[ SECRET POST #6330 ]

[case]

⌈ Secret Post #6330 ⌋

Warning: Some secrets are NOT worksafe and may contain SPOILERS.


01.



__________________________________________________



02.



__________________________________________________



03.



__________________________________________________



04.



__________________________________________________



05.



__________________________________________________



06.

















Notes:

Secrets Left to Post: 02 pages, 29 secrets from Secret Submission Post #905.
Secrets Not Posted: [ 0 - broken links ], [ 0 - not!secrets ], [ 0 - not!fandom ], [ 0 - too big ], [ 2 - repeat ].
Current Secret Submissions Post: here.
Suggestions, comments, and concerns should go here.

Comments

Let's talk numbers

[(Anonymous)]

Okay, so there are about 336 million people in the US, and roughly 162 million who are registered to vote. Gallup Polls typically have a sample size of 1000. I believe that's common to a lot of overnight polls. Longer polls, taken over weeks, could have a sample size of tens of thousands. But let's say that you had 10,000 respondents, that would still be a fraction of a percent. They do try to get representative samples, but it's all based on self-reporting. There are margins of error that are given, but I'm not sure some things are truly accounted for or even could be. And that's not even thinking about the way the questions are asked.

Re: Let's talk numbers

[(Anonymous)]

I mean polling as a general industry has a lot of demonstrated successes. It's pretty well established that, in principle, using relatively small sample sizes to understand the opinions of larger groups is something that works.

It's possible that there are specific problems with polling methodologies that are being used right now. In particular, it's possible that the way people communicate have changed and so pollsters are now unable to actually contact a representative sample of voters. But polling fundamentally can work in the right circumstances, even if you're only asking 1200 voters or whatever.

Re: Let's talk numbers

[(Anonymous)]

There is such a thing as a sample that's too small, but after *n* gets above a few hundred, what really matters is whether it's representative of the general population.

Let me give you a quick and easy US-based example:

You've managed to get a whooping 50 million respondents to your poll. However, nearly all of them are white, Republican-leaning men.

Your competitor gets only 1000 respondents to her poll. However, she carefully selected her respondents, making sure that the demographics of her sample roughly match those of the broader US: i.e. the sample is about 50/50 male/female; a third identify as Democrats, a third as Republicans, a third as independent or something else; 12% are Black; 7% are LGBT; etc. She also solicited respondents from each region of the country.

Of these two polls, which is likely to be more accurate?