Bsphil wrote:
Yes, it is. Polling 100x as many people is incredibly cost inefficient and the benefit is nigh microscopic. If it's close, then it's close. The poll is showing exactly how people actually feel: they're evenly divided on the issue. Again, deal with it. Those are the numbers.
Once again, I don't need a poll to tell me that the people are divided on the issue, just like I don't need a poll to tell me that people think rape is bad. That isn't my focus, once again, I'm looking to find out EXACTLY which side is above the other. If all your polls can do is tell me that the people are divided, then they are no good to me.
Bsphil wrote:
Why do you care so much about the word "majority"? It's about a split, slightly in favor of SSM. Stop thinking of supporting/not supporting as this massive switch in public opinion just because a different check box gets 50% + 1. In fact, why don't you do that for everything?
BECAUSE THAT'S MY WHOLE FREAKIN MOTIVE.. I want to know what the majority is. People like you and Belkira, will source these polls to say "The U.S. supports SSM" when in fact, that could very well be false.
That's how this whole thing started. I said that majority of the people don't support SSM and she responded with poll. That's how these polls are used in debates.
Besides, you haven't answered, how do you know it is slightly in favor of SSM and not the other way around? You admit that it's practically split, there is margin of error and you also admit that you could find 500+ people out of a 1000 person survey that DON'T support SSM, how can you say it's in favor of SSM? You can't! You're simply saying it is because that's what you want it to say.
Which was my initial argument. People like you will just continuously re-poll until it says something else. It wouldn't surprise me if they did that phone survey a couple of times first and disregarded all of the results that concluded otherwise and only published the one that was in favor of SSM.
Bsphil wrote:
Lulzy English aside, that should be even easier to find, then. Chop chop.
Here, I'll even save time and assume you're not just missing my point with your summary:
You're not flipping this around on me. You said it. I told you exactly where to find it, which isn't that hard since it was right above the text you quoted me from. If you don't want to address the issue, then fine.
Bsphil wrote:
If the poll is biased, then it doesn't matter how many people you asked, you're going to get slanted results. You could poll every single adult in the country, but if the questions are biased to subconsciously favor one answer, then even polling the entire US adult population will result in inaccurate poll data. This is why I said the size doesn't matter. It doesn't, ASSUMING YOU PICK A SAMPLE SIZE LARGE ENOUGH TO FULFILL THE WEAK LAW OF LARGE NUMBERS. That's why 10 doesn't count and 1000 does.
My counter to that was how did you assume that the poll was biased? It appears that you were claiming any poll opposing SSM as "biased".
I stated how random polls work. Random polls don't take in consideration Democrats vs Republicans or any other criteria, it's completely UNBIASED. Doing so is good in the sense of not being biased, but it also opens the door for the strong possibility of an unfair representation. If your poll randomly selects more Democrats than Republicans, then your poll will more than likely favor SSM. If your random poll selects more Republicans than Democrats, then your poll will more than likely not favor SSM. Being random is unbiased, but it isn't a fair representation of the U.S. MATHEMATICALLY speaking, in order to guarantee more coverage in a RANDOM poll, you need larger numbers.
Bsphil wrote:
So, let's review:
Difference between 300,000,000 samples and 1,000 samples = not important
Difference between 10 samples and 1,000 samples = extremely important
I can only assume your continued ignorance of statistics means that you didn't read through the wiki page on WLLN?
We're going in circles. You're right, I didn't read it, because I'm not denying that it supports the WLLN. What I'm denying is that it's enough people to answer the question that I'm wanting to know, "Do majority of the U.S. support SSM or not?". Once again, if the question was "Do you support rape", I'm sure every 1000 person poll would probably accurately represent the U.S. population. Given that this is a nearly split opinion, IT IS NOT ENOUGH TO KNOW THE ACTUAL ANSWER. That is proven by contradicting polls.
You keep trying to paint this picture that I'm arguing against statistics, when I'm arguing against the application of statistics.
Nadenu wrote:
ITT: Alma wants every single person in the country polled. GG.
Belkira wrote:
He has also argued a position that he has now admitted is impossible to hold.
Almalieque The Man of Pleasure wrote:
My goal is to logically, not scientifically, figure out what the ACTUAL majority believes, not a small sample. The only "scientific" way to do that is to poll everyone or the majority of the population where it is statistically impossible for the minority to overcome. As you know, that is practically impossible and or not feasible to do, hence the "1,000 person polls".
Samira wrote:
Better poll 'em every day, in case they change their minds.
ALmalieque The Awesome wrote:
Here's the thing, people's opinions just don't change in general. What happens is, generations have different opinions and as older people pass away and younger people grow older in a "newer" society, the general opinions change. You act as if 270 people just changed their minds from last month, which caused the overall percentage to tip. No, that's not what happened. Those 270 people had the same beliefs last month as well, they just weren't polled. A different group of people answer the question.
Wow, looky there.. Funny how reading works.