With less than one week to go until the Nevada caucuses, it’s not clear whether any of the eight remaining Democratic presidential candidates has a distinct advantage in the state, because Nevada — referred to as “the stepchild among the first four [Democratic primary] states,” by Nevada Independent editor Jon Ralston — gets shockingly little attention from pollsters.
The reason? Although the caucuses — which will take place on February 22 this year — have been the third Democratic primary contest in the nation since 2008, the state is notoriously hard to poll.
We now have exactly one recent poll of the state, courtesy of the Las Vegas Review-Journal, which shows Vermont Sen. Bernie Sanders in the lead among likely caucus-goers with 25 percent support, trailed by former Vice President Joe Biden with 18 percent, and Massachusetts Sen. Elizabeth Warren with 13 percent. The poll was taken February 11 to 13, and has a margin of error of 4.8 percent.
These results should not be taken as absolute fact, however. Not only are Nevada polls challenging to do, they’re challenging to get right. In 2010, for example, polls suggested that then-Senate Majority Leader Harry Reid was on course to lose reelection. Instead, he cruised to a nearly 6-point victory.
Ralston, who founded the nonprofit Nevada Independent in 2017, has minced no words when it comes to polling the state. In 2012, he wrote a piece titled “Why most polls done in Nevada are garbage,” and eight years later, he says the state’s polling problems have not been resolved.
“Nevada is a quirky state,” he told Vox. “And the new caucus rules, with early voting, combined with the impact of the Iowa disaster ... make polling very, very difficult.”
Changes to caucusing process has made polling more difficult
The Iowa caucuses are a well-established tradition dating back to the 1970s, but Nevada’s caucuses are relatively new. The relative newness of the caucuses means that pollsters have not had time to build the sort of infrastructure and state expertise that was developed over decades in Iowa.
For a pollster to provide an accurate snapshot of the state, Ralston points out that they need to “know how to weight the results to fit the picture that will exist on Election Day — that is, what the turnout actually will look like.”
As Vox’s Matt Yglesias has explained, pollsters typically give certain responses more weight than others in order to make sure their sample accurately reflects the likely electorate. But given the dearth of historical caucus day data, developing formulas to accurately weigh different responses becomes nearly impossible.
Further increasing the difficulty of developing turnout models is the fact that, as Ralston mentioned, the caucus rules will be different this year.
As Vox’s Andrew Prokop writes, Nevada’s caucus rules, like Iowa’s before it, have been changed to provide greater transparency. State party leaders will report three sets of caucus numbers: a pre-realignment vote total, the post-realignment vote total, and the delegate distribution. As in Iowa, caucus-goers whose first choice candidate does not receive at least 15 percent support will be asked to realign their support to a candidate who has met that threshold.
The state also has a newly instituted four-day early voting period that attempts to model caucusing by using a system modeled on ranked-choice voting.
The new rules mean pollsters are faced with new questions when it comes to predicting turnout: Will the early voting period increase participation by making the caucus process more accessible? Will the new caucus day rules — and the confusion similar rules caused in Iowa — lower enthusiasm among potential caucus-goers and reduce turnout?
For now, these are questions that can’t be answered until the caucuses are over. The answers the caucuses reveal will allow pollsters to build more accurate models for the 2024 election cycle. But even once these unknowns are solved for, pollsters in the state will still have to overcome another challenge that has made polling difficult: the uniqueness of Nevada’s caucus-going population.
Nevada’s workforce means normal best-polling practices aren’t effective
Again, concerns like those outlined above can be overcome, as Iowa’s polling has proved. But Nevada’s voting-age population has particular characteristics that lead to additional polling challenges.
It’s important to keep in mind that accurate polling requires surveying a representative cross section of a given population, as Vox’s Matt Yglesias has explained:
The most basic idea of polling is that you can get a pretty good idea of what a population of several million people thinks by asking a sample of just a few hundred of them.
The trick is that for this to work, you want a random sample of the state’s population. If you sample a few hundred people coming out of an exurban megachurch, you’re going to get a sample that’s quite biased toward Republicans. If you sample a few hundred college students, you’ll get a sample that’s quite biased toward Democrats. Traditional telephone polls avoid this by calling people at random.
Usually a representative sample of the population can be polled through this method, and any undercounted groups can be corrected for by the weighting process Ralston outlined. However, random phone calls don’t work well in Nevada.
Ralston says that is because the state has a “fluctuating population,” in large part due to its heavy reliance on the tourism and casino industries. Nevada’s population can be transient, and can wax and wane seasonally. Telephone polls become difficult to conduct in these conditions because as FiveThirtyEight points out, a sizable segment of likely voters does not have local cell phone numbers, numbers in public records may belong to those no longer living in the state, and newer residents are less likely to be registered — or even eligible — to vote.
And those who work in Nevada’s entertainment industry — there are around 60,000 members in the state’s Culinary Workers Union alone — can be difficult to reach by telephone due to their hours. Many jobs in casinos, hotels, or restaurants are filled in shifts over a 24-hour business day instead of a more traditional 9-to-5 schedule.
In 2016, Mark Mellman, a pollster who’s worked with Reid — the top Senate Democrat until his 2017 retirement — told FiveThirtyEight’s Clare Malone that failing to contact these workers, particularly those who belong to unions that encourage caucusing, can badly skew polling results.
“If you’re only polling at night,” Mellman said, “you’re missing a fair share of electorate ... in the casino and entertainment industry that’s highly unionized, that’s very Democratic.”
The need for pollsters to have staff conducting field work over a 24-hour period rather than in evenings like in other states, as well as the costs involved in keeping up-to-date phone records and developing new models, means polling in Nevada is substantially more expensive to produce than in other states.
Patrick Murray, the director of the Monmouth University Polling Institute, told the Las Vegas Sun in August that polling the state could be twice as expensive as elsewhere because of the difficulty of contacting voters alone.
Without the body of reliable polling that has been available in the contests so far, it is hard to say who is truly ahead in Nevada. Polls or not though, the caucuses are coming: Early voting began Saturday.
Correction: An earlier version of this story misstated Harry Reid’s Senate title in 2010. He was majority leader.