It's a request, not a demand. I even said please. I have said before, and I mean it, if you can show a legitimate use for these data I would gladly support your request. You may (and will, likely) ignore my request, but frankly you act against your own interests in doing so. I honestly think that the data are inappropriate for any use beyond their challenge purpose (for reasons of self-selection, and of the bias that may emerge in combination, as previously explained); if you honestly think they are of use, I have no clue why you would not say how. Claus dismisses you out of hand; I am more patient. I don't care what I might think of you, if you have good ideas, I am for them. If you do not share them, I cannot support them. Your loss.
Willing to think? Certainly. Willing to suspend what I know about biased samples? Sorry, no. Thinking about how test results relate means critically examining the possibilities. If the data don't work for a particular purpose, it is better to obtain good data than to argue about bad data.
But of course, that number is meaningless. It does not tell us if there are more dowsers, because we simply cannot know if the reasons for self-selection were constant, or whether there was some unrelated reason that there were more dowsing challenges. In fact, if we picked up an apparent trend that the dowsing challenges were increasing, it would tell us nothing; is it a representative sample? We simply cannot know--not only why there were more dowsers--but even whether there were more dowsers, in anything but a self-selected sample. An ad in a dowsing website could easily throw off any claims of "trend", and no one has control over that.
Well, if I had a question to answer, I would have a much better idea of how to collect the data. That is pretty much the whole point of the "hypothetical as yet to be collected datasets"; they are collected to answer a particular question. Obviously, you have some reason to keep looking at the data which have already been collected, and I cannot fathom why. What question can be answered? You will not help with that, and I ...apparently...lack imagination.
So you're not going to do this for me? Why not?
I am sorry that you see scientific inquiry as spanish inquisition. You see no point? Funny, I thought I was the one with no imagination.
This has been addressed. A faulty answer is as bad, or worse, than no answer. What does the self-selected sample give you that a proper sample would not? There are, of course, many things that a proper sample would give you that a self-selected sample will not. Even your level of understanding of statistics and research methods will show that.
Yes. It is not a representative sample.
The data we collect on those are intended for different purposes. And are collected with those purposes in mind. Unless you wish to suggest that a self-selected "reportyourweather.com" will be a better source of data than systematically collected data, these examples are useless for your argument.
So far, the answer is "no, there is no reason to look at these data."
I'm curious - does this sort of "screw you, answer my questions!" approach work for you normally? It seems a very strange way to elicit answers.