By John Zogby
There have been three national polls since the first Presidential debate. Two polls (one for Investors Business Daily by TIPP and the other by John Zogby Strategies-EMI Research Solutions) show a 3- and 2- point lead respectively by the former VP. The other by the Wall Street Journal and NBC with a 14-point lead by Mr. Biden. Let’s get the obvious out of the way. TIPP is an excellent polling outfit, one of the most accurate covering Presidential races over the past decade. The WSJ/NBC has been around a long time and is truly bipartisan – a joint project by prominent Democratic and Republican pollsters. And my Zogby polls have an unbroken record of near-perfect accuracy from 1996-2012. (I sold my company and was not directly involved in 2016).
First, the Zogby Strategies-EMI Research Solutions Poll of 1006 likely voters nationwide was the only poll taken start-to-finish after the President entered Walter Reed Hospital. (October 2). The others were taken after the Tuesday night debate but before any report of the President’s positive Covid-19 testing. It is hard to definitively address the impact of the President’s test, but it is an additional variable in our polling as opposed to that of our colleagues.
Second, is the issue of sampling. I am a lifelong Democrat and have mainly voted that way over the decades. But I have also felt that many of our colleagues have oversampled Democrats. In the WSJ/NBC Poll, 45% of those polled identified as Democrat and 36% as Republican. Even with a Clinton victory in the 2016 popular vote, the exit polling showed 38% saying they were Democrats or leaning that way, while 34% were Republicans. A nine-point difference between the current poll and the 2016 result. TIPP has a 40%-40% partisan split, while the Zogby Strategies has stuck with the 2016 model of 38% Democrat to 34% Republican. Given that our polling shows Mr. Biden leading 91%-8% among Democrats and Mr. Trump has climbed to a 94%-6% lead among fellow Republicans, it looks like partisanship and voting intent really matter.
Just for the record, all of us pollsters use what we term a “turnout model” – i.e. results that reflect what the actual demographics of Election Day turnout will look like. Of course, this year we have a strange situation with so many opting for mail-in ballots, but the point is that we all engage in weighting of some variables. In the Zogby Strategies poll, we project a solid turnout of Hispanics (13% of the total vote) and Blacks (12% of the total vote) – both healthy figures for Mr. Biden. He receives 61% of the Hispanic vote and is up to 87% of the Black vote – both solid but not quite enough just yet.
With 51% of our turnout model as women voters, Mr. Biden’s 56%-41% lead is very significant. For the President, it looks like he has consolidated his winning base – at least for now. He is back in the lead 50-%-47% among voters 65 and older – as opposed to the 10-point deficit we saw back in late August 29. He also leads among men 53%-42% and whites 56%-40%, as well as among Evangelicals (68%-32%), Conservatives (78%-18%) and Very Conservatives (92%-8%).
Finally, the TIPP Poll and the Zogby Strategies Poll included only “likely voters” as opposed to the WSJ/NBC Poll of “registered voters”, which can also explain the Democratic lean in their results.
Which poll is right? All three of us – and several others out there – have great track records. For now, I think it is mainly about timing before and after the President’s hospitalization. It is not a time to make any projections about the outcome except that this will seesaw back and forth as the polling always seems to do.