Quote:
Originally Posted by Delta  I'll say it again -- IF the sampled fraction is small, THEN the simplied math applies in which population size is "completely irrelevant to the accuracy" of the poll. And that's the case we're actually in, the case misunderstood by the prior poster, the case of any standard poll.
Whether the population of ENWorld is 25,000 or 75,000 or 1 million or infinite, the math is exactly the same in any of those cases. Any larger population number disappears from the calculation formula.
Which is why I was the one who brought up the above-5% correction factor in the first place. You want a small sample fraction, such as we have, to use the standard math, in which population size is irrelevant. |
Just to be clear, you use the word "accuracy", but that's not true. The "math is simpler" with a smaller percentage simply because those polls have a bigger margin of error. A larger percentage poll is more accurate due to a smaller margin of error, but then the m-o-e must be calculated with more complexity.
There is no margin of error in the EN poll, and it's not a representative sample or a valid poll for any use other than "hey, I was just curious". The fact that Drader's remarks are basically "I know internet polls are invalid, but this one proves I'm right!" is all I'm commenting on.
Either way, a smaller margin of error means more accuracy. A larger percentage of the base will result in a smaller margin of error and more accurate results, but the math will be more complicated.
And yet another word from our sponsors
