Hacker News Reader

Comment by htag | Hacker News Reader

> Children who score [IQ]145 or above would then be offered a life-long genius award.

> They could start at, say $75,000 a year and go up by $25,000 increments every decade. The payments could expire if our geniuses decided to take paid employment but then resume if they decided to “drop out” again.

1. I don't think UBI expires, this is more like unemployment benefit which include perverse incentives against employment.

2. IQ is a controversial measurement. It's validity is questionable. It's been used or purposed to be used for truly horrible state sponsored actions in the past. I have a bias against using it for discrimination across populations, simply because of it's history and proponents over the years.

View on HN

> Children who score [IQ]145 or above would then be offered a life-long genius award.

Sounds like it would create MASSIVE incentives to cheat on the IQ test or pay off the graders to give you a good score as well

It would literally set your kids up for life

Also, it doesn't have to be overt "cheating": I suspect that whatever metric(s) would be both defined and gamed by the affluent, like college admissions criteria.

This is the only cogent criticism I can find in this thread. I'm predisposed to like the proposal, but I don't really have a good response for you.

On the other hand, don't most IQ tests have elaborate anti-cheating mechanisms? E.g., handwriting an ungraded written portion?

You insinuate you have a high IQ and that is the only cogent critisms you can find here.

That sounds a lot more like the SAT. That was also supposed to have anti-cheating measures and be resistant to the practice effect. Yet, students who take SAT prep often end up scoring higher (but dramatically higher) than those that did not take such prep courses.

It seems perfectly plausible that repeatedly taking IQ tests would likely result in a higher score due to practice and familiarity. Especially if there is a significant financial incentive involved and someone is on the border of qualifying for it.

LSAT does this as well, and it's generally accepted as an IQ test rather than a knowledge test.

I'm not concerned with trainability. Lots of minimum bars are trainable, like the bar exam, or the Marine Corps PFT/IST. The fact that they're trainable doesn't mean the measures are invalid w/r/t minimum competence as a lawyer, or minimum fitness as a Marine. Similarly, I don't think trainability on tests of logical reasoning, pattern recognition, reading comprehension, memory, etc. necessarily meaningfully reduces the effectiveness of the tests that measure such things. On the contrary, I think it shows that intelligence increases when exercised.

> "IQ is a controversial measurement. It's validity is questionable."

IQ is the most reproducible finding in psychology and psychometrics by far.

It's also one of the most poisoned by history. It has, for instance, been used in eugenics movements and forced sterilizations in the US.

It is influenced by all sorts of factors, from parental income to nutrition.

My family has folks in psych, and they universally regard IQ as "pretty bad, but useful for roughly measuring one form of intelligence".

> It's also one of the most poisoned by history. It has, for instance, been used in eugenics movements and forced sterilizations in the US.

What a weird line of reasoning. So are rocket propulsion, gold, and English language.

rocket propulsion has been used in forced sterilizations?

It was developed as a weapon. I guess the parent is saying IQ tests are also a weapon that has found some not weapon uses?

IQ tests reliably measure IQ, where "IQ" is the name of the metric invented to describe what IQ tests measure.

And yet it strongly correlates with tons of other meaningful metrics, like work success and income.

I have no idea what you could be possibly referring to. IQ is known to be susceptible to practicing the exams. IQ itself is measured w/ different tests that don't necessarily completely agree.

If the exams are known beforehand, the test is flawed. Notwithstanding, what people deem to be an expression of IQ requires practice anyway. At any rate I don't understand the value in the tests. They're redundant.

> most reproducible

https://en.wikipedia.org/wiki/Flynn_effect?wprov=sfla1

you must be joking right

please refer to the "reliability and validity" section contained herein https://en.m.wikipedia.org/wiki/Intelligence_quotient

It's a much more complicated topic than your comment would suggest. The very definition of "intelligence" in regards to the iq test is fuzzy and ill -defined.

Not to mention that Alfie Binet (created the iq test) believed the test itself to be insufficient, and asserted that any true measure of intelligence could not be quantitative.

Please don't use Wikipedia to argue your points. We should all be beyond that. It's not a source.

The references in the article might be good sources, but they too often do not actually support the article.

That's a very incestious way of looking at it. Psychometrics in itself starts off with an a priori assumption that mental capabilities (and implied, human value) can be measured. One can always introduce another metric, say Integrity-Quotient, and then develop the test for it. Now we not only have this made-up metric, and its corresponding test (look at how objective it is), but now because it can be captured in a number, it is information in the narrowest computational sense. It's reproducibility is irrelevant in this dynamic. The positive trait (integrity) has now shriveled to fit the parameters of the machine.

This type of circularity is willful submission to the algorithm. We reduce the richness of human cognition only so it can be bureaucratized (e.g. tie it to some fiscal benefit like UBI). It is not fair in any humane (as opposed to mechanical) sense.

I would argue that The Big Five personality traits is just as reproducible but with a much higher construct validity than IQ. It measures a similar thing in a similar way, in that it's an abstract concept about human psychology (personality and intelligence) that is measured with questionnaires.

There's good science in Psychology, even if the field as a whole is suffering the replication crisis. Some parts of cognitive psychology such as studies in reaction times, memory, and perception are very reproducible. Classic conditioning is a standard in animal training.

Psychology as a whole has a bed reputation, and it has received solid criticism. I think a lot of it has to do with a few high profile frauds, bad translational science (theory to application), perverse publication incentives in academia, and the tendency for non-scientist to use questionable findings to justify their opinions.

This only tells you in what dire conditions lay psychology and psychometrics.

IQ is only propaganda for an ideology.

Even as a High IQ person, if someone gave me this kind of money as a teenager, I would have spent half on liquor and wild women and the other half, I would have wasted…

Lex Fridman had an interesting podcast on IQ a few months ago:

https://www.youtube.com/watch?v=hppbxV9C63g

> IQ is a controversial measurement. Its' validity is questionable.

There's a loud chorus saying that. Mostly they're making excuses for some racial group not doing well.

Stephen Jay Gould wasn't making excuses.

Reply on news.ycombinator.com