Saturday, November 05, 2022

Hypothesis Dashed

In discussing intelligence and other abilities, and how they relate to performance and success, I have often taken the analogy of basketball because the attributes height, speed, and gross-motor coordination are all extremely important, but not equally so - and other attributes contribute as well. To succeed at the highest levels more than one of these is going to be needed, even if you have one in abundance.  Consider Tacko Fall, who is 7'6", but even he has needed some hard work and engineering-level intelligence to overcome his lack of speed and general coordination. One could also use music as an analogy for similar reasons. 

I have consequently believed that in the same way that height is the #1 requirement for basketball but is a threshold phenomenon, and after a certain point other characteristics take on increasing and eventually equal importance, that intelligence is the same in it's operation across a huge number of activities.  Yes, one needs an IQ of 90 to get into the USMC, and probably 110 to be an officer, but beyond that other characteristics such as self-discipline begin to gradually take over as predictors of success. Certain hard sciences may have very high floors, up into the 130 range or even more, but even there the ability to work with others, the capacity for hard work, etc become of increasing importance.

That is at one level true, because IQ alone won't get you much.  But the effect of the other characteristics such as the modern favorites "grit" (the Angela Duckworth book), growth mindset, EQ, 10,000 hours (with or without focused practice), or privilege - or the old favorites hard work, charisma, luck, training, or connections turn out to matter only a little each according to Michigan State researcher Zach Hambrick. Even at high levels of expertise, intelligence remains the best predictor of the ability to next level up. Not EQ.  Not growth mindset. There does seem to be some sort of generalisable sales ability, but it has proven very tough to measure.  which doesn't stop these guys from selling you theories about what it is and what works.  Selling sales is big business.

I had expected a different result.  But his methods seem good and the samples are large, including the ASVAB data in some cases.  Hambrick is on the committee to improve the ASVAB and AFQT, and he is at pains to say that he thinks they are already quite good but could be better, not that the test should be discarded after 100,000,000 tries.

His explanation is not universally accepted, but this is always true in IQ research.  People's priors are that something else, maybe anything else, should be true and there is motivated reasoning. Even I hoped that the hypothesis above would prove out and the starkness that g-factor is the single largest predictor in field after field would be relieved at least a little.  

People do have trouble understanding even a two-factor model when it goes against their priors, and they immediately accuse one of claiming that those other things don't matter at all. No, it's just that they vary from task to task whether they are important, with little generalisation among them, and even at highest levels do not match intelligence. Speed in basketball is big, but not as big as height. Beauty is big in Hollywood, but sustained excellence over decades is better predicted by intelligence, as in Phi Beta Kappa Glenn Close.


David Foster said...

So maybe Hedy Lamarr's high intelligence had something to do with her success as an actress, as well as with her inventing activities?

james said...

Since g doesn't explain everything, we want to look for g' that explains the rest. Maybe it's a combination of things that have to work together. Maybe Brady was good-but-not-great until he played alongside some guys who used some practice he needed.

I know luck plays a little role (which they don't deny, of course). I remember a math class I took. I was having trouble remembering which orthogonal function was which and why, but the prof's oral exam question was about the gamma function. I'd been fascinated with it and playing around with it on my own for a month, and had obscure facets of it on the tip of my tongue. That was the one question from the course that I could really ace. I wouldn't have failed with a different question, but I looked a little better than I really was.

David Foster said...

Generalized sales ability but difficult to factor here is that there are so many different ways to be a great salesman. The extraverted-loudmouth stereotype doesn't always apply.

Zachriel said...

Interesting that the original post uses basketball as an example, where height, hand size, and reflexes are at a premium. That is not to say there is no intelligence involved in basketball, but g will probably not correlate as well for basketball as for many other fields.

Hambrick: One important point to make here is that G is far from perfect as a predictor of job performance or any other outcome, so there’s value in searching for additional predictors.

Hambrick: The average correlation across a wide range of jobs between G and performance is about 0.5, somewhere in the neighborhood of 0.5.

About half, so, g is not destiny.

Hambrick: it’s possible that an individual can, in a sense, defy that trend and reach a high level of performance in becoming a helicopter pilot or something through just doggedness.

Doggedness is often a factor in success.

Hambrick: Well, and I would just add that they’re {high-IQ nonconformists} probably not going to enlist in the first place.

Hambrick uses military testing as the basis for most of his findings. But this group is self-selected. A non-conformist may not do well in the military regardless of g, and will be less likely to be included in Hambrick's stats. Presumably, non-conformists would do poorly in the infantry regardless of g, suggesting the actual correlation with success is lower than its measured value. Non-conformists may do better in certain fields of art or design, but success in art can be very subjective and subject to popular trends.

Hambrick: They stipulate that full concentration is necessary for deliberate practice. Well, how do you know whether somebody is concentrating fully? Can anybody ever concentrate fully on something? Does that mean that 100% of your attentional resources… How would you ever know that?

This misunderstands deliberate practice. It's a specific type of practice that breaks complex tasks down into simple components. A piano player may practice the thumb cross. A baseball player may repeatedly practice swinging low and wide. A chess player may practice knight and pawn endgames.

Hambrick: We can make predictions about what individuals can do, but it’s possible that an individual can, in a sense, defy that trend and reach a high level of performance

So, g is not destiny.

Christopher B said...

People do have trouble understanding even a two-factor model when it goes against their priors, and they immediately accuse one of claiming that those other things don't matter at all.

David Foster said...

Discusion at Twitter: Is it true that Successful People = Boring People?

Ganzir said...

"Yes, one needs an IQ of 90 to get into the USMC"

I thought the reputation of Marines is that they're not so bright.

Ganzir said...

I say this not to disparage the Corps, but just to relay the stereotype as it reached me.

Assistant Village Idiot said...

The numbers fluctuate depending on wartime and recruitment needs, but in general it 90 for Army, 91 USMC, 93 Navy and Coast Guard, 95 USAF. Obviously, officer corps are different, and if you want to be in many specialties your AFQT is going to be higher than that minimum. 90-110 is considered average intelligence.

james said...

According to wikipedia:
In 1824 Sir Walter Scott used the phrase "Tell it to the Marines – the sailors won't believe it" in his novel Redgauntlet.