(C) Daily Kos This story was originally published by Daily Kos and is unaltered. . . . . . . . . . . Four Reasons We're Going to Win! [1] ['This Content Is Not Subject To Review Daily Kos Staff Prior To Publication.'] Date: 2024-10-22 There’s been a lot of panic lately regarding Trump’s “surge” in the polls. Most of the aggregators and all of the election markets are now predicting, with modest confidence, a Trump win and a subsequent end of civilization. You shouldn’t take anything for granted, of course (and vote and volunteer and give and all that — but please, if your only contribution is to say something like, “the only vote that matters is on election day” then this might not be the diary for you). But, and this is the important point, Trump is definitely not winning. He may win, but even just based on public data, he is behind at the moment, and here’s why. 1. A Plain Reading of the Polls Take my home state of Pennsylvania, considered by most people to be the lynchpin to the election on both sides. As of right now, 538 has Trump up by 0.3% in the aggregate. Not a lot, though the goons on predictit think that it equates to a sure thing. I’ve grabbed a screenshot of the current polling data from the last 2 weeks, and posted it in the top image. And what do you notice? First, visually, it’s easy to get freaked out because a lot of pollsters will include 4 variants of their topline numbers so you see Harris losing 4 times, which is obviously silly, but hard to just tune out. But secondly, you have an awful lot of unambiguously partisan pollsters. Trump’s own campaign, Daily Mail, American Greatness, Trafalgar, Quantus. Josh Marshall at TPM has a nice discussion of this although I wrote a number of diaries (as did others) in 2022. Partisan pollsters ended up something like 5 points to the right of the final outcomes in the midterms, and have performed badly in both the primaries and in special elections. They should be out and out ignored. There are other polls, like RMG (cheapo polls that come out too often, headed, eponymously, by Scott Rasmussen), AtlasIntel (very outliery numbers, and no real track record), and Emerson (which clustered with the RWers and were 5 points to the right of the final outcomes, even if the organization has no clear bias). I have no a priori reason to dump these guys, so I keep them. But even so, what do you see when you get rid of the explicit partisans? The picture changes significantly. Harris leads in most of the recent PA polls. The mean lead is 0.7, and the median (probably a better estimator when there are outliers) is 1. The point is though, that she’s ahead, and clearly so. And this is true across the board. Over a one month timeline, taking only the median of nonpartisan polls, we find: State 538 Margin Nonpartisan Diff AZ -1.9 +0.6 +2.4 GA -1.5 -1 +0.5 MI +0.2 +1 +1.8 NV +0.3 +1 +0.7 NC -0.8 -0.5 +0.3 PA -0.3 +1 +1.3 WI +0.3 +2 +1.7 Even though 538 (and other aggregators) are supposed to be aware of the effect, every single state is bluer when you just look at the nonpartisan polls, by an average of over a point. She leads in the Blue Wall plus NV and AZ. If you’d like to look at the polling 2 weeks out (selected for this) in a for you can use, good news, I made you a spreadsheet. You can (and I do) put this into a model, and assuming sampling errors for individual polls, and the possibility of (large) overall systematic errors, I find just a simple reading of the polls gives Harris a 57% chance of winning if the election were held today. Not overwhelming, by any means, but a hell of a lot better than the doom and gloom we’ve been seeing. Here’s my map: Using median nonpartisan polls from the last 4 weeks for each of the swing states. Nothing is certain, but I’d rather be us than them. By the way, there are polls that I’d be disinclined to use simply because they have terrible or limited past performance, and there’s no indication that there have been improvements, like AtlasIntel and Emerson. But even taking those out only bumps the estimate up to 58%. More significantly is the assumption that I (and everyone else) make that there might be a big systematic error. A lot of the pundits implicitly are assuming that a polling error will favor Trump, but as the midterms, the primaries, and pollster adjustments suggest, that error could be in either direction. For what it’s worth, if we assumed that the only noise is independent sampling noise (nothing systematic), Harris’s odds go up to a very nice 69%. But as I’ll argue, I think that if anything, the polls are underestimating her. 2. History and Digging Beyond the Toplines The democrats won the 2020 election — a non-controversial opinion around these parts, I know — and Biden won by 4.5 points nationally. Indeed, Dems have won the popular vote in every election since 1992 except for 2004, and won only 1 between 1968 and 1988. What’s more, after ignoring some large 3rd party elections (1968, 1992, 1996) or one offs like a very popular incumbent in 1984, there’s a pretty clear trend where the Dems improve by a point or so each cycle — but with some scatter. As the electorate gets more diverse — and white voters will make up about 2% less of the total than they did in 2020 — Dems improve. Most demographics are relatively stable in their voting patterns. A 10 point shift is unusually large, even within a demographic group. Looking at exit polls from 2020, for instance (CNN and Pew provide somewhat different numbers, but they amount to similar pictures), here are some big picture numbers from the last election: There is roughly a 10 point gender gap (women are to the left, naturally) White voters preferred Trump by about 15 points (17 CNN, 12 Pew) Black voters