I understand the idea of statistics but I do question calling 1006 people randomly on the phone and asking them, then applying the results to 150,000,000 people. But more importantly, national polls don’t take the electoral college into account.
Yeah, there’s a lot less randomness to it than you seem to think… Anything over a thousand is statistically usable, though not definitive.
That being said, there IS some inherent bias in the methods of some pollsters, such as calling landlines and only talking to people who pick up the first time, both of which skews the people polled older and more conservative than the general public.
Assuming for a moment that they got the popular vote as it stands now right, though, the EC is likely even worse, since it’s very much stacked in favor of the GOP…
Would the threshold of 1000 people be different if it was intended to represent say, 2,000,0000 vs 150,000,000? Why is 1,000 the magic number there? It seems like it would become more reliable with a larger sample size. And yeah, the population of people who answer unknown numbers with no warning and are willing to take a poll probably doesn’t represent the entire voting public accurately.
The EC is stacked in favor of the GOP due to smaller rural states, sure, but the extra dumb thing about the way it works in the US is that only about 10 states really matter. The rest are, of course, considered a sure win for one side or the other, and since the system is winner-takes-all, it’s not even worth campaigning or caring about how people on the other side vote in those states.
Believe it or not, the threshold for a reliable confidence measure can be a lot lower than you think. And the difference between a valid sample size for 1,000,000 people versus 350,000,000 people is almost certainly zero. Once you are confident in a result concerning many thousands, that usually extends to effectively everyone.
If you want 95% confidence with an interval of +/- 5% for every human being on the planet, your sample size only needs to be 385 participants. That gets way larger really quickly where you need to consider cohorts like gender, geography, age groups, income, and so on, but for a generalized read you need surprisingly few people. A higher confidence and smaller interval also increase the number significantly.
1000 is pretty much a compromise between the ideal and what’s realistically possible. Most opinion polls have a sample size between 800 and 1500, with those under 1k considered much less reliable, around 1k pretty much standard and 1500 extra rigorous.
There’s a lot of technical details on how to select those 800-1500 people to be reliably representative of a much larger population and different poll takers use different methodology, but that’s all too deep in the weeds for a lemmy reply and some of it is outside the scope of my knowledge as a curious and thus pretty well-informed layperson.
And you’re absolutely right about the EC on all counts. That and the filibuster are both examples of a supposed democracy being EXTREMELY undemocratic.
I guess for more info about the mathematics they use, I could ask my uncle - he’s a university statistics professor.
Yes, also the Senate was specifically designed to be undemocratic as to not give larger states more representation than smaller ones. At this point it’s absurd that states with 12 million people have the same representation as states with 800,000.
It’s a ridiculous compromise. It’s all they can do, but it’s just wrong.
And that’s not even the main problem; the main problem is with the reporting of it. That’s where the wrong end of the stick is used to beat around the bush.
I understand the idea of statistics but I do question calling 1006 people randomly on the phone and asking them, then applying the results to 150,000,000 people. But more importantly, national polls don’t take the electoral college into account.
Yeah, there’s a lot less randomness to it than you seem to think… Anything over a thousand is statistically usable, though not definitive.
That being said, there IS some inherent bias in the methods of some pollsters, such as calling landlines and only talking to people who pick up the first time, both of which skews the people polled older and more conservative than the general public.
Assuming for a moment that they got the popular vote as it stands now right, though, the EC is likely even worse, since it’s very much stacked in favor of the GOP…
Would the threshold of 1000 people be different if it was intended to represent say, 2,000,0000 vs 150,000,000? Why is 1,000 the magic number there? It seems like it would become more reliable with a larger sample size. And yeah, the population of people who answer unknown numbers with no warning and are willing to take a poll probably doesn’t represent the entire voting public accurately.
The EC is stacked in favor of the GOP due to smaller rural states, sure, but the extra dumb thing about the way it works in the US is that only about 10 states really matter. The rest are, of course, considered a sure win for one side or the other, and since the system is winner-takes-all, it’s not even worth campaigning or caring about how people on the other side vote in those states.
Believe it or not, the threshold for a reliable confidence measure can be a lot lower than you think. And the difference between a valid sample size for 1,000,000 people versus 350,000,000 people is almost certainly zero. Once you are confident in a result concerning many thousands, that usually extends to effectively everyone.
If you want 95% confidence with an interval of +/- 5% for every human being on the planet, your sample size only needs to be 385 participants. That gets way larger really quickly where you need to consider cohorts like gender, geography, age groups, income, and so on, but for a generalized read you need surprisingly few people. A higher confidence and smaller interval also increase the number significantly.
1000 is pretty much a compromise between the ideal and what’s realistically possible. Most opinion polls have a sample size between 800 and 1500, with those under 1k considered much less reliable, around 1k pretty much standard and 1500 extra rigorous.
There’s a lot of technical details on how to select those 800-1500 people to be reliably representative of a much larger population and different poll takers use different methodology, but that’s all too deep in the weeds for a lemmy reply and some of it is outside the scope of my knowledge as a curious and thus pretty well-informed layperson.
And you’re absolutely right about the EC on all counts. That and the filibuster are both examples of a supposed democracy being EXTREMELY undemocratic.
I guess for more info about the mathematics they use, I could ask my uncle - he’s a university statistics professor.
Yes, also the Senate was specifically designed to be undemocratic as to not give larger states more representation than smaller ones. At this point it’s absurd that states with 12 million people have the same representation as states with 800,000.
It’s a ridiculous compromise. It’s all they can do, but it’s just wrong.
And that’s not even the main problem; the main problem is with the reporting of it. That’s where the wrong end of the stick is used to beat around the bush.
States with 38 million people have the same representation as states with 600,000, even.
deleted by creator