Apparently, if I go around asking my friends for a strand of their hair, about half will just give one to me, and almost all the rest will give one when I explain that I'm going to see how thick it is under a microscope. Your mileage may vary.
My parents have not yet chanced upon the card with taped and labelled hair strands, but if they do then they might have some, uh, pointed questions. I'll just tell them I'm trying to diversify my clone army.
Hey, at least it's not like that time a friend of a friend's ex once drew someone's blood at a party.[1] Unfortunately, I never got around to asking her for one of the blood vials she was known to possess.
Methods
To measure length, I broke off a tiny piece of a clear ruler and put it under the microscope with the hair. Then, I had Free Claude (from 2025?) write some code so I could compare two lengths easily by clicking points. Then I modified some Claude analysis code because I was too lazy to write it myself.
Here's what mine looked like (At the time I had some black hair and some bleached bronze-ish hair, and both seemed about as thick, so I just used the latter):
The reference is 1 mm, so my hair was about 85 microns thick.
Results
Here's an anonymized bar plot, where my id is 16. Id 5 and id 6 come from the same person, but the 6th was a sample of hair dyed red and the 5th was of black hair. The colors are according to biological sex (most are trans women because friendship selection pressures):
The mean's 78 microns and the (Bessel corrected) sd's 12. That's in line with this international study[2] (I live in America and most participants were some flavor of white):
Male group got , female group got , doing a (Welch) t-test for the difference. The difference of 1.3 microns has a Welch's t statistic of with degrees of freedom, which gives a large 95% CI of about and a p-value of . Personally I suspect some assumption is broken that makes the CI quite a bit tighter. Regardless, they have basically the same mean.
Let's keep trying to make a mountain out of a molehill of data and test the male variability hypothesis, using a Brown-Forsythe test. The W statistic is 0.735 and p = 0.4. Running some magical bootstrap code that Claude gives me spits out a 95% CI for the sd ratio of [0.661, 5.391]. Wow, doesn't this nothingburger taste like nothing?
Here's the empirical distribution function in blue, compared to a normal distribution in green, and confidence bars from some magic frequentist statistic gadget that I didn't end up caring about because clearly it wasn't saying anything interesting.
Now let's eyeball the result of a related frequentist gadget! To do a Kolmogorov-Smirnov test, we take the max difference between those two curves, multiply by the square of the sample size, and then we'll get something that in the limit looks like this magic distribution:
(there's a magic correction[3] to speed up the convergence that I'll actually be using).
The max difference looks to be at about the part at 82% on the x-axis, and the difference looks to be about 15 percentage points. So sqrt(16) * .15 = 4 * .15 = .6, magic calculation says that the area under the curve (adjusted by magic to become .6354) is 0.186, which means that if we truly had a normal distribution there's about an 19% chance we'd get something smaller than this, which is small, which is (checks notes) good.
Conclusion
The person with the thickest hair said something like "That's good, I think?". The person with the thinnest hair happens to be pretty shy, so she was probably embarrassed.
A study with a larger sample could try getting actual results for the sex differences, could try comparing bleached vs unbleached hair, dyed vs natural, natural hair color, and region of the head/body. A study with a smaller sample would be faster to conduct. A study with an imaginary sample would help you test your statistical methods and nothing else. The generalization to quaternionic sample sizes is left as an exercise.
For some reason everyone asks whether I collected pubic hair. I can neither confirm nor deny that pubic photons have reached my retinas. I will gladly look at your stats if you want to know how they compare.
In the table they sum up the sample size among all ethnic groups, and then for some dumb reason they also sum up the... hair diameters... and color tones... and curliness...
Intro
I have a microscope. It can see hair.
Humans have hair. Some of my friends are humans.
How thick is my friends' hair?
Apparently, if I go around asking my friends for a strand of their hair, about half will just give one to me, and almost all the rest will give one when I explain that I'm going to see how thick it is under a microscope. Your mileage may vary.
My parents have not yet chanced upon the card with taped and labelled hair strands, but if they do then they might have some, uh, pointed questions. I'll just tell them I'm trying to diversify my clone army.
Hey, at least it's not like that time a friend of a friend's ex once drew someone's blood at a party.[1] Unfortunately, I never got around to asking her for one of the blood vials she was known to possess.
Methods
To measure length, I broke off a tiny piece of a clear ruler and put it under the microscope with the hair. Then, I had Free Claude (from 2025?) write some code so I could compare two lengths easily by clicking points. Then I modified some Claude analysis code because I was too lazy to write it myself.
Here's what mine looked like (At the time I had some black hair and some bleached bronze-ish hair, and both seemed about as thick, so I just used the latter):
The reference is 1 mm, so my hair was about 85 microns thick.
Results
Here's an anonymized bar plot, where my id is 16. Id 5 and id 6 come from the same person, but the 6th was a sample of hair dyed red and the 5th was of black hair. The colors are according to biological sex (most are trans women because friendship selection pressures):
The mean's 78 microns and the (Bessel corrected) sd's 12. That's in line with this international study[2] (I live in America and most participants were some flavor of white):
And look, 5/16 or about 31% of my
victimssubjects are outside of one sd of the mean, and 1/16 or about 6% got quite close to two sds. I guess the statisticians weren't lying after all!Male group got
Let's keep trying to make a mountain out of a molehill of data and test the male variability hypothesis, using a Brown-Forsythe test. The W statistic is 0.735 and p = 0.4. Running some magical bootstrap code that Claude gives me spits out a 95% CI for the sd ratio of [0.661, 5.391]. Wow, doesn't this nothingburger taste like nothing?
Here's the empirical distribution function in blue, compared to a normal distribution in green, and confidence bars from some magic frequentist statistic gadget that I didn't end up caring about because clearly it wasn't saying anything interesting.
Now let's eyeball the result of a related frequentist gadget! To do a Kolmogorov-Smirnov test, we take the max difference between those two curves, multiply by the square of the sample size, and then we'll get something that in the limit looks like this magic distribution:
(there's a magic correction[3] to speed up the convergence that I'll actually be using).
The max difference looks to be at about the part at 82% on the x-axis, and the difference looks to be about 15 percentage points. So sqrt(16) * .15 = 4 * .15 = .6, magic calculation says that the area under the curve (adjusted by magic to become .6354) is 0.186, which means that if we truly had a normal distribution there's about an 19% chance we'd get something smaller than this, which is small, which is (checks notes) good.
Conclusion
The person with the thickest hair said something like "That's good, I think?". The person with the thinnest hair happens to be pretty shy, so she was probably embarrassed.
A study with a larger sample could try getting actual results for the sex differences, could try comparing bleached vs unbleached hair, dyed vs natural, natural hair color, and region of the head/body. A study with a smaller sample would be faster to conduct. A study with an imaginary sample would help you test your statistical methods and nothing else. The generalization to quaternionic sample sizes is left as an exercise.
For some reason everyone asks whether I collected pubic hair. I can neither confirm nor deny that pubic photons have reached my retinas. I will gladly look at your stats if you want to know how they compare.
Ceterum autem censeo frequentistae esse delendam.
This led the host to institute the first party rule: do not draw blood. Don't worry, it was consensual... though the drawee seemed drunk.
Thus the principle that if I am the voice of caution, you should run.
In the table they sum up the sample size among all ethnic groups, and then for some dumb reason they also sum up the... hair diameters... and color tones... and curliness...
Using the one given by Wikipedia, where you replace
which has a max error of like .2%.