Is 17 the most random number between 1 and 20?

[pne]

A couple of days ago, I asked people to pick a number between 1 and 20 (supposed to be inclusive, though perhaps that wasn't quite clear), in order to find out whether 17 is the most random number.

Turns out that indeed, more people picked 17 than any other number: 12 out of 68, or 17.6%. The runner-up was 7, with 8 out of 68, or 11.8%. Most of the other numbers scored 1 or 3 votes.

Since the average number of votes for each number, if the 68 people had truly picked at random, would have been 3.4 if 1 and 20 were counted, or 3.8 if they weren't, having 3 be the most common number of votes is consistent with that. Interesting, then, that 7 and 17 scored so many more votes.

Comments

[blue-hours-too.livejournal.com]

That's really interesting. I must be an alien because 17 doesn't appeal to me at all.
3, 7 and 11 do, so does 13.

[(Anonymous)]

This may be because 3, 7, and 13 are viewed as lucky or unlucky numbers, eleven is one of the three numbers left in the mix if you remove prime numbers and "lucky" numbers, which makes it a less predictable random number, that could explain why it appeals to you.

[diffrentcolours]

Question is, how biased is your sample set? I didn't participate in the poll because I knew it referred to the "17 is the most random number" phenomenon which I'd recently seen linked from someone's LJ. I wonder how many people who took the poll were similarly aware.

[pne.livejournal.com]

Well, if anything, I would have expected such biassed readers to specifically pick a number other than 17.

[diffrentcolours]

I'd expect them to be biased towards 17 myself.

[spodlife]

I was also biased not to vote. Unfortunately it is difficult to repeat the experiment with different number ranges because more people will be wise to the nature of the test.

[ubykhlives.livejournal.com]

I've noticed a couple of other interesting things about these data (which you may have noticed, too!).

- Aside from unity, the numbers that were chosen the fewest are all divisible by five, the number of fingers on one hand.
- The five numbers that were chosen more than the median number of times (median number of times selected = 3) are all odd: 17, 7, 19, 11, 3.
- Moreover, these five numbers are all prime.

The fact that the two highest choices both contain a seven might be a statistical artefact; perhaps "select a number from 1 to 100" would be a more useful exercise to determine whether there's a bias towards the digit 7.

[arthur-sc-king.livejournal.com]

I didn't look closely at the results, but I suspect that this (and most similar tests, at least with numbers under 100) would be biased towards prime numbers.

I can't explain it, but I suspect that when asked to pick a "random" number, most people would shy away from composite numbers.

And, of course, now that I'm looking up at ubykhlives's comment for the first time, I see that my suspicions played out with this poll, at least....

[bluewingedcat.livejournal.com]

Woo! I was the ONLY one that answered 20! Unique and beautiful snowflake go me!

[nyssa.livejournal.com]

This is an interesting post; I might do a poll in my LJ and see what kinds of responses I get. My flist isn't so great on reading things like this. ;) I might get a better sample.