Here's the latest YouGov breakdown of the presidential race:
Kamala Harris started out three points behind Trump and is now three points ahead.
I should note that I routinely show the YouGov poll for a couple of reasons. The first is that it's a pretty good poll: reliable, weekly, and with good crosstabs. The second is that by tying myself to a single poll I avoid cherry picking the polls that are most favorable each week and thereby fooling myself about how well things are going. It's a discipline thing.
And for those who are going to insist that only state polls matter, that's not really right. State polls are certainly useful, but they also tend to be smallish and error prone. And while it's true that we don't elect presidents via a national popular vote, the states do generally follow the national vote. If Harris gets to a five-point lead, it will be very unlikely that she loses the electoral vote. That's the target.
That's a 6 point turn around. The Dem convention ought to be worth a couple of points. This should send Trump over the edge. Nastier, meaner, bigger lies, stupider lies. Good times.
I'm pretty sure the Harris-Walz rally in Milwaukee had more people than Trump did during his acceptance speech.
He's never been able to controll his worst impulses. It should get very interesting. Unfortunately in a train wreck sort of way.
There have been news reports that RFK JR is going to drop out and endorse Trump. As he and Trump have been splitting the stark raving loony vote, that might tighten things up.
But a good number of those loons might say "I am voting for my loon, not your loon."
Agree. It will help trump a little. But I suspect a lot of rfk backers dont think of him as nearly as crazy as trump. They may not be willing to go out that far. I also suspect some of them couldn’t back Biden but are willing to come back for Harris. (Like me, although I was never a rfk backer.)
They don't think RFK is as crazy as Trump? ?? Trump is a crass opportunistic con man. RFK is a conspiracy riddled crackpot. Of the two RFK is definitely the craziest.
Yes, but the tiny number of people who support RFK probably don't think he's crazy.
The question is what does an RFK partisan do when deprived of the opportunity to vote for RFK? Since the average person is not an RFK supporter, we can't really figure that question out by thinking about what most people would conclude about either RFK or Trump.
The only RFK votes that matter are in the 7 swing states. Has anyone figured out where RFK’s support has been coming from? If it’s CA, TX, NY etc, whatever he does is meaningless. His polling numbers have gone way down since Biden dropped out (hence his decision to sell himself to the highest bidder). And does anyone think there are large numbers of disaffected Trump voters in swing states who had decided to vote for RFK because he was a viable candidate? Of course not. If you voted for Trump before but won’t vote for him again and won’t vote for a Democrat, you will find someone else for your protest vote.
The only RFK votes that matter are in the 7 swing states.
This.
Has anyone figured out where RFK’s support has been coming from?
No idea, but I'm tentatively hopeful his withdrawal doesn't mean much either way. Some of the polling suggests his presence in the race hurt Trump, although some polling suggested otherwise. My guess is the fuzzy data means his withdrawal is mostly a wash: if Trump ends up benefiting, it'll be a tiny effect.
a sizable number might sit out the election or vote for one of the other minor candidates on the ballot. question is, did these people support rfk jr. because they thought he was the best candidate or because they found the two major candidates to be disagreeable? or are they a bunch of contrarian cranks?
"The only RFK votes that matter are in the 7 swing states."
He's not on the ballot in 7 swing states. And of the five-ish swingy states (AZ, NM, WI, MI, and PA), he's on the ballot in just two of them (AZ and MI).
RFK Jr. is bad, but here's a quote from his running mate. "Do we trust Trump and his personal sincerity to really do the right thing for our country, end chronic disease, balance the budget, end these forever wars?". His personal sincerity??? She could not be more ignorant and naive about who Trump really is. So much so that it's genuinely frightening anyone would pick such a person for V.P. But money talks. And RFK is nutty enough to make such a bad choice.
I can translate that for you: "I'm a billionaire and Trump will lower my taxes, and Harris will raise them"
RFK has been polling lower and lower. Not that many supporters left. And that was not a rumor but a quote from his very wealthy veep candidate.