preferred Biden by about 80 points (75 CNN, 84 Pew) Hispanic voters preferred Biden by about 25 points (33 CNN, 21 Pew) Under 30 preferred Biden by 24 in both polls — though this is the only one of the demographics that by definition is a significantly different group of people in this election compared to the last. Plugging in numbers like this, and assuming a ~70% white electorate (along with, for instance, Dems winning other minority groups by 20 points), you get a 4.3% popular vote lead; very similar to the actual 2020 outcome. What’s more is that for every group besides white voters (who make up a majority), we have other polls of just that subgroup — and recent ones as well. Black voters: Harris+70 (NY Times +63, Pew +71, Howard U. +76) Hispanic Voters: Harris+20 (Pew +18, GQR Swing state +33, NY Times+19) Under 30: 30 (Harvard Youth +31) These polls aren’t the be-all and end-all, but they’re targeted, and for hispanic polls, for instance, they are conducted in both English and Spanish. They are aiming to get a fair sample of a group that will normally make up a subgroup in a poll. They are similar to the actual outcome of a previous election (which had, after all, both of the current candidates in it), where Harris does a little worse with black and hispanic voters, and a little better with young voters. So, for instance, if you have a poll that has Harris only winning black voters by, say, 40, that is almost certainly a huge underestimate. Because they make up 13% of the voters and the margin is ~30 points too low, you’re almost certainly underestimating her overall margin by about 4 points! More in places like NC and GA. And the effect is significant, especially in polls where people self-identify their demographics (random dialers). Across 4 recent national polls (Marist, Fox, HarrisX, YouGov CBS), Harris’s black voter margin was only 52, suggesting a systematic underestimate of 2 ½ points overall. The hispanic average was about 15, a little better, but still erring in the same direction. Meanwhile, there’s evidence that Trump is actually losing noncollege white voters (in addition to the women and college educators voters he’s lost in a post-Dobbs world). All this, while the fraction of white voters is decreasing. Another problem with some of these polls is that they have unrealistic white voter fractions. That weirdo Fox poll that had Trump ahead by 2? They had a 72% white voter share. It’s likely to be much closer to 67, which itself entirely reverses the lead. Finally, putting in realistic numbers (H+75 with black voters, H+20 with Hispanic and other minority voters, 68% white vote, 53% women, T+9 with white voters — suggested by an average of the 4 pollsters above), you get H+7.5 overall — a popular vote that will carry the day no matter what the EC advantage. 3. The Enthusiasm Gap and the Ground Game Beyond picking apart polling, there are facts on the ground. Kamala Harris has raised over a billion dollars in small donations (and Trump has not). She fills giant stadiums wherever she goes (and Trump cannot). People are canvassing and organizing in a way that’s reminiscent of our ground game for Obama. The 4 most popular people in the contest (in order) are Walz, Harris, Vance, and Trump. People genuinely like the folks on our ticket. Here are Harris’s favorables: She started strong, and is still improving. But perhaps even more significantly, according to Emerson, at least, among people who’ve made up their minds in the previous month are breaking 60-36 for Harris. In other words, this is a very different situation from 2016. 4. Actual Data so Far Finally, there is the reality that people have already started to vote. TargetSmart tries to capture what vote by mail and early in person voting can prognosticate about the final outcome, but it’s tough because the environments are so different from either a midterm or the pandemic environment of 2020. But here’s what we do know: Female voters represent a larger fraction of the 14.4M early votes than in either 2020 or 2022 (elections in which we did very well), in a post Dobbs world. The same is true for black voters. But, for what it’s worth, the same is also true for white voters. Hispanic voters have (relatively) decreased. Dems, unsurprisingly, are dominating the overall early votes compared to republicans, by something like 16%. In PA, out of the million early votes, Dems outnumber republicans more than 2:1 (as you’d expect given the reality of the voting modes we’ve been stressing), but more importantly, even of those who requested ballots, Dems have a 60.1% return rate while republicans are only 53.6. We really are more eager to vote, and plenty of republicans look like they’ll leave their ballots on the shelf. Similar patterns are seen in WI and MI. In Georgia and NC, they are smashing early voting records, and that includes votes by younger (18-24 yo) voters. These are banked votes for Dems. There are lots of unanswered questions about what early voting means when there’s so little basis for comparison, but if the turnout is high, and large numbers of women, young people, and black voters (relative to previous elections) are out there, then those are very, very good signs for us. In short, nothing is certain in this world, but I think we’re gonna win. Maybe by a lot. [END] --- [1] Url: https://www.dailykos.com/stories/2024/10/22/2278591/-Four-Reasons-We-re-Going-to-Win?pm_campaign=front_page&pm_source=more_community&pm_medium=web Published and (C) by Daily Kos Content appears here under this condition or license: Site content may be used for any purpose without permission unless otherwise specified. via Magical.Fish Gopher News Feeds: gopher://magical.fish/1/feeds/news/dailykos/