Idk man, our confidence intervals don't mean much if we're using a biased sample. I think the best way to gather data like this would be to ask homebrew clubs to do poll, I think we'd get a more representative sample of the average homebrewer.
Your assuming I'm interested in the average homebrewer. Actually, I'm not (well I am but not for the present purpose). I'm doing it to help John and Colin with their book. The kinds of people who are likely to buy their book are the sort who are likely to participate in a forum like this. IOW I'm interested in a population that has more intense interest in home brewing than the average homebrewer.
I accept that this population is biased (in, as it turns out, exactly the direction I want it to be) and am trying to estimate a statistic based on a sample. I want to know not only what the most probable value for that statistic is but how it is distributed.
This is not to say that data on the universe of homebrewers wouldn't be interesting but obviously I'm not going to get that. And, of course, we'd need to recognize that data from a home brew club in Texas would be sharply biased relative to data from a homebrew club in the Pacific North West.
Overnight another 9 people voted and that reduces the confidence bounds somewhat. What is interesting is that the average use or plan to use fraction has hovered right around 2/3 from the beginning. Here are the latest numbers:
Most likely Probability: 67.29 percent
90.0023 percent confidence interval: 59.91 to 74.67%
95.0026 percent confidence interval: 58.49 to 76.09%
99.0002 percent confidence interval: 55.70 to 78.88%
99.9000 percent confidence interval: 52.34 to 82.24%