Numerous outlets have reported per the exit polls conducted by Edison Research that 29 percent of Latino voters cast a ballot for Donald Trump, suggesting he did slightly better than Mitt "self-deportation" Romney who landed at 27 percent in 2012. The same exit polls suggested Latino turnout in critical states such as Arizona, Colorado and Nevada was down. Sorry, but both of these conclusions are quite literally unbelievable.
First of all, they counter nearly every pre-election poll designed to conduct enough interviews with Latino voters to get a statistically significant sample. Here’s the level of Latino support for Trump reflected in several polls leading up to the election.
- 19 percent: Univision/Washington Post
- 17 percent: NBC/Telemundo oversample
- 14 percent: NALEO/Telemundo Tracking Poll
- 13 percent: FIU/New Latino Voice
Second, it's impossible to know how many Latino voters the exit polls even surveyed because they were clustered in specific precincts, which poses real hurdles to accurately sampling minority voters. As Nate Silver noted, Hillary Clinton may have underperformed pre-election mainstream polls in the Midwest, but she over-performed those polls in states with a heavy concentration of Latino voters like California and New Mexico.
Why does all this matter? Partly because it would be political malpractice for Democrats to ignore a constituency that likely played a large role in ensuring that Hillary Clinton soundly captured the popular vote. Latino Decisions’ election eve poll of 5,600 voters showed Clinton getting fully 79 percent of the Latino vote to Trump’s 18 percent. And partly because not accurately counting Latinos disenfranchises voters who voted like their lives depended on it. As America’s Voice writes:
This is not a mere academic exercise or a story simply about methodology. At a time when the Latino and immigrant communities are vulnerable and fearful given the election of Donald Trump, an accurate assessment of the community’s performance in the 2016 elections and its electoral power is of huge consequence. Selling the Latino vote short, as the national exit polls do, essentially diminishes – even disenfranchises – Latino voters at a moment of maximum peril.