california recall

Did One Bad Poll Help Gavin Newsom Mobilize Recall Opponents?

Things are now beginning to look up for Gavin Newsom. Photo: Justin Sullivan/Getty Images

The Public Policy Institute of California, a well-regarded public opinion outfit with a tangible Republican lean in their polling history, has just released its final survey on the Newsom recall election in which voting will end on September 14. The poll shows the effort to recall the governor trailing among likely voters by a hefty 39-58 margin.

Given all the recent talk about Democrats potentially losing the governorship of this very blue state, the PPIC survey shows that momentum has swung back to Newsom. You can already envision the articles on how Newsom benefitted from recent disclosures about front-running replacement candidate Larry Elder, or from national Democratic efforts to fire up the base via ads featuring Elizabeth Warren and Bernie Sanders. And maybe that’s all correct.

Or maybe it’s at least partially an illusion.

PPIC’s latest numbers are virtually identical to how they showed the contest in March (40 percent yes on the recall, 56 percent no) and May (40 percent yes, 57 percent no). And if you look at the 14 public polls of the recall election in the RealClearPolitics database, only one shows Newsom going down: a SurveyUSA poll for KABC released in early August. Indeed, it showed “yes” on the recall leading “no” by a shocking 11-point margin (51-40). It’s not an exaggeration to say this poll set off alarm bells throughout California politics, exciting Republicans who have wanted to smite Newsom from the day he was inaugurated and helping Democrats convince voters the recall wasn’t some nuisance effort they could safely ignore. The same poll raised eyebrows by showing the leader in the replacement contest was little-known Democrat Kevin Paffrath, a minor YouTube phenomenon as an advisor on real estate investment and management. This suggested that perhaps Democrats were not obeying party instructions to skip the replacement ballot question entirely.

But then SurveyUSA released another poll for KABC at the end of August that showed “no” leading “yes” by a robust 43-51 margin, a 19-point swing in just a few weeks. Was this proof Newsom was “surging” or replacement candidates like Elder were swooning? Possibly, but SurveyUSA, to its credit, offered a clear methodological explanation:

SurveyUSA subtly changed the language it used to identify a likely voter. In previous polling, voters were asked how certain they were to vote or not vote “in the recall election to remove the Governor,” wording which may have led some respondents opposed to the recall to misidentify themselves as being unlikely to vote. In today’s release, voters are instead asked how certain they are to vote or not vote “in the recall election.”

I don’t think the misleading nature of the earlier word choice was all that subtle. Another methodological change evaporated the startling lead for Kevin Paffrath in the replacement race:

[In the later poll] SurveyUSA inserted an additional question, before the list of candidates, to ask if voters would leave the second question blank, better mirroring, in SurveyUSA’s judgment, the way those voting “No” on the recall would approach their ballots. Today, in the newly added question, 28% of voters initially tell SurveyUSA they will leave the 2nd question blank; another 9% then tell SurveyUSA they will leave the question blank after being shown all 46 names on the ballot.

2nd, SurveyUSA today names all 46 candidates; in previous polling, SurveyUSA named 7 viable candidates only, including a single Democrat, Paffrath. When respondents today are shown the full list of candidates, which includes 9 Democrats, Paffrath’s support reverts to what may be its truer level, 6%.

So much for Governor Paffrath, who fell into the pack of 41 other candidates on the ballot behind the top five in the final PPIC survey.

SurveyUSA also mentioned two other non-polling factors that might help explain a better performance for recall opponents down the stretch: the galvanizing effect on Democrats of actually receiving a ballot in the mail (they were sent to all 22 million registered voters, just the second time the state has done that), and a common aversion to confusing ballot initiatives that often drags down support for them in the end.

Now, it’s possible there really was a change of momentum in the contest recently, not just an illusion based on a single faulty poll. A Berkeley IGS/L.A. Times survey in mid-July showed “no” leading “yes” by a spare three percent; an early May poll from the same outlet had “no” up by 13 points.

But more than likely Newsom has been in the lead, if not entirely in the clear, the whole way, with the “no” campaign’s heavy spending (along with media scrutiny of Larry Elder) countering any late burst of Republican excitement. By all accounts Elder does still lead the replacement field, but with that and six bucks you can get a pretty good latte at California’s expensive coffee shops.

You probably shouldn’t spend much time polishing obituaries for Gavin Newsom’s political career, unless two weeks is enough time for Democrats to become highly complacent.

Did a Bad Poll Help Gavin Newsom Mobilize Recall Opponents?