RFK never polled well, and never had even a plan to get on the ballot in enough states to have even a technical shot at winning. Nor was he a specific issue candidate who was forcing the major parties to (for good or for ill) confront something voters cared about that the Rs and Ds would otherwise ignore.
Accordingly, he should have never been covered by national media; certainly not to the asinine extent that he has been. He's a non-story.
i disagree. his name carries some weight and i believe some early polls showed him picking up as much as 10 percent of the vote. he certainly had no chance of winning but he could have influenced the outcome of the election. he's crank and that needed to be exposed for anyone who might have taken his campaign seriously.
"his name carries some weight . . . he could have influenced the outcome of the election."
The idea that somebody who cannot win and has no plan to win should be covered as if he was a legit candidate because he's related to famous politicians from a half century ago is borderline offensive.
And, as I noted above, he was not running on some issue such that his presence in the race forced the other candidates and the election more broadly to highlight it; accordingly, the only mechanism he had to "influence[] the outcome" was by getting a lot of free media from bored national news outlets covering him as if he was a real candidate. He wasn't.
"he's crank and that needed to be exposed"
That's not how this works at all. Covering his "campaign" lended it legitimacy, and thereby buoyed it in the polls (for chrissakes, that's what happened with Trump in 2016).
There are millions of cranks in this country, and in the 1200+ people officially* running for president in 2024, cranks are undoubtedly statistically way, way overrepresented. Should the national news media have expended increasingly scarce journalistic resources on credulously reporting the campaign events and youtube ads from those folks, just to ensure that Americans got a chance to be "exposed" to their crankery?
And if you say "Well, RFK has a famous name!" there were three Washingtons, two Carters, one Nixon and two other Kennedys on that list of official candidates, too.
Bottom line, if a candidate has no chance of actually being on the ballot nationally in a national election, then it is goddamn malpractice for the media to waste the country's attention span on reports about their candidacy.
_______
*https://ballotpedia.org/List_of_registered_2024_presidential_candidates
I’ve been watching the Democratic convention each night and it has been a great show. Biden’s speach was historic. His last turn on a national platform like that. Michelle and Barack both gave amazing speeches. Hillary had a standing ovation that wouldn’t end. AOC gave a commanding talk. Bernie Sanders was his usual grumpy self. Tonight Tim Walz came across as such a regular guy like I used to see at our neighborhood cookouts.
This has been really well produced and managed. I’m curious to see the surprise speakers tomorrow night. They only listed Kamala herself. We already got to see Oprah. Who is left? The Pope? Taylor Swift? Beyoncé? Gavin Newsom?
This was really a love fest. Every day there have been a handful of Republicans talking about how they can’t stand Trump and will vote for Kamala this time and they were welcomed with open arms. The roll call of the state delegations was a great party with a DJ and music and was really fun to watch.
Bill Clinton looked rather old, yet he is 78, same as Trump.
It was also interesting that the crowd shots and most of the speakers looked like America. Men and women of all races and orientations. This really looked like the future of the country.
By contrast, I heard that while Fox carried the convention they didn’t show their viewers any black speakers. This was up to yesterday. I wonder if they showed the Obamas?
There were some protests outside the hall but they didn’t amount to anything.
It was an amazing contrast to look at this versus the recent RNC. Outside of Trump who do the R’s have? Who are the up and comers after Trump? Maybe Nikki Haley. She could run in four years.
On the Democratic side it is an embarrassment of riches: Newsom, Shapiro, Whitmer, AOC come to mind easily. AOC came across really well and she is just 34 years old. In 8 years Newsom will be 64 years old. Maybe he could run with AOC?
Lawrence O'Donnell kept complaining about how the speeches ran really late, past 11PM Eastern time. It was just fine for me here in California. LOL
I would be shocked if Kamala doesn’t get a significant boost in the polls next week.
Some say conventions don’t help that much, but this year is different ( have you all noticed?). Much fewer opinions of the nominee were set before the convention than usual. More of the voters just now tuning in will be pleased than not by what they have seen this week.
Bill Clinton looked rather old, yet he is 78.
