He put that proposition to the test in PPP’s next poll, of Arizona Republicans. It turned out that Jensen had been mistaken. Snyder was viewed favorably by 5 percent of respondents; Egan, the reporter, actually outpolled him with 6 percent support.
Farris, who teaches an upper-level class on polling, was puzzled by the inclusion of an unfamiliar name. PPP explained its logic. “Well, next time throw my name in the poll as someone with no national profile,” she tweeted to PPP. “It'll amuse my Survey Research students.” And then, on Thursday, she spotted reporters trying to figure out why a quarter of Republican voters had opinions about “Emily Farris,” when they’d never heard of her.
It’s well-established that voters will volunteer views concerning wholly fictitious issues or candidates. One landmark study asking voters for their opinions on the Public Affairs Act of 1975 found a third of them willing to voice opinions on the non-existent statute. Subsequent surveys have replicated that result. Voters can be reluctant to confess ignorance, or conflate invented laws or candidates with real ones.
The reaction to Farris, then, is a useful reminder that polls this far in advance of an election have little to no predictive value. That may be particularly true this year, with a GOP field so crowded that automated pollsters like PPP can only include half of the candidates when they ask horserace questions. It’s easy to forgive voters who failed to spot the ringers in that enormous lineup. Few are paying much attention. And so their responses to the actual candidates should be taken with an equally large grain of salt.
But if it’s understandable that voters offered opinions, that still doesn’t explain why they were so uniformly negative. Voters were evenly split on the Public Affairs Act, and on Egan, the reporter included in the earlier PPP survey. So why did they so dislike Emily Farris?
Perhaps they take issue with her scholarship on social capital and political participation? Maybe her sky-high unfavorables were just a statistical fluke? Her name itself seems anodyne: Emily is one of the most common first names in the country, and Farris ranks high among surnames.
In fact, about the only thing her name reveals is her gender. With Hillary Clinton leading the race for the Democratic nod, it’s striking to see a generic female name generate such strongly negative views in a poll of Republican voters. Might gender have played a role? Farris was appropriately cautious. “We'd have to see this as a pattern across multiple polls,” she said. “We never want to draw too many conclusions from just a single poll.”
She’s in luck. "I do think it's something we need to test more,” Jensen told me. The next time PPP puts a poll into the field, it may include two generic names—one identifiably male, the other female—to see if it can replicate its results.
As for Farris, it turns out that her poor polling is the least of the hurdles standing in the way of a presidential bid. She’s only 31, not yet old enough to run. But she’ll be eligible by 2020. Would she give it another try then? “I don't have any aspirations to yet be president,” she protested. “I'm happy commenting on the news in politics, not being a participant.” And when classes reconvene in the fall, she’ll have a new case study to teach to her students.