Hacker News new | ask | show | jobs
by WhitneyLand 3457 days ago
The results are really impressive. It's also comforting to see them dealing with some of the subtleties that were intuitive but not always practiced. That labeling people high and low can both be damaging. That putting children in college could sometimes save them emotionally rather than cause problems. That IQ tests are bullshit when used improperly and even the best tests are still missing people who should be nurtured.

Other factors seem very important anecdotally, have these been dismissed or just not fully studied yet?

1. Socioeconomic background - they mention it's not as predictive, but how sure are we Microsoft would exist if BG's/PA's parents were illiterate and couldn't afford to give them access to computers?

2. Mental health - Someone could be a 4th standard deviation genius yet be debilitated by anxiety, depression, or some other condition.

3. Occasionally mental "disability" can seemingly be related to success. So many great scientists have been on the Autistic spectrum, or suffered from severe OCD. Can the success of the person be separated from the disorder? Could "cures" for Autism and OCD actually negatively affect scientific progress? Or could their condition cause them to be incorrectly screened out?

4. How do curiosity, motivation, and ambition affect success? Do high scoring kids automatically have these traits or is it a separate variable?

2 comments

> how sure are we Microsoft would exist if BG's/PA's parents were illiterate and couldn't afford to give them access to computers?

We can be entirely sure it would not exist. This is not a hard prediction.

I think you're right. But I don't understand why they implied its predictive ability is low.
> Someone could be a 4th standard deviation genius

What does this mean? I am not familiar with that. Is that related with IQ or some other metrics?

I googled 4th standard deviation, and it pointed me to Wikipedia article https://en.wikipedia.org/wiki/68%E2%80%9395%E2%80%9399.7_rul... Reading that article, I am still not sure what "4th standard deviation genius" means - does that mean "1 in 15787" of people based on some intelligence metrics?

Yes. A common way to talk about IQ scores is as a normalized distribution with mean 100 and standard deviation 15.

So a +4SD IQ score would be 160, or about 1 in 15000. Sounds pretty good, but that makes 20,000 just in the US. There are only 450 NBA players and usually less than 20 A-List actresses at a given time as a comparison.

An aside, anyone who throws out their high IQ unsolicited doesn't seem to make for interesting conversation. Maybe it's too hard to keep up. :)

As a follow up to this, yes, 15 tends to be the standard deviation. However, most, if not all tests have a cap to the highest score they can return, which is about 160. I believe this is the case for both Stanford-Binet as well as Weschsler (WISC):

https://en.wikipedia.org/wiki/Stanford%E2%80%93Binet_Intelli...

https://en.wikipedia.org/wiki/Wechsler_Intelligence_Scale_fo...

I don't know for sure, but I suspect that one of the reasons that 160 is the max score is that the tests need to be normed and that's difficult to do once you pass so many standard deviations.

As another note, the overall composite score can qualify a student for services in a public school, but a good diagnostician will be using the test to assess areas of strength and weakness. Basically, a student who has high spatial reasoning, but poor verbal IQ requires different assistance than the reverse.

Anyway, yes, I'll agree that I hate IQ measuring contests because they're just like the other anatomical measuring contests. Further, if even someone wants to have one, they're almost never about the results from a properly normed test like those from above and, further, those tests contain a lot more information than a single number.

Thank you for clarification. I also like your comparison with NBA and A-actors. I am Mensa member which means my IQ is in top 2% of the general population of my country. Because of the membership and my active involvement, I know a lot of people with high IQ. And my conclusion is that IQ doesn't mean too much. Yes, I think that some of them are very smart, but not genius level smart. Not even close. In my opinion, the usefulness of the IQ metric is very questionable.
I'd say the comparison is not quite right, as everyone can be assigned an IQ (however useful that may be), but not everyone aspires to be an NBA player or A-List actress.
How do you assign an IQ to someone who doesn't want to be tested?
You don't have to, you can just use basic statistics for that, because the pool of people who have an IQ is everyone.

The pool of people who aspire to be NBA players or A list actresses is not everyone though, so you have to estimate first the respective pools.

False. Aspiration is a red herring.

You can just as easily use statistics to assess whether an individual possesses the traits to become an NBA player or actress as you can assess their IQ without their cooperation.

Conversely if someone doesn't want to participate in IQ testing, it is equivalent to them not wanting to play in the NBA.