Comment by btilly | Hacker News Reader

Comment by btilly | Hacker News Reader

Parent Full thread

When did the officials find this out?

It's only been known for close to a century. It is the reason why standardized tests were accepted in the first place. It has been confirmed in many ways since. If they were ever ignorant of this fact, it is because they were willfully choosing to be ignorant of it.

Research also found that IQ tests are a better predictor of job performance than other available measures, such as interviews. This was first established in the military, then in the 50s and 60s was confirmed for a variety of jobs. Unfortunately the case Griggs v. Duke Power Co. make it illegal to use IQ tests, because it results in hiring fewer blacks.

But IQ tests can still be done indirectly. For example coding boot camps can use what is essentially an IQ test to decide who will be a student, and then graduation from said boot camp serves as a signal to companies that this inexperienced person is smart, motivated, and willing to learn. That signal is likely to be more valuable to the company than the training itself.

Griggs did not make the use of IQ tests illegal. It simply required businesses to prove a job-related business necessity for requiring such a test. (And this was codified into law by the Civil Rights Act of 1991.)

Some businesses are able to prove a job-related business necessity for IQ tests, and do require applicants to take those tests for certain positions. Most businesses can't show a business necessity for any positions, and so don't require IQ tests.

To prove a job-related business necessity for IQ tests, you need a large group of people who have been through IQ tests, and whose performance you can measure. The US military has collected this data and can use that.

Your average company with < 100 employees almost certainly CANNOT have collected this kind of data. And larger companies will have trouble putting large groups through an IQ test with no apparent purpose, and so are unlikely to have ever collected the data needed to attempt to make the case. The practical effect is that such tests are usually illegal to use. Even though we know that they are very often useful.

That's still false. You don't need to have a large group of people go through IQ tests to prove a business necessity, you need to prove that IQ specifically is relevant to the specific job in some way.

And the thing is, that it's very difficult to show that IQ (as measured by an IQ test) is relevant to a specific job versus something like experience, education, etc., because "IQ" is a made up number that doesn't actually test intelligence or problem-solving abilities.

Businesses test applicants all the time. But instead of a useless IQ test, they test job-specific performance through things like coding tests, draft patents or legal documents, etc.

You ALMOST get it, but then entirely missed the point.

You need to demonstrate that the test is consistent with a business necessity. If you have a large enough people who have taken the test, and enough data on how they perform, you can make that demonstration. You've got the data, and it is enough to cover the military despite people like you who think that it "doesn't actually test intelligence or problem-solving abilities." Because, whatever it actually tests, it predicts real world performance well enough to pass legal muster.

The US military is the only organization that I'm aware of which has an IQ test with enough data that they can pass legal muster. Very specifically I'm talking about the AFQT, which is the core of the ASVAB. See https://www.military.com/join-armed-forces/asvab details on the ASVP. And see https://asvabmilitarytest.com/asvab-compared-other-iq-tests to see that it really is an IQ test to tell if you're smart enough to be in the military. And not just whether you can join, but which jobs inside of the military they will allow you to pursue.

The military has documented that this is an effective way to find people who can do things like fill out paperwork properly, and follow orders in the field. The reason why other businesses can't do the same is not because the test is useless, but because they don't have the data to demonstrate how useful the test actually is.

You're simply wrong. It's not a data set size issue. The issue is that most businesses can't show a sufficient connection between an IQ test and the responsibilities of a job to establish a business necessity for having the IQ test.

If they can show a business necessity, they don't need to show that the IQ test actually predicts job performance.

Notably, the army's test is not an IQ test. It is an aptitude test which is a very different thing. An aptitude test specific things are that directly and strongly related to the job (or jobs) for which the applicant is applying.

It is an IQ test, they will tell you that. IQ predicts job performance and wealth. Sorry if that offends you but it is reality. Good luck fighting it!

I appreciate your point, but the nuance to remember is that IQ tests are also influenced by education (or, as you point out minorities, access to education as well). IQ tests do have value only within the context of what they essentially require to be able to take them and have a shot and demonstrating your innate problem solving and reasoning ability, etc.

> IQ tests are also influenced by education (or, as you point out minorities, access to education as well).

Presumably, so is job performance?

I suspect we'll get to scientific racism sooner or later in this thread.

The cure to that would be letting colorblind standards work over time, but that's politically unacceptable because too many people derive their income from nonsense like "merit is a white colonial construct". Quit treating people differently based on race, and racist kooks will be irrelevant.

I agree with the spirit of this comment, but note that quitting treating people differently requires fixing the school system so that minorities don't consistently start off in second rate schools.

It isn't just one side that finds talking about real solutions politically unacceptable.

I think it's more precise/actionable to say that quitting treating people differently "includes" fixing discrepancies in the primary and secondary school systems, rather than "requires" it.

I strongly disagree.

But saying that it is required, I'm drawing attention to the fact that it must be done to succeed. I believe it must be done because, starting with Piaget's work, we have evidence that children go through important developmental stages. Some lessons missed at specific ages, can be missed forever.

If I had merely said included, that makes it easy to walk away thinking that it is just one of many things that advance the goal. Which takes away from a must, to a nice to have. And then we can excuse not doing it based on the price tag. While imagining that the other things we're doing somehow will add up to a real fix. A piece of imagination that we make easier by discounting the evidence of standardized tests which demonstrate exactly how badly we, as a society, are failing.

My whole point is that, since the popular rejection of school busing, neither party has been willing to try to honestly tackle this problem.

> IQ tests are also influenced by education

[citation needed]

Reply on news.ycombinator.com