Ah. Generally when I am speaking to someone who I feel doesn't understand what I'm saying, I will make an effort to help them understand, and if that fails, I will simply remove myself from the conversation while trying not to be rude.
As for my lack of understanding of the statistics... I understand them just fine. It's really quite simple to find out exactly how large of a sample you need to achieve an accepted level of accuracy. I also repeatedly agreed that the sample size this poll has is almost certainly too low if you wish to get a result within that accepted level of accuracy.
If you really did understand the statistics, you wouldn't have been pulling meaningless numbers before. It wasn't until I introduced you to the concepts like confidence interval and degree of confidence. You may have worked for Gallup, but you didn't understand what was being done afterward.
Generally I'm a nice guy, in and out of game. But I dislike when people don't know what they're talking about -- not the people themselves, but what they're trying to say. I have no compunction to pointing things out as silly, inane, misguided, or plain untruthful.
There seems to be a simple disagreement on whether or not anything can be inferred if you do not reach that acceptable statistical level. To me a matter of do you want a specific representation with the highest level of statistical accuracy possible... or do you just want to make as the poll puts it "a guesstimate".
You're speaking nonsense here. There's no more disagreement here than on whether the sky is blue. This poll is as meaningless as a poll question, "How many foggy days do you think San Francisco will have next month?" None of the poll targets would have the vaguest idea, and those who would know aren't being asked.
Would I stake my professional name or my or anyone else's livelihood on a poll with an accuracy level of 85% (or whatever)? Of course not. Would I call 85% (or 80% or 75%...) "good enough" if I were limited to an undersized sample pool in a poll regarding a hobby or a game I played the general outcome of which is unsubstantial to any involved? I would, as I think would most. Thats really all I have been trying to say.
Don't you want to go by something accurate, or are you content to go by approximate numbers? What if your auto's gas gauge was plus or minus a quarter tank. Good enough? When I did some baking earlier, there was no "good enough" when measuring shortening or sugar, but quite precise measurements. It's actually quite easy to be well within ±1% of a cup.
I wouldn't mind a UO poll with reasonable accuracy, but this poll was silly just from its question. Until we're given numbers that can be believed as realistic, I'm going to call out such things for the absurdities they are.
Surely one can question the value of this poll,
Then why are you defending it at all?
but personally I think the value is inescapably subjective. While it may mean nothing to you or me, it may mean something to the post creator, or some other person participating. I don't think its simply pointless or not.
And you're going in circles again. In one post you've support the poll, in another you've said it can't do what it aims to. Now you're doing it in the same paragraph.
And you don't "NEED" those numbers to make a general guess. If the poll swings 5% it doesn't really matter. It's not an election. It's a video game. And I agreed with you that 90-100 sample size will have a wide swing range, but it will still be around 85% accuracy.
