A 5% lead in a poll is probably still within the margin of error (or very close to it); it likely is still a “coin flip.”
She’s doing worse than Biden in 2020 and that was a coin flip if there ever was one. Biden won Arizona by 10k, Georgia by 12k and Wisconsin by 20k. Random chance absolutely played a part in his win, everything down to traffic and weather played a part.
With Trump, unpredictability and illogic are such that, IMO, polls are meaningless and irrelevant. They truly don’t matter. Trump could, and possibly will, win.
Yeah, cuz he’s MAGIC, baby!
Guys, Trump LOST in 2020–it CAN happen, LOLOL!!!
Trump WON in 2016–it CAN happen, LOLOLOLOLOL!!!
But, seriously… @Aeschines, I know you really want Trump to lose, but it seems like, every time there’s a poll number, or a news story, which sounds good for Harris, you have this tendency to get really excited, and read it in the best-possible light.
Also, a 5% lead in a national poll is, honestly, meaningless. It’ll come down to the swing states, and in polls in those states where it looks like Harris has a lead today, nearly all of them are very close, and well within the margin of error for polling.
No magic necessary.
It’s the same story as the last few election cycles - the election will come down to thin margins across 3 or 4 states adding up to ~100k votes, even if the overall popular vote won’t be competitive at all.
And when the results will be that close, they’re well within the polling error margins.
Am I “really excited” about this new poll? No. I just think things are trending in the right direction, and I’ve thought that since I wrote the OP. And I don’t need to see a good poll in a good light–it’s just a good thing. Bad polls are bad things.
The context of 2016 and 2024 is totally different. If Trump were to win, we would not be surprised as we were in 2016, nor are we going to let that happen through lack of effort and discipline.
There are two polls I am seeing that are completely post-debate, Reuters and Morning Consult, and both have Harris up by five points. Sample sizes are large.
I think it highly likely this is a real debate bounce.It might not be durable. There could be a shy Trumper effect. But I’m reasonably confident that there was a debate bounce.
What I am extremely sure of is that the pollsters are doing their best to be accurate.
Even with a “large” sample size, results from political polls are still typically +/- 2 to 3 percentage points – and following your link, both of those polls are reporting an MOE, or “margin of error,” of 3 points.
That means, for example, that Harris’s 47% in the Reuters poll is projected to represent an “actual” score of somewhere between 44% and 50%, if you were able to survey everyone in the U.S. And, similarly, Trump’s 42% would be somewhere between 39% and 45%. This means that is absolutely possible that the actual margin, among the entire population, is less than a 5% lead for Harris – and it’s entirely possible that Trump is actually still ahead by a small amount. (And, yes, it also means that it’s possible that Harris’s lead is actually bigger than 5%.)
And that is why, yes, the election really, really, still is effectively a “coin flip.”
Maybe. But the previous results of those two polls were Harris +4 and Harris +3. So if it’s a bounce it’s in the 1-2% range. Could just as easily be noise. We need more post-debate polls to make any sort of conclusion.
If there is any indication of a debate bounce it sure explains why Trump wants another debate.
Is it significant that Harris seems to be on the up side of almost all of the polls that are within the margin of error? In other words, wouldn’t there be more Trump leaning polls if the error was random?
A good question. When you have a number of polls which are all giving similar results, it does suggest that the “real” number is probably something that’s pretty close to the poll numbers.
OTOH, as has been beaten to death on this board in recent years, political polling, as a whole, has real accuracy issues these days, due to sampling and contact techniques, increases in non-response rates, the possible phenomenon of “shy Trump voters,” and polling organizations which may be introducing their own ideological biases to their polls.
Speaking as a trained statistician and market researcher, as much as I would love to see good empirical evidence that Harris is doing well, I have to remind myself that trying to see what the electorate is thinking, via poll numbers, is like trying to study something through a dirty, distorted pane of glass.
We touched on it many posts back, but I think the “bounce” concept is dead. Certainly seems so for this election, perhaps going forward as well:
- No Trump post-assassination attempt bounce.
- No post-RNC bounce.
- Polls went up a bit after DNC; not clear that this was a “bounce.”
- Not even really clear how things “bounced” after the Trump-Biden debate.
Thus, I don’t know if Harris is getting a “bounce” from the debate. My guess is that the higher national polling will stick and not go down.
Also, it now seems pretty clear that Trump is banging Laura Loomer. Since Trump is out of control, he’s barely trying to hide it. He brought her to the 9/11 ceremony, and it seems that he’s with her all time now (not necessarily continuously but very frequently). I was wondering how Trump was getting laid with Melania manifestly despising him; well, that’s the answer. Ick for both of them! Ick ick ick!
Here is a NR article about how hideous Loomer is, which we knew, but it’s in the comments that the issue of the relationship is dealt with (provided as a cite; not saying you should read it).
(Trump was also openly macking on the Black chick who posed as a Black voter in support of him but who morphed into something like a staffer. Michaela something. Now she’s clearly an awful person but rather hot IMO. Loomer, noooooo.)
Yes it’s just two polls so far, but MC has consistently run H+3 and this really sticks out of its line. A bump of 2 if real is definitely more than I was expecting at least. Also yes it may be just be a bump and not stick … but this an unprecedented cycle and it might just last.
Yes if all we were concerned about was random error. The error of larger concern is systemic error, the sort that underestimated Trump’s vote the last two presidential cycles. It being three one direction or the other is a reasonable expectation.
Trends are believable though.
No, I’m sorry I was vague. I meant the system has to preserve the ability of the EC to act as a safety check for unqualified populists.
I have a question. Does anyone have any information or a link that explains how polls are conducted and analyzed to provide any sort of reliable results, even theoretically? I’m just wondering how data is collected, what steps are used to try to make sampling statistically meaningful, and how pollsters overcome the modern ability to avoid polls.
Thanks.
Due to the margin of error and the EC effect, it really is a coin flip right now, and there really is a large element of random chance because turnout is unpredictable, and the turnout numbers in battleground states will be crucial. This will remain true right up to election day unless Harris somehow gets a big boost. But the DNC and the debate are now behind us.
Thread is closed awaiting further review on this post by the OP.