I'm the person who plays the vielle, and it looks like the second of the pictures ShiroKuro posted...although mine looks a lot less fancy!

Brilliant!

casinitaly, the Central Limit Theorem basically says that the accuracy of a randomly sampled poll depends on the number of respondents only, and does not depend at all on the total size of the whole population.

So whether there are 590 or 59,000 members of PW, it doesn't matter. Once you've got 150 replies to the poll, you've got a good level of accuracy. Assuming of course that the responders are representative of the whole group (that is, equivalent to randomly chosen).

I think this is one of the most brilliant theorems in all of statistics.

I can get my head around that. Thank you

Great!

A quick addendum: the closer the results are to a 50-50 split, the more replies you need for accuracy. But it still doesn't matter if there are 590 or 59,000 in the total population.

I quite agree, but the methodologist in me is compelled to point out that the words randomly sampled are absolutely critical here. The extent to which one can generalize from a sample to the population depends completely on how it was sampled. And while, as PianoStudent88 points out, a random sample of 150 people would be more than enough to be representative of AB forum as a whole, the fact remains that these poll results are from a self-selected sample. As such, we can generalize only to the population of AB forum viewers who choose to answer polls.

In other words, Sam's data are an interesting snapshot of the 150-odd folks who took the time and effort to respond to the survey, but we can't assume they tell us anything about the people who didn't.

There are also these problems with this simple poll:

- you can take it multiple times
- anyone can take it if they can find the link

So it's a fun poll, and my gut feeling is that it's pretty representative of us, but it can't stand up to rigorous scrutiny.

And now that I have revealed those flaws, please don't send the link to the poll to all your facebook friends that have never been to the ABF forum!

Sam

My gut feel is that most of the participants were "regulars" - like myself, Monica, cas, etc. and hence a fairly accurate snapshot of the ABF. I wonder then if anyone from the other forums (like Pianist corner) came in a also filled out the survey.

Sam, I'd suggest to re-open a fresh new poll and see if we can get another 150 "fresh" replies. We all need to use the honor system so don't "retake" the new poll.

I think this survey was fine for our informal purposes. I would suggest waiting for this topic to disappear way down the list of topics before starting another one - maybe next year. We could come up with better questions too.

Sam

A better way to get a random sample is to pick random 150 persons from the database and ask them to reply? Yet even then you will have only those that reply. And given that half of the people here never have posted anything the percentage of replies might be very small indeed

I find I'm more interested in the replies of those who do come and post here regularly enough to have spotted the poll. So the 150 self-selected at-least-semi regulars who replied are exactly the people I want to know about.

I will look forward to next year's poll. I answered one thing wrong, I realized afterwards, but I won't cheat and fill it out again . I said I only knew theory from a college course, but I forgot I self-taught myself a lot of theory from Edly's Music Theory for Practical People. Edly's is what made the three forms of minor scale make sense, and the relation between major and minor scales and other modes. The college course taught me all I know about tonal analysis, up through secondary dominants. Despite my years in band and chorus, I never learned any theory there. (Well, I guess I learned to read high treble leger lines from playing flute, but the rest of reading music I taught myself, and when I think of music theory, I usually think of the parts beyond simply reading the notes on the page.)

Hmmm...let me make one more quick observation that I just noticed this morning. During normal waking hours, there are usually 130-160 people actually viewing (or active) in the ABF. Thus it makes sense that the number of respondents falls right in this range.

Normal waking hours for. . .Who? The continental U.S.? Morning people? Night people? People who aren't on during working hours? We may be largely U.S. members, but it seems to me the rest of the mix is pretty wide. One of the people I PM on the most is almost never on when I am - besides the time difference, she's a morning person and I'm a night person. And I'm on more during my working day more than she is.

So I'm not sure there's a correlation. Maybe. But not sure.

Cathy

I have to agree with Cathy - Being one of those who is GMT+1, my "normal waking hours" are wildly different than quite a few members here, but I also see that when I'm up and on line at 6 or 7 am (yes, that can happen, not that often, but it does)..... there are still a lot of American/Canadian night owls on-line.

In our next survey maybe we can include a question on which time zone members are from!

Yes, and while I'm thinking about it, because I will forget before we decide to do the next survey, here are some more questions to consider:

- time zone
- hours of practice daily
- electronic or acoustic?
- gender
- playing level self-assessment
- favorite genre
- can you play by ear?

Any others?

Sam

Funny, but on my way home during my lunch break 12:00PM CST US, I predicted that there'd be people poking holes on this post from this morning 7AM CST US because I know this is a global forum and "normal waking hours" don't apply to everyone. Anyway.... At this hour, 12:30PM CST US, there are currently 180 members online!

Now...i've been on this forum practically all hours of the day and night (occassional insomniac), and noticed during the wee hours of the night, the active members usually slips to below 100.

I guess the original intent of my post --- isn't it cool that the number of respondents roughly equals those who are online during "peak hours".... this confirms the "central limit theorem" and also confirms Monica's assertion that respondents are more-than-likely regular and active members of the ABF.

Sam, a re-survey would be great.

I don't see any confirmation of the central limit theorem. (Nor disconfirmation either, of course.)

I thought as we were coming up to 1 year since we did this, it might be fun to do another one.
(Said she, who only has to fill in the answers and not do anything else )

Ideas for new questions=
Did you participate in the last survey?
How many recitals (ABF) have you played in?
Have you played in for a public
--- if yes: live recitals ?
house parties?
nursing homes?
volunteer gigs?
paying gigs?

oh, and as I have just re-read the part where anyone can do the survey - maybe ask :

Are you a member of PianoWorld?

Thanks for reminding me! I actually have a survey sort of ready to go that I added some questions to - I'll get it updated.

Any other questions that should be included?

Sam

Here are a couple of suggestions for the poll:

--I'd be interested in what the responder's current number of posts is. Zero posts = lurker; 10,000 = long-timer with a lot to say. At any rate, one can make some inferences from the number of posts.

--If the responder takes lessons, the cost per 30 minutes. (If they take longer lessons, do the conversion to cost/30 min. Use $ if you know it, otherwise your local currency and someone can do the conversion.)

Whether the person mostly plays on a digital or an acoustic piano.