Dem
51
GOP
49
Last week, we wrote a fairly lengthy item about that allegedly-grim-for-Joe-Biden poll from CNN. Our opinion, in case you have forgotten: It doesn't mean anything. We are too far out from the election, and there are too many unknowns between now and then to be reaching firm, or even semi-firm, conclusions about the dynamics of the race.
Of course, the nation's most prominent psephologists are not paid to say "pay no attention to the polls," so they have to say something when new polls come out. And so, we would not expect to see too many of them agree with us 100%. That said, the other Nate (Cohn), the one who works for The New York Times, did write "Polls taken 15 to 27 months out don't necessarily augur much..." in his latest piece. That's actually pretty close to "pay no attention to the polls," especially since Cohn works for an employer who actually pays for polls and expects them to attract eyeballs.
Cohn also agrees with us that the demise of Joe Biden has been greatly exaggerated. In fact, looking closely at the numbers, he thinks that there's actually some pretty good news in there for the President (and some bad news for his predecessor). It is true that the two presumptive nominees are running roughly equal in national preference polls. However, don't assume that 2024 is the new 2020 or the new 2016. While acknowledging that the data is not entirely reliable, given how far we are from the election and how tough it is to tease fine-grained conclusions out of polling results, Cohn observes that Trump appears to be gaining strength among minority voters, while Biden is improving with swing state voters.
If this is true, then it could well be a zero-sum game (or close to a zero-sum game). That is to say, the number of minority voters moving in Trump's direction could be offset by the number of swing-state voters moving in Biden's direction. That would cause national preference polls to look the same, even if the actual coalitions were of a different character. Meanwhile, even if you didn't already read the Cohn piece, you probably see where this is headed. Minority voters tend to be concentrated in states that are already in the bag for one party or the other (blue, blue California and New York or red, red Texas and Alabama).
On the other hand, swing-state voters are, without exception, found in swing states. So, a "same old, same old" result in the top-level national number (say, 46% for Biden, 46% for Trump) might obscure a meaningful change in the projected electoral totals. The bottom line, as Cohn writes, is that "the demographic foundations of Mr. Trump's Electoral College advantage might be fading." Put another way, it may no longer be correct that if Biden and Trump are tied, then Biden's in trouble, because he has to win by at least 3-4 points to win the EC.
That said, for every action there is an equal and opposite reaction. And so it is that while Cohn sees silver linings for Biden, CNN number cruncher Harry Enten sees rain clouds. While we have no reason to think Enten is in the bag for Trump, and significant reason to believe that he's left-leaning, he's also made something of a cottage industry of pumping out "good polling news for Trump" pieces. In just the last 6 weeks, for example, he's published items under the headlines "The chance of Trump winning another term is very real," "Republican voters think Trump is electable. They may be right," and "The group that won Trump the election in 2016 may win it for him again in 2024." That's quite a lot for a guy who only produces 2-3 pieces a week.
The major argument of the last, and most recent, of those pieces is this: Among voters who really don't want to vote for Trump or Biden (which is about 20%), Trump is slightly more appealing (or, if you prefer, slightly less odious), roughly 53%-47%. Needless to say, if those voters ultimately decide to suck it up and vote for the lesser of major-party evils, and if they break the way they are feeling now, and if the other 80% of the electorate sticks with their current horse, then Trump would probably win. As Enten points out, the numbers here are similar to 2016, when there were also two widely disliked candidates on the ballot, and where the so-called double-haters broke pretty decisively for Trump, thus handing him the win.
If you are interested in which analysis we put more stock in, the answer is Cohn's. First, his analysis just generally tends to be better and more thoughtful than Enten's, especially since he spends most of his time writing, and does not spend a bunch of time on TV hits. Second, Cohn shows his work to a far greater extent than Enten does. That may be a byproduct of the outlets for which they work, and the audiences they are trying to reach, but the fact is that Cohn's piece takes you through the math with graphs and charts and everything, whereas Enten's is positively breezy. Third, and finally, we specifically don't love Enten's argument. Beyond the fact that 7 points is a relatively small spread, considering all the other X factors (something that the author himself concedes), it is also the case that double-haters tend to favor the status quo. There was no status quo in 2016, since neither candidate was a sitting president. There is now.
With that said, the real point is this: Because the numbers are so fuzzy right now, it's easy enough to find analysts who will argue that bad numbers are good, good numbers are bad, bad numbers are even worse than they seem, good numbers are even better than they seem, up is down, left is right, black is white, no is yes, and the clocks are striking thirteen. Ok, maybe not that last one, but the rest of it. And so we're back at the same place we were last week: Polling numbers are fun to talk about and dissect right now, but don't take them too seriously. (Z)