Tuesday, June 08, 2021

IQ/EQ Update

I must not have been as clear as I hoped, so I will have another try at it. When discussing Intelligence as measured by IQ I am not much talking about individuals. As I have mentioned before, the success of any one individual is dependent on many things, and I have listed many: resilience, emotional self-control, charm, and diligence, for example. IQ is better at making group predictions.  If you send 100 people with an IQ of 120 to med school, compared to another hundred with IQ 130, and another 100 with IQ 140, there will be a difference in success. There may be some washouts in the 140 group or remarkable successes in the 120 group, but the general trend will be that success, both at school and in careers, will be greater as the IQ goes up.  This will be at least partly true for all fields as well. Waiting tables does not require as much intelligence as electrical engineering, but it requires some, and more is better. Memory and focus are both components of measured intelligence. 

This is similar to my analogy to basketball prowess and height. Take a hundred players who are 6-1, another hundred who are 6-5, and a hundred who are 6-9, how far they go in the game will show a trend favoring height. I could have picked speed, and that would also show a trend, but less dramatically, just as I could have picked foreign language grades in high school as a predictor of college success and seen some trend as well.  Just not as strong as IQ, which is the strongest single predictor.

Yet somehow the conversation always goes to individuals we have known or have read about who have high IQ's but did not do as well as some other people we have known or read about who have some advantage in social skills. I don't know why we can't shake that, but it is largely irrelevant to the overall discussion of IQ and intelligence. Individual lives have too much variance.  Opportunity differs, wisdom differs, health differs. Yet these average out over large numbers, and we can measure what qualities are good to have.

There is also a continuing assertion that what IQ measures is not really intelligence, though there might be some association. This is not so, not unless we switch what we mean by "intelligence." The informal category of smart is usually more encompassing and includes some social skills, and wisdom is immediately recognised as a better thing to have and includes judgement, self-control, and perspective. I have no problem with that.  Yet if you put all of us in a fourth-grade class for a week and asked at the end who the most intelligent six kids were we would have very similar lists. And...they would be the kids with the best test scores. We would all see who was picking up new math concepts, new vocabulary, and science information more quickly and putting it to some use. We would notice the other attributes - which were charming, or musical, or diligent. I include a caveat to this. It would just kill some people not to "reward" one of those children with other attributes with a higher ranking, because they just liked the more charming or hard-working ones better. But if you held their feet to the flames and insisted they stick to intelligence, there would be agreement.

This would happen again at every level: later in school, training someone to be an EMT or fireman, bringing in new software that has to be used for the job. We would all generally agree who were the most intelligent of the last hundred trainees we had, and that list would correlate with IQ.  We might miss a few in either direction because their enthusiasm was so great or so terrible that it disguised ability. And again, it would just kill some people not to rate someone who was a favorite higher than deserved or another who was an arrogant jerk lower, but you can usually find objective measures of some sort. 

Finally, there is the continuing matter of EQ.  I thought I had been clear and even repetitive that individual people skills, or social skills, or emotional skills were valuable, often more valuable than raw intelligence. Yet somehow there is some impression that I am devaluing these things because I don't think EQ is real. Let me try again. There are many social skills, all of which have some importance and in some instances critical importance. Yet these do not measure well. There might be rating systems based on responses from hundreds of independent evaluations that could somewhat identify wisdom, or charm, or evenhandedness, or tact, or kindness, and each of these might leave a trail of some type of success in any profession that we could try to evaluate, but with each of these, levels of evaluation, whether letter grades or numbers, are going to be uneven and difficult. We would have to say by the objectivve measure of being elected president that both Donald Trump and Barack Obama must have a great deal of charisma.  Yet there are those who are repelled by either (or both) of them. So what's the charisma score going to be?*

If the objection comes that we aren't really talking about a score of any kind, just an informal acknowledgement of some ability, then I will ask why you are bringing in Quotients, or why you feel the need to call the virtue you prefer by that name. Does the implication of prestige or power attached to the mathi-ness of the term seem too much of a gift to one ability at the expense of others you like better?  I confess I don't quite get it. 

