[quote=“kilozombie, post:60, topic:2974”]Lazy’s actually got a point on the second part:
No, he doesn’t. Making a “poll” into a statistically valid sample requires very careful controls to prevent bias. In this case the sample is tiny, self-selecting, and more strongly self-selecting for people with a particular agenda.
All you can say (maybe) is that 32 people say they want an option, or at least did when they answered as such. There isn’t sufficient rigor here to project that result across the entire forum population, or even a significant subset of it. For one thing the number of people who viewed the poll dwarfs the number who answered at all, indicating that the vast majority is actually “don’t care enough to answer”. This is a problem systemic to forum polls, if the participants of a poll self-select, and even more so if they have a string stake in the outcome, it throws all statistical applicability right out the window.
Specifically regarding, “expect next 50 people will be much different on average”, actually I do expect that the next 50 people would answer differently, because the only people with a strong incentive to answer are in the “yes” group, and that group is tapped out, whereas the “don’t care” group is huge and underrepresented in the poll.