- The Harris campaign is showing new strength in must-win states ahead of the party’s convention.
- In Michigan, Pennsylvania, and Wisconsin, Harris leads Trump 50% to 46% among likely voters.
- It’s a reflection of the continued reset of the 2024 race after Biden’s exit.
We don’t know the sample size; so we don’t know the margin of error.
I mean, click a couple links and it’s right there
MI: 619 PA: 693 WI: 661 All of registered voters
Using the amount of total registered voters in each respective state and a 95% CI, we get the following margins of error MI: ±3.939% PA: ±3.723% WI: ±3.811%
Depending on the exact lead (NYT only shows round percents, not specific numbers for each response), all of those are potentially within the top end of that margin of error.
Am I trying to claim that a swing from being down by ~4% to being up by ~4% means nothing and is indicative of nothing? Of course not. But man, most people really do not at all understand how statistics work, and I really wish people would stop talking out of their ass about it.
It was paywalled for me 🤷♂️
“I don’t see it” =/= “the information doesn’t exist and you don’t know so I’m right”
Next time try this asking.
Next time try to not take a statement as an insult.
I’d like to reiterate my larger point but I could’ve done it less abrasively, yes. Sorry about that.
So which links did you click? The one that goes to NYT is paywalled.
NYT polls aren’t pay walled
https://www.nytimes.com/interactive/2024/08/10/us/elections/times-siena-poll-likely-electorate-crosstabs.html
It’s easy. When Kamala is down we say that polls don’t matter as much they used to, but when she’s up polls are obviously right. The margin of error is just a thing we use after the fact to justify whether the polls are useless (Kamala losing) or absolutely correct (Kamala winning)
If I remember this correctly, the square of the error for the sum of (or difference between) two independent measurements is the sum of the squares of the individual errors. Gauss something.
That would make the error for the 8 point swing be sqrt(2×3.8²) or about 5.4. So at least the swing is significant in each state.
Also, the error for the average of 3 variables is sqrt(e1²+e2²+e3²)/3 or 2.2 so the average lead in the 3 states is significant.
But we can’t make a significant claim about the lead in each state.
I love how any time someone wants to argue about stats they act like this information can’t possibly be obtained and talk as if the sample size must be like 7 people.