David Goldman / AP
Thursday, Feb. 11, 2016 | 2 a.m.
After Donald Trump won New Hampshire’s Republican primary Tuesday night and Vermont Sen. Bernie Sanders handily swept the Democratic one, the national gaze turned to Nevada.
The state’s caucuses are nine days away for the Democrats and 12 for the Republicans, who will square off Feb. 20 in South Carolina before coming to Nevada. What happens here could shape the trajectory of the campaigns. But beyond pundit projections and conventional political wisdom, it's hard to gauge how Nevada might alter the race — at least due in part to the limited number of polls conducted in the state.
Unlike the run-up to Iowa and New Hampshire, where there were 70 to 100 polls apiece before the states' respective caucuses and primaries, there were only six polls last year and none this year in Nevada.
The most recent Nevada poll by Gravis Marketing, conducted in late December, showed Trump in the lead, ahead of Texas Sen. Ted Cruz by 13 points. On the Democratic side, the same poll showed former Secretary of State Hillary Clinton ahead of Sanders by 23 points.
The first reason pollsters offer for the dearth of polling — and the conventional explanation — is that what happens in Nevada depends on what happens in Iowa and New Hampshire, so it often doesn’t make sense to make predictions before the contests in those states are finished. Gravis, which conducted four of the five Nevada polls last year, plans to poll again between now and the Nevada caucuses.
Still, there’s been a comparative abundance of polls in South Carolina — 31 for the Republicans and 20 for the Democrats. Pollsters say there are likely to be more Nevada numbers in the days before the caucus, but the difficulties of polling here in the state make the extent of that polling uncertain.
Mason-Dixon Polling and Research, which has polled in Nevada in the past, has no plans to poll in Nevada before the caucus. (The firm has conducted media polling exclusively for the Las Vegas Review-Journal since 1990.)
Brad Coker, managing director of the firm, pointed to the difficulty and high costs of polling in caucus states compared to primary states as a possible reason for the shortage of polls this year.
A specific challenge in caucus states relates to the format of the events, said J. Ann Seltzer, whose firm has conducted the Des Moines Register’s Iowa Poll since 1997. In the Nevada caucuses, as in the Iowa ones, supporters of each candidate can stand up and give a brief pitch for their candidate before the actual voting happens, which can prompt voters to make last-second changes of heart. In primary states, voters go to the polls armed with all the information they'll have to make their decision, and are less likely to shift their allegiances. Primary voting is less volatile, in other words, meaning voters are more likely to stick with the choices they make in polls.
“Things happen in caucus rooms that change the vote,” Seltzer said. “It’s one of many ways that caucuses are hard.”
To compensate for that unpredictability, Seltzer’s firm polls respondents’ second preference candidates. What they found this year was that while Trump was many people’s first-choice candidate, he was almost no one’s second choice.
“One of the things we knew going in is if you weren’t already a Trump supporter, nobody liked you,” Seltzer said. “With Trump, there was only likely to be outflow with no compensating with inflow.”
Turnout has also been unpredictable since Nevada won its status as an early-caucus state. In 2008, about 118,000 Democrats turned out to caucus compared to 44,000 Republicans. Four years later, even fewer Republicans, 33,000, turned out. Democrats did not have a competitive caucus that year.
By contrast, Iowa saw 180,000 Republicans and 170,000 Democrats turn out this year. (Iowa’s population is about 3.1 million compared to Nevada’s 2.8 million.)
“You look at it in Iowa and you say, ‘You can’t possibly poll this,’” Seltzer said. “Nevada is even worse.”
In Nevada's first year as an early-caucus state, 2008, there were 18 Republican polls and 20 Democratic ones. Four years later, there were 11 Republican polls. Seltzer said it was possible that pollsters were willing to take a chance on Nevada in 2008, but decided to take a pass this time around due to the unpredictable turnout and the difficulties of polling caucuses.
Beyond the numbers, it’s also difficult for pollsters to guess the demographics of those who turn out. Not only does Nevada have a diverse population, but a transitory one, too. Young voters, or those who are new to the state, may not make it through the screening process for likely voters, said David Damore, UNLV political science professor, adding that polls in Nevada tend to skew older and more Republican.
“It’s a very challenging environment for Nevada,” Damore said. “It’s a microcosm of all the issues that are facing the polling industry.”
Two factors that increase the cost of polling are needing to conduct polls in both English and Spanish, and polling both over landlines and cellphones. Damore, also a senior analyst at the Latino political opinion research firm Latino Decisions, said one of the group's polls typically would cost about $30,000 to sample 400 to 500 people.
Not only is polling in the state expensive, but it’s often difficult to get someone to find someone to front the money. The well-known Field Poll in California, for instance, relies on financial support from media outlets, the University of California and California State University systems and other nonprofits. In South Carolina, Clemson University, Winthrop University, Monmouth University and Marist College have all polled voters ahead of that state’s upcoming primary.
“You don’t have a university-affiliated constant poll like the Field Poll built in here,” Damore said. “You’re relying on whoever wants to pay for it.”
Until the next polls come out, it’s a waiting game for those trying to predict what will happen in Nevada. The Democratic caucuses will be Saturday, Feb. 20, with the Republican caucuses on Tuesday, Feb. 23.