I'm not talking in circles at all. I've said the same thing like 3 times now. It depends on what one want to accomplish with this poll. If you want to know how people feel about the population, it can give you that indication. If you want to know how many people are playing it cannot, and the poll is pointless toward those ends.
Yes, you have been saying the same thing, over and over, all in circles. What is it you want, and don't you realize the very question of the poll is meaningless? People are pulling numbers out of the air and elsewhere. You'll notice I haven't even bothered to give a number, except to show what sample size is needed for such and such a population. My home shard is pretty quiet, and I'm rarely on Japanese shards during their primetime.
I know that. I was saying 10000 as a random number to make the point that you don't need the exact number to come up with a fairly accurate sample size, you just need to know a rough range. UO somewhere within the 5000-200,000 range (being generous). So you need like 600. Though I think 300-500 would do just fine, since the outcome is not THAT important. Like if they were polling on "Would you rather have us work on a or b" a 5% swing is no big deal.
As I said I would look for 500ish off the top of my head, and the numbers would say around 600. So while 90-100 sample size will not give you the highest accuracy, I would be willing to bet that its not THAT far off from the target 600. Or if you want to widen the swing a bit 300.
Don't use random numbers. If you really wanted to be scientific about this, if you really wanted to understand the underlying statistics, you'd know you can't just go off the top of your head. That's why a prominent political strategist of certain notoriety got the 2012 presidential election all wrong, when prediction markets again were on the money.
Even if UO had just 5000 players, a mere 87 people is barely enough to get
±10.5% with a 95% confidence level, which is pretty bad on top of the meaningless question. To get ±4% with a 95% confidence level, a population of 5000 still requires 536 poll subjects. The sample size, as I said, growths logarithmically.
So as I've been saying all along, simple statistics shows that 87 respondents are completely meaningless, on top of a meaningless question. That's all there is to say.
If you are looking to find out how many people are playing, a poll is not going to do that for you. I don't think the question is "pointless", as I think it would be a some indicator of how confident players are in the game's future, which is important.
And if only a few had replied, would that mean we should be lemmings and close our accounts simultaneously? I prefer what Petra and Goldberg said, keep playing and have fun. As long as my guild and our allies are around, I'm happy.
If the poll is a matter of determining people's confidence in UO's longevity, then the question needs to address
that, not some silly question that starts from nowhere and goes nowhere. You say you worked for Gallup. Do you understand why polls directly asked people who they would vote for, not who they think would win?