with Cuban-Americans, this was a very strong reason for them not supporting the Democratic candidate.

    1. Anonymous9 months ago

      Yay you are right