On: Than Average

[geraneum]

I think this experiment is about perception rather than finding the true statistical average. The sample space might not be balanced for studying poverty, intelligence, relationships, etc. but I believe that's not the point here. The test is trying to show you how is your perception of yourself compared to others who participated.

Take for example this question: "Are you smarter than average?" and (last time I checked) 72% think they are. Just imagine how the distribution looks like. By comparing this number to the IQ distribution (assuming correlation) which is a normal distribution, you could infer that either [0] a sizable portion of participants are wrong about how smart they are, [1] the sample size is too small, or [2] the people who've landed on the page are exceptionally smart!

[0] I have a "feeling" that this is the case!

[1] Maybe this can be taken into account with better statistical analysis.

[2] The assumption of participants coming mostly from HN, does not prove that this is the case.

[strken → geraneum]

There's probably a bias towards sightly higher than average IQ, because extremely low IQ usually means intellectual disability, and because IQ correlates with the ability to apply effort (at least at lower IQs). I also suspect but have no evidence that low IQ users tend to stay within walled gardens like Facebook more.

I think a lot of HN would be surprised at what an IQ of 100 looks like.

[LASR → geraneum]

Yes. But isn’t this point obvious?

There is no mathematical way to determine the true average of anything unless if there is some true measure that exists outside of perception.

So regardless of the sample size, you will never arrive at any true measure, unless if that measure is of the perception of self, on average.

Which is what this experiment is doing, as you’ve stated.