Allow me to revise and extend my remarks on polling skepticism in our annus horribilis by citing … another poll. Yes, yes, I get the irony of this, but yesterday’s AP/Ipsos poll points out the dichotomy facing election polling in every cycle — but especially in this one. And that dichotomy is how to calculate preferences against enthusiasm in identifying just which voters are likely to participate at all, especially in the middle of a pandemic.

The toplines of the poll are hardly cheery for Donald Trump or Republicans. At the moment, 41% of respondents plan to vote for Joe Biden, while only 30% plan to vote for Trump — which might be a new low in presidential polling. The leaners break roughly evenly, adding up to a 46/34 advantage for Biden. Eight percent don’t plan to vote at all — a point to which we’ll return in a moment — and another 10% think they’ll go with a third-party candidate in November.

Trump’s approval ratings, both overall and on issues, are dismal in this series. He peaked this year in March at 43/56 at the beginning of the COVID-19 pandemic, but has slid ever since. In this iteration, Trump gets a 38/61, which is notably above his share of the vote in the same sample. Trump is now underwater on the economy (48/51) after peaking at 56/41 in late March. He only gets 32/68 for his handling of the pandemic, 36/63 on health care overall, and 36/63 on education. Trump tends to do poorly in this series anyway, but these numbers look especially bad.

However, those are only the toplines. The Hill got access to the crosstabs on a question about enthusiasm for the upcoming election, and that changes the picture quite a bit:

Supporters of President Trump are more enthusiastic about their candidate than those who back presumptive Democratic presidential nominee Joe Biden, according to a new poll.

In The Associated Press-NORC Center for Public Affairs Research survey, 42 percent of Trump supporters polled described themselves as excited heading into the election, while 31 percent of Biden supporters said the same.

Nearly three-quarters of Biden’s supporters, 72 percent, also described themselves as anxious about the election, compared to 52 percent of Trump backers who said the same.

And 65 percent of Biden supporters surveyed said they are frustrated about the upcoming election, compared to 45 percent of Trump supporters.

So what does that mean? It strongly suggests that Trump will get his voters to the polls in larger numbers, as well as get volunteers out in more strength to help his campaign. If true, that would have significant down-ticket impacts. But what would they be, and would they be enough to overcome a twelve-point lead — likely narrower in the battleground states that truly matter?

Those questions exist in every cycle, but they are especially acute in 2020. Not only can’t we identify likely voters by enthusiasm, we don’t know who’s likely to overcome all of the obstacles in the COVID-19 pandemic to casting ballots at all. Biden and the Democrats are pushing hard on mail-in balloting, while Trump is actively discouraging it, which might end up cutting against the enthusiasm. On the other hand, states that decide mass-mailed ballots are too problematic might end up giving Trump the edge based on willingness to come to polling precincts.

Add all that up, and you get … utter uncertainty. And that’s true even if one concedes that the polling itself was accurate in either aspect. It’s a fool’s errand to consider polling as a predictive indicator of anything related to elections at the moment. These are better used as temperature-taking of the moment rather than harbingers of doom.

If Trump can’t boost his approval ratings on issues, especially COVID-19 and the economy, then he will be doomed. Let’s wait until after Labor Day and more vision on what voting processes will be before reaching that conclusion.