It's the voice, more than anything, is the thought that struck me when I listened to his speech. In his prime Bill Clinton was a highly effective public speaker (maybe not quite at Obama's level, but very solid). His speech was still pretty good, I thought, but his voice is a pale imitation of the Bill Clinton of 25 years ago. Similar to Pelosi, or Joe Biden for that matter: the passage of the years is apparently hell on the larynx. Aging sure beats the alternative, but so far Father Time is undefeated.
Jasper, you make a good point about the ravages of time. And a phrase you used really brought it home:
"...but his voice is a pale imitation of the Bill Clinton of 25 years ago."
It's hard for us old-timers to think of it this way, but 25 years ago Bill Clinton was reaching the end of his second term.
Lordy, the time sure flies.
It was a great show...
...if you watched the convention feed and not the network tv one which was selective, skipped all the music and celebrity, and much of the introductions to talk over them often about needless things.
Negative Nellie! Here’s the polls of a 5-way race since the start of August: Harris by 2, Harris by 3, tie (Fox), Harris by 4, Harris by 3, Trump by 4 (Rasmussen, which is accurate only in a low turnout), Harris by 2, Harris by 3. ….
Polls of 2-way race since start of August: Harris by 1, Trump by 2 (cnbc), Harris by 3, Harris by 3, Harris by 2, Harris by 1, Trump by 1 (Fox), Harris by 4, Harris by 4, Trump by 4 (Rasmussen), Harris by 3, Harris by 4
…..
Last polls listed, where Harris does better, are latest polls. All before convention….
No, not victory. But tell me that this means nothing.
Notice Harris does better in 2-way race. So
Much for rfk effect.
Rasmussen's final poll will exactly match the poll of polls and will probably have Harris up by 4. The will be 6 points more Trump up til then.
The second is that by tying myself to a single poll I avoid cherry picking the polls that are most favorable each week and thereby fooling myself about how well things are going.
You can get the same effect—I'd imagine with greater accuracy—by using the same aggregator or average.
Quoting from memory, but Silver has Harris up 2.5 points. 538 has her up up 3+. RCP has her up by 1.8 (these are averages, not individual polls). Big improvement, though I won't sleep easy until after it's over. And, needless to say, if the general take of these polls is correct and the polling is off by magnitudes similar to 2016 and 2020, Trump is going to win the Electoral College. Let's hope they're more accurate this time around (or her lead widens by at least 3-4 points). 2.5 points for me is the tripwire: it gets increasingly difficult for Harris to win if her margin over Trump is much below that, I think, and increasingly challenging for him if Harris's lead is over that. Thanks, Hamilton!
Follow up: Per Nate's latest update, the Democratic ticket has made gains in all seven swing states over the last month. But he has TRUMP making (modest, to be sure) gains in fully six out of those seven over the last week.
I sure hope Harris gets a convention bounce!
https://www.natesilver.net/p/nate-silver-2024-president-election-polls-model
Memories of state polls: https://www.cnn.com/2016/08/10/politics/clinton-leads-trump-wisconsin-poll/index.html
His averages let known polling bias pull the weekly tracking number around.
It's California. It's the only very large state that isn't fairly closely divided politically. Florida and Texas are reliably Republican, but only by about ten points (55%-45%). California is THIRTY-SEVEN points for Harris over Trump.
She's a Favorite Daughter on top of being a Democrat. So she has to win by at least five points in the PV to be confident of winning the EC.
Biden got that range in 2020, but just squeaked through in MiWiPa by 100K votes. Unless Arizona stays Blue, Harris has to win all three Blue Wall states.
Arizona and Nevada will have a woman’s right to choose on the ballot. So will a number of states that are not normally contended. Although Florida is one, and the Democrats ought to at least put some effort there. Besides the abortion issue, they might remind Floridians that GOP climate-change denial hasn’t kept the Atlantic and Gulf out of their streets and yards.
If Harris gets to a five-point lead, it will be very unlikely that she loses the electoral vote.
