At first I saw that IQ seems to correlate with less children (a not uncommon observation):
Number of children/ACT score: -.279 (269)
Number of children/SAT score (2400): -.223 (345)
But then I noticed that number of children obviously correlate with age and age with IQ (somewhat):
Number of children/age: .507 (1607)
SAT score out of 1600/age: -.194 (422)
So it may be that older people just have lower IQ (Flynn effect).
Something to think about:
Time on Less Wrong/IQ: -.164 (492)
This can be read as smarter people stay shorter on LW.
It seems to imply that over time LW will degrade in smarts.
But it could also just mean that smarter people just turn over faster (thus also entering faster).
On the other hand most human endeavors tend toward the mean over time.
Time on Less Wrong/age: -.108 (1491)
Older people (like me ahem) either take longer to notice LW or the community is spreading from younger to older people slowly.
This made me laugh:
Number of current partners/karma score: .137 (1470)
So it may be that older people just have lower IQ (Flynn effect).
In the data set older people have a significantly higher IQ than younger people. The effect however disappears if you start to control for whether someone lives in the US.
US LW users are on average more intelligent and older.
This can be read as smarter people stay shorter on LW. It seems to imply that over time LW will degrade in smarts. But it could also just mean that smarter people just turn over faster (thus also entering faster).”
Alternatively: higher IQ people can get the same amount of impact out of less reading-time on the site, and therefore do not need to spend as much time on the site
So it may be that older people just have lower IQ (Flynn effect).
The 1600 SAT was renormed in 1994, and scores afterwards are much higher (and not directly comparable) to scores before. As well, depending on how the ‘null’ is interpreted, the youngest are unlikely to have a SAT score out of 1600, because it switched to 2400 in 2005. The line between having a score out of 1600 or not is probably at about 22 years old.
Some thoughts on the correlations:
At first I saw that IQ seems to correlate with less children (a not uncommon observation):
But then I noticed that number of children obviously correlate with age and age with IQ (somewhat):
So it may be that older people just have lower IQ (Flynn effect).
Something to think about:
This can be read as smarter people stay shorter on LW. It seems to imply that over time LW will degrade in smarts. But it could also just mean that smarter people just turn over faster (thus also entering faster).
On the other hand most human endeavors tend toward the mean over time.
Older people (like me ahem) either take longer to notice LW or the community is spreading from younger to older people slowly.
This made me laugh:
Guess who does the voting :-)
In the data set older people have a significantly higher IQ than younger people. The effect however disappears if you start to control for whether someone lives in the US.
US LW users are on average more intelligent and older.
“Time on Less Wrong/IQ: -.164 (492)
This can be read as smarter people stay shorter on LW. It seems to imply that over time LW will degrade in smarts. But it could also just mean that smarter people just turn over faster (thus also entering faster).”
Alternatively: higher IQ people can get the same amount of impact out of less reading-time on the site, and therefore do not need to spend as much time on the site
Wait, this means that reading less wrong makes you dumber!
Hmmm, there was something about correlation and causation… but I don’t remember it well. I must be spending too much time on less wrong.
The 1600 SAT was renormed in 1994, and scores afterwards are much higher (and not directly comparable) to scores before. As well, depending on how the ‘null’ is interpreted, the youngest are unlikely to have a SAT score out of 1600, because it switched to 2400 in 2005. The line between having a score out of 1600 or not is probably at about 22 years old.