Moby Dick is a useless guide to polling, vol 3.

ONCE MORE, WITH FEELING

The hour grows late in this election season, and we continue with this series on why Herman Melville was right that we should not stare into polls for too long, lest we lose ourselves to phantasms and grim fancies, but totally failed to tell us what in fact we should do with polling. You can read my previous missives here and here, and also Jason Stanford is taking Melville’s advice and you can read why in a good piece here.

Here is the quickest of recaps: when confronted with a poll, we must ask ourselves who is moving (as in, changing their expressed opinion) and also why. Polls are optimized for the first, not for the second, but most media coverage either pretends that they are or just declines to deal with that problem entirely. The result is confusion and disorder, and by god we can’t have that.

This man will not help you understand polls. Do not ask him, for he is mad.

I’m going to speak in broad terms here: most voters know who they will vote for before their party chooses a candidate. The single most powerful heuristic in politics is party identification (this is true at every level, though it can manifest in different ways at state & local versus congressional versus presidential). Sometimes, a party will screw up so outrageously that habitual party voters will cross party lines to get rid of them; that’s how you get Obama 2008, among other landslides. Sometimes a candidate will be so obviously limited that a slice of that candidate’s own party doesn’t turn up to vote at all (this, I believe, is what would have happened if Joe Biden had stayed on the ballot after the debate).

For the most part, though, that party-affiliation heuristic immediately and drastically winnows down the universe of voters who will make up or change their minds at any point during the election. The remainder are the voters candidates fight over. Swing voters; undecideds.

What we need to understand about swing voters is this: in a normal election, the number of voters who are swing voters in the sense that they might vote for one major party candidate or the other major party candidate is extremely small. Again, I’m speaking in general terms and (as is my habit) violently compressing a lot of science here: by this time in the election cycle, when a voter tells you they’re undecided, what they mean is that they have not decided if they’re motivated to vote yet at all, not that they have no idea who they would vote for if they did.

We have an existing permission structure for voters to take a long time deliberating over who they will vote for, but no such thing for not voting at all. So “I haven’t decided yet” or “I need to learn more about them” acts as the socially-acceptable version of “I am not yet motivated enough to declare an affiliation, much less vote.”

A certain stripe of liberal, in my experience, spends a fair bit of time fulminating about undecided voters - “how could you possibly not be able to tell the difference between Kamala Harris and Donald Fucking Trump?” This, with its implied “how could anyone be so stupid” is wasted energy and smacks of a particular kind of righteous onanism that does the political left no credit at all. If a poll presents you with undecideds breaking in one direction or another, you’re not seeing people who just couldn’t decide between Kamala Harris and Donald Trump finally making a call. You’re seeing them declare that they have been engaged, been activated - by the candidates, or by the political environment - enough to at least declare an affiliation, but they’ve known for a while how they feel about at least one of those candidates.

Actual undecided voters in the sense that we think of them are generally pretty rare. There are always some, and their number is significantly greater with candidates who are not as well known.

This election, of course, features the best-known presidential candidate in American history. Trump’s very nature is totalizing, or at least striving for that effect; there’s a damned good chance that this whole political epoch was occasioned by nothing more than his compulsion to get on TV and stay there for as long as possible, a project at which he succeeded wildly. At the start of his third campaign for president, virtually every likely voter had a fully-formed opinion of the man.

I cannot emphasize this enough: the segment of voters who do not know if they’d vote for Donald Trump, but could be persuaded to, is so small as to barely exist at all. One of my firm’s most consistent findings across the year (full disclosure: we are a qualitative firm, not quant, so we don’t deal in statistically significant samples) has been that voters invariably know what they think of Trump, and those who dislike him do so for personal, not political, reasons. There’s a hard limit to the number of people who will vote for him. Biden’s ceiling was not quite as hard, and his opposition less personally-motivated, but he also had the limited upside that comes with being extremely well-known but not hugely popular, for what that’s worth.

So who, exactly, is really up for grabs now? Let’s start with who Kamala Harris needs. Really two broad segments, as I see it.

The first debate scrambled everything; it blew a (potentially huge) hole in the Biden coalition as a bunch of people who had decided (with varying degrees of enthusiasm) that they would vote for Biden suddenly felt that they simply could not. In a large panel of Michigan voters - conventionally identified as “swing” voters by their partisanship score, which is method in the political industry for identifying people based on their likelihood of voting Dem (someone who watches MSNBC every night is a 99; the unquiet ghost of HR Haldeman is a 1) - we found that Biden lost a significant portion of his Dem voters and essentially every Independent and disaffected Republican on the panel. Our assessment, based on the texture of their feedback, is that he had no hope of reclaiming them, because voters can feel a lot of things about presidential candidates but it’s not clear if it’s possible to come back from being pitied.