Since Kevin's post is about polling—not the actual vote—I would stress it's really not about a five point lead in the polls that counts, but a five point win in the popular vote. Those are two very different things. It wouldn't be unheard for the polls to underestimate Trump's support by 2-3 points. If the polling inaccuracy is toward the higher end of that, a 5 point polling lead for Kamala Harris could easily translate into a Donald Trump Electoral College victory.
As I understand it, the factors causing the polling to be off in 2016 were different from those in 2020. So my hope is the polls will be much more accurate this year. I don't think that hope is unrealistic—I just think it's highly uncertain at this point.
But yes, the national polling is a very useful tool—at least if it's reasonably accurate—in estimating who's going to win the electoral college. Trump probably needs to keep within three points of Harris's popular vote margin to have a hope of winning the Electoral College. And she probably needs to be above that to be confident of victory. Where it gets dicey is when the popular vote margin is between two and three points.
I also wonder if polling organizations have made changes to counter the errors that were made in 2016 and 2020. I found this article about that:
https://www.cnbc.com/2024/05/04/why-election-polls-were-wrong-in-2016-and-2020-and-whats-changing.html
Some people will start a poll, they’ll tell you who they’re going to vote for and then they say, ‘I’m done. I don’t want to talk to you anymore. Goodbye,’” Don Levy, director of the Siena College Research Institute, which helps conduct polls for the New York Times, told CNBC. “In 2020 and 2022, we didn’t count those people.”
But this time around, Levy says they are counting the “drop-offs.”
They found that if they had counted those impatient respondents in 2020 and 2022, their poll results would have moved “about a point and a quarter in the Trump direction,” Levy said, eliminating roughly 40% of their error.
In addition, the article says that non-college-educated voters are being weighted more this time.
That's interesting. This topic has been on my mind. My guess is this year's polling will end up being more accurate, for the simple reason that, one would imagine many/most pollsters would have learned at least something from the last two cycles (and made adjustments accordingly). But it's really just a guess.
The major pollsters have made adjustments for underestimating the Trump vote before. But have they over adjusted? Also since 2022 the polls have tended to underestimate the Democratic candidates in off year elections. This Democratic overperforming has been to the tune of 6-8%. Hard to know what this means. Some argue it isn't the same because Trump wasn't on the ballot but it could mean that the enthusiasm of the Trump cult is waning. It also could reflect the role of abortion in many of these elections.
I think Trump will have difficulty breaking 46% and will lose by 5-6%. Harris needs to win by at least 3% to carry the electoral college
Yes, it's the Pissed Off Dummies Poll Deflator, the "PODPD".
"As I understand it, the factors causing the polling to be off in 2016" The 2016 vote missed by half a point in the individual's votes. No poll can be ever expected to do better. "So my hope is the polls will be much more accurate this year. " That's a delusional hope with respect to 2016.
538's weighted aggregate says Harris is up 3.3 points. When Biden left the race, he was down 3.2 points; that's a 6.5 point swing.
Not sure what the benefit is, sticking to a single poll rather than stick to a single aggregate that adds weights relative to the quality of the polls it includes. And of all polls, YouGov is so tedious, which becomes a problem with the quality of answers.
I think this convention will generate a ~+1.5-2.0 point semi-permanent bounce, half of which is actual movement and the other half being engagement/participation bias -- the effect of increased excitement of Democrats that were always going to vote for Harris but are now willing to answer the phone and engage polls.
I have to look this up but, if the polls consistently show the same gap between candidates, do the normal error bars still apply. In normal probability and statistics, the error is based on an random sampling in which case the results should bounce around within the margin. The polls haven't done this. For Biden the were consistently 1-2 points below Trump. OR do the error margins try to capture poll/model BIAS as well which will be in a certain direction.
I would apply the same questions to a consistent move in the poll results. If Harris swings 6 points up and it is consistent, then do we really say, "There is ALWAYS a 5% chance that she is actually be 3 points BEHIND Trump?+"
It's kind of similar to research studies. When all your not ideal studies show the same corrolation, you accept that cigarettes cause cancer.
J.D. Vance called Tim Walz's speech last night 'dark and ominous'. These idiots can't see past their own hand mirrors.
