Catching up on polling

Yesterday there was but one lonely national poll. The head-to-head survey was taken among likely voters; the surveyor, considered highly reliable by 538.

“The 2024 presidential race currently mirrors 2020, with the Democratic lead narrowing from four points to two in national polls,” said Spencer Kimball, Emerson College Polling's executive director. Why he said "national polls," I've no idea. Numerous polling operations have put Harris at +4 or +5. He might have meant only Emerson's national polls. If so, he didn't say what he meant.    

On Kimball's site, he first presents both the unfathomable and the heartbreaking: "Those who say the economy is the top issue facing the country break for Trump, 62% to 36%, as well as immigration, 84% to 13%. Those who say threats to democracy break for Harris, 84% to 12%, along with housing and abortion." 

In the land of Compos Mentis, Trump would outflank Harris only among those who desire grift, graft, unremitting falsehoods and ground-breaking incompetence from their central government. On the economy, Trump has top executives and most businesspeople nearly paralyzed with fear because of his idiotic tariff proposals. Yet there's a frighteningly large voting bloc that would prefer Trump over Harris: Those who nostalgically itch for a second Great Depression. And immigration? Only sadists have a strong preference for Trump and a liking for detention concentration camps. 

As for the generic congressional ballot, Emerson found Democrats leading Republicans by 48% to 44%. That's outside the MoE, so hope for an intelligent or at least a more intelligent Congress remains grounded. Also notable and somewhat disturbing is that generic Democrats are slightly ahead of Harris.

This line from Emerson reads as though its pollsters had read "Don't Trust the Election Forecasts," Prof. Grimmer's Politico piece, here, and which I commented on, here. "Survey results should be understood within the poll’s range of scores, and with a confidence interval of 95% a poll will fall outside the range of scores 1 in 20 times."

Its polling must be fairly correct, since Rasmussen, same day, released this.

There is one other poll — again by Emerson, commissioned by The Hill — that's fresh, but bewildering. The survey of Harris vs. Trump was mostly conducted in the readily predictable states of Florida, Texas and California. Why The Hill wanted to squander its money on Shakespeare's original coinage — a "foregone conclusion" — is also bewildering. Only the polling in Ohio was of interest. It showed Trump leading by 10 points. Even that was rather predictable, but still worthwhile because of some pundits' remarks that the state might be in play this year. 

Comments

[VoiceOfReason]

Nate Silver seems to think everything has flipped back to favoring a Trump win: https://www.newsweek.com/nate-silver-election-model-shows-donald-trump-surging-1949000 -- I must admit to finding that hard to believe.

[PM]

Silver seems to enjoy being seen as "provocative." It's part of the ego factor in modeling that Prof. Grimmer wrote about in his slam of the profession. Still, as long as his readers accept his findings for what he calls them — election "forecasts" — then maybe his doomsaying will energize them.