When Harris took over the ticket, the Dems effectively returned en masse, and brought a large chunk of the Independents with them. Again, these voters were not saying “I was thinking about voting Trump but now I’ll vote Harris” (speaking just from our study again, essentially none of the voters who abandoned Biden considered voting Trump); they were saying “I feel fine voting Dem again” or its sharper cousin “Thank god I have someone I can vote for against Trump with a clean conscience.”

But not all of the Independents and disaffected Republicans came back. These voters are the first segment that is up for grabs. They’re not likely to vote Trump, but they were uncertain about Harris for reasons ranging from the understandable (she wasn’t nearly as well-known as Biden) to discreditable (race and gender are two great tastes that taste great together). They are mostly middle-aged and older (so, Elder Millennials and up); mostly White; and - based on my reading of the data - a touch more female than male.

What makes all these voters belong to one segment, in my view, is that they share the common trait of having existing predictive systems in their identities for figuring out how to evaluate whether a presidential candidate deserves their support - essentially, they might be difficult to please, politically, but they have some idea how a candidate can do it, even if they’d struggle to articulate that themselves. After the ticket change, they were waiting to see if Kamala Harris could satisfy the requirements of their predictive systems. Mostly, she had to show them that she’s competent and reasonable enough not to crash the country. The data suggests that she substantially has done so, particularly with the women in this segment; she’s clearly qualified herself with a chunk of the men, too, or she wouldn’t be tied or ahead in any battleground poll.

But the remainder of that segment, still technically out there to be won, is going to be very hard going. Again, most of these voters have fully-formed and personally negative views of Trump (if they didn’t, they wouldn’t be “undecided”; they’d be Trump voters); if they’re not already moving toward Harris, it’s not clear what else she can do. This is what the recent Republican-oriented push of the campaign is about, exemplified by Liz Cheney as a surrogate. And there’s an entire cottage industry in politics around persuading exactly these holdouts that Harris really is okay and they absolutely must vote because of how bad Trump is. It’s not clear how effective that effort has been; we probably won’t know until after it’s all over.

The other group that is up for grabs are younger voters. I cannot believe I am writing these words. If I have seen one progressive campaign break itself on the rocks of youth turnout, I have seen a thousand. It is very hard to get young people to vote.

The reason they are up for grabs - and the reason I group “young people” together despite the fact that they are not at all a monolith, hahahah goodness no - is that they do not yet have predictive systems for deciding whether they affiliate with a candidate or party. Based on work we’ve done this cycle, I’d say that GenZ (and some younger Millennials) not only do they not have a pre-built rubric for figuring out which party or candidate they like, by their lights they have never seen a working political system at all.

This might seem like an argument against the idea of young people as a critical segment; why would they engage politically at all if they’ve never seen politics work? And indeed, that’s how this election was going - until the ticket change.

Since then, new voter registrations have skewed staggeringly young - and female. Prior to Harris’s accession, for example, the Democratic Party had an existential threat in a devastating lack of enthusiasm with Millennial and GenZ Black women; that problem essentially resolved overnight and has been headed in the other direction, with spiking registrations that typically indicate heightened enthusiasm within the broader segment.

Some of this is the simple energy of having someone who is not A Very Old White Man on the ticket, and especially someone who isn’t Trump. While the young folk tend to have less formed political identities, we found in them the same thing we found in the older voters: rock solid views on Donald Trump, most of them negative (more on the exceptions to that later).

But just because they don’t like Trump doesn’t mean they’ll engage in politics; they didn’t like Trump before the ticket change, either, and were pretty switched-off. The Harris campaign and the larger progressive ecosystem have made some extremely canny choices in how to connect with young voters. Some of that is around media methods (I wrote about one small example here), but a lot of it is around substance.

For years, the conventional wisdom in Democratic circles was to message young voters with social progress, because the younger generation is generally quite socially progressive (yes, even a good portion of the men). What we found in our own work with GenZ is that those attitudes are not politically motivating; what they care about is material politics in general and home ownership in particular. And the Harris campaign, along with a number of big progressive orgs, have been ruthless about hitting exactly those points. What this communicates to young voters is that, for the first time in their lives, someone understands what they care about. This could have an outsized effect on turnout among younger voters.

Indeed, for what it’s worth, I think the larger political analysis ecosystem is underestimating young people in its models, especially GenZ women and in particular GenZ women of color. This is understandable, given that young people typically do not overperform in elections, but this is not a typical election.

This will not appear in most polls, and coverage of those polls will continue to present the election as being decided by a small group of voters who just can’t make up their mind between Trump and Harris. Listen to Melville - don’t look to long.

Coming soon: the voters Trump needs. That’ll be the last in the Melville series.