Democrats have been absolutely killing it so far this week in Chicago. A truly masterful performance. Vance and other Republicans are whistling past the grave yard, and they know it.
I'd be surprised if Harris doesn't pick up another 2-3 points in most of the polls by Labor Day. If she's up in most of them by 5-7 points in early September and if the polling is a bit sounder this cycle, she's in a pretty strong position, and I'd guess her odds will be trending well about 60% at that point, facing an opponent who's aging, erratic, undisciplined and widely loathed (with a running mate who is literally the least popular since polling has been done on this topic).
I was at the Milwaukee rally. What I noticed in line to enter was that the crowd looked to be at least two-thirds female. Just ahead of me was a group of four women from Kenosha (suburban women?). Inside the energy was through the roof. Kamala wasn't my first choice, but she's more than adequate to beat Trump who is a deeply flawed candidate. And I'm warming up to her. One of the strongest crowd responses was when she touched upon reproductive rights in her speech. The crowd's response was deafening. As a woman she is well-suited to champion this issue which is perhaps the Democrats' strongest. I think you will see women coming out of the woodwork to vote for her this election.
From the moment Dobbs was leaked, I expected it to set off a generational earthquake in electoral politics. Or maybe more like a tsunami-- developing deep, strengthening over time, leading to big changes. I think it will surprise a lot of the pundits and politicos who've thought it's all been assimilated by now, but you don't suddenly decide that more than half the populace doesn't, after all, have a fundamental constitutional right they've built their lives on and not get a reaction. "Oopsie, Roe was a silly mistake, now let's all move along" just doesn't cut it.
Kamala packed them in in Chicago, Milwaukee and the Colbert show simultaneously.
That's a level of enthusiasm no other convention has ever shown.
Colbert has become such a political commercial for the Democrats that I have been watching MSNBC clips instead of Colbert on YouTube*. At least the political analysis is better. I can't stand listening to politicians. They are so transparent and predictable that I feel dumber after they talk.
*- For the record - I am a total Colbert Head
Check out Seth Meyers segments on YouTube. His A Closer Look bits are consistently great, not just for political comment and laughs but in theatrical sophistication.
Of course it is impressive when someone can be in three places simultaneously… 🙂
All the polls tend to be "smallish" - that is they take at most one or two thousand samples to get the raw statistical uncertainty down to a few percent. Getting larger samples has diminishing returns with respect to reducing the uncertainty. There is nothing about state polls that necessarily makes them more "error prone" than national polls, and presumably they are not wasting a lot of sampling on areas that are not in contention.
Just as an incremental vote in Pennsylvania will count much more in the electoral college - there is no doubt at all about which way California will go - so the polls should sample accordingly. If the total sample is only a couple of thousand (more than most polls) taking them all from swing states rather than nationally would reduce the actual uncertainty as opposed to the reported +- uncertainty, which is an absolute minimum (even disregarding bias in sampling).
If Kevin can find reliable swing-state polls with sample size equal to or even a sizable fraction of the usual national polls he should in theory use them and aggregate the results and errors. Maybe some pollster is doing this - they get paid for this and Kevin doesn't (nor do I).
It should not be necessary to rely on the five-point fudge factor, or whatever approximately relates popular vote to electoral vote. What is the uncertainty on that and how do you determine it? That must be added to the raw error of the national poll.
If the national polls (or state polls for that matter) are wrong about the popular vote, that is if there is a bias, that is an additional contribution to the uncertainly.
If Harris gets to a five-point lead, it will be very unlikely that she loses the electoral vote. That's the target.<
This is ridiculous. If she gets a 10 point lead it's also unlikely that she loses the electoral vote. Should that be the target?
Hillary Clinton won the national vote by about 1 percentage point, yet she lost due to a total of a few thousand votes over a few states, which is microscopic. Yes, as you say, "states do generally follow national vote", but the "generally" is doing a lot of heavy lifting. They might deviate from the overall trend only a little bit, but a little bit is all that's necessary to swing the election.
