> So, given this information, I would think that the numbers from 1 to 10, listed by popularity would be something like 7, 3, 9, 8, 4, 2, 6, 5, 10, 1.

You're making a mistake here: You're supposing most people think like you do. That may prove dangerous. (Well, not in this case, I guess...) You'll have to take the randomness of human psychology into account.

Many people would probably vote for exactly the opposite numbers, just not to not do anything normal. (Yes, both the "not"s are intended.) I suspect the extremes would score pretty high anyway for that reason.

But the premises change a bit when you increase the number of available choices:

> If there were a poll asking people to pick a number between one and one hundred, I'd be willing to bet a graph of it would appear low at both end and near the middle, and that there would be 'hills' around 25 and 75.

Personally, I think the "hills" would be closer to 38 and 62 -- around the golden section. (Which Monkeyman (and Fuzzpilz too) already has proved. ;-) There have been experiments conducted where classes have been told to put a mark on a plank of wood, dividing it into two parts, where they felt it was in balance and harmony. Most people did not put the mark in the middle, but around the golden section. I believe there would be a tendency towards the golden section also here.

But even with a hundred choices, the extremes would probably get more votes than others, just because people would like to be extreme, even though they wouldn't be doing anything original. (Again the "not not normal" thing.)


> I'm also willing to bet more people would chose odd numbers than even number. I'd even go farther than that and say that if you took the number of votes for prime numbers and divided it by total numbers of primes from 1 to 100, it would be more than the total votes for non-primes divided by the total number of non-primes from 1 to 100.

I don't think primes would show up more frequently than other numbers, unless people were given time to think first, that is. But even then, I'm not sure. (To me, most primes are "uncomfortable" numbers.) And I can see no reason why odd numbers would be more desirable than even ones. Personally, I think I'd go for even. It's in my nature. I like evenness. (Unless you ask for a "random number", that is. I think most people feel odd numbers are more "random" than even ones.)

If this poll were to be done, I think there should be two consecutive, equal questions, the only difference being that the first one is to be answered impulsively, while you're supposed to choose your number carefully the second time.

I'll gladly host a poll like this if anyone cares to put together a Perl script that lets the user enter a number from 1 to 100 and, as the RinkWorks poll does, keeps track of IP's so that people can vote only once.

Travholt.