IMHO, a better way to look at this is where the "tipping point" states (where you cross 270 EV) are relative to the "tied" states. If you're up in the tipping point states by 2-3 points relative to the tied states, you're in pretty good shape. She's almost, but not quite at that point now, but it's still August and there's a long way to go.
3% up on the popular vote probably puts her about even chances on the Electoral College.
Sounds about right. Clinton came in 2.1 points over Trump in 2016 and lost. Biden beat Trump by 4.3 points in 2020 and had a bit of breathing room in the Electoral College: he would have won even if he had fallen short in his two closest states, Arizona and Georgia. There's been some talk that the GOP advantage in the Electoral College isn't quite as strong this cycle, but I'll believe it when I see it.
"...the states do generally follow the national vote..."
?! Kevin, what nonsense is this? In what sense did WV, AL, TX follow the national vote in 2020? Even if we squint and say "well, we sort of know what he means", it is only qualitative and not quantitative. It is still entirely possible that most of Harris's gains will ultimately take the form of running up the score in blue states, and it is still overwhelmingly likely that the final result will be decided by statistical noise, so that no one will be able to say what it means -- and an election whose meaning is not clear and agreed upon is net harm, regardless who "wins".
It’s not nonsense, and it is quantifiable: Joe Biden ran 2 - 3 percentage points better against Trump in those states in 2020 than Secretary Clinton did in 2016 (state-by-state results tallied in Wikipedia), close to his gain in the national popular vote — because state votes correlate with the national vote. Here’s a discussion on Andrew Gelman’s very good site: https://statmodeling.stat.columbia.edu/2020/08/31/problem-of-the-between-state-correlations-in-the-fivethirtyeight-election-forecast/
"poll site:jabberwocking.com"
Some joy and hope for the dems after all the negative repulsive doo doo for years. You rock Kamala.
A no-cost sign-up to Nate Silver's Silver Bulletin will get you daily updates on poll numbers for the battleground states, but not the details and commentary for subscribers. I track those daily.
I did a simplistic state-level analysis of the Nov 2020 vote shares (Biden, Trump, 3rd parties) versus the final (Nov 1) polling averages (Biden, Trump, Undecided (excluding 3rd parties). It showed me that Biden's vote share was 94% of his polling number plus 44% of the undecided. Trump's vote share was 103% of his polling number plus 15% of the undecided. In other words, Trump had a 3% higher vote share from people who did not talk to the pollsters.
Rolled together for battleground states, Biden needed to go in with about a 2.5% lead in the polls on average to actually win a close race on election day. In AZ, he was up 48-46 in the polls with 4-5% undecided right before election day. He won by a 0.3% margin over Trump with 1.6% going to 3rd parties. In GA, he was up 48-47 with ~4% undecided before election day. He won by 0.3% over Trump with 1.6% going to 3rd parties. There is reason to believe (noted above, inclusion of people who did not complete the polls, better representation of non-college voters) the bias may be smaller now.
Nevertheless, Harris probably needs to have a 2+% margin in enough battleground states to win the electoral college. Also, what happens to the vote for 3rd parties dwarfs the Dem-Rep margin in those states. It will be hard fought and will come down to the very end unless Trump tanks.
This is great.
Hopefully we dont see a repeat of 2020 where a 7-9 point polling lead in late October turned into a contest determined by less than 50,000 votes across a few states.
The polls have been fairly useless predictors over the last 2 election cycles.
We've all been burned by polls we wanted to believe were "right". I get it.
I believe Harris would win of the election were held tonight. In fact, I don't think it would be close.
The problem is that November is still a "lifetime" away and a lot can go wrong.
On the involvement level, Harris has gotten more people involved and interested in the election. More people have donated and the excitement level is very VERY high. But these are soft metrics - or, are they?
Policies have been largely ignored by the candidates from both parties. (I won't even discuss RFK). Fundraising has gone solidly DEM as of this week.
My beliefs are just that - my BELIEFS. If Kamala can remain mistake/gaffe "free" she should win. I don' think this one will be as close as Bidens win - but we aren't in landslide territory yet either.
A Trump loss will trigger some anxiety - this I'm sure of