Next, these various people skills are not strongly associated.  There is some association between leadership and followership in that people who understand those roles can apply different skills at different times, but one can see that those two skills will just as often be offsetting. Is being easygoing a good social skill? Well, very often it is. An easygoing surgeon may not be setting a good example for his team, though. Charisma and tact may conflict. Self-confidence and humility are imaginable in the same person, but we can see how they might not coincide. Friendliness and kindness would seem to be general goods, but sometimes they distract from getting work done, or from justice in the workplace. All these attributes of emotional wisdom don't cluster into some general category. They don't measure well individually, so they certainly aren't going to measure any better by lumping them together and trying to extract a number from that. There's no quotient or any other mathematical term that's going to fit in the least for a general category of people skills. 

So there's no such thing as EQ. If all you mean is "people skills are important," then I have never objected. But I think what folks are saying is closer to "IQ tests don't measure social skills, and I admire social skills, so we should stop paying attention to IQ scores." Eh. The tests never pretended to measure anything other than candlepower. But they are very good at measuring that. Yardsticks don't measure speed, but they measure height very well.

* 3d6 plus bonuses

31 comments:

Narr said...

I'm not sure where my own comments fit in your critique, but just to clarify-- I'm an IQ Believer and High Score-Achiever, but I don't really think EQ is a thing as you define the terms (or by anyone else's).

And obviously, I missed the whole group/solitaire framework.

I took the GRE twice for masters programs. The first time (1978 or so) I hit 770/610, the second time 760/590 (about ten years later). No brag, just numbers that others thought important for their purposes. (All my SATs and ACTs, and the one MAT I took, likewise.)

FWIW I'd hesitate to take them today.

Cousin Eddie

Assistant Village Idiot said...

No problem. I intentionally did not go back and take the comments here, at Althouse, and at Maggie's apart in detail. I absorbed the overall impression and forgot the names. I have not repented of that decision and am not going back. It's a false precision that leads to argument rather than discussion. While it is sometimes valuable to note you chose this word instead of that one and I find it significant, it is more usually an infallible marker of someone who wants to win rather than arrive at the truth.

I mentioned about a year ago that we think my wife was once the record holder on the MAT with a 98. I never took it myself, but I know I wouldn't have hit that number. I have people all over my family with stunning scores. But I can also tell you where every one of them has areas that are infuriatingly obtuse or unwise - as they likely could also say about me. IQ is what it is, and I defend it against those who say it is more and those who say it is less.

Sam L. said...

"* 3d6 plus bonuses" OK, AVI, WHAT does that mean? What does it apply to?

stevo said...

I heard something once that seemed true. No matter what the test is, the high scorers have higher IQ. Any test is an IQ test and the ones who do well have higher IQ. Makes sense to me.

Thomas Doubting said...

Sam L, that's a D&D joke. Charisma was one of six characteristics that characters in the game had, and you rolled 3 x 6-sided dice to see what your character's score was. Some types of characters had bonuses.

Thomas Doubting said...

AVI, thanks for posting this longer clarification. I really appreciate it, and it's made me think about why this particular topic seems to get more heated than it seems to deserve. If I get some time, I may post about it over at the Hall.

Assistant Village Idiot said...

@ Sam L - Thomas is right. Inside joke.

@ Stevo - that's an astute observation. I had a psychiatrist friend who put up his hands and said to me "But there's just a sense that you get, taking those tests. Once you know the code, they aren't that hard," and a testing psychologist chuckled and said "Oh yes. But there are people who take those tests repeatedly and don't get any better. So they measure intelligence, one way or the other."
If you can't figure out what key turns the lock...

Erick Arnell said...

It's a reference to rolling dice while making a character in Dungeons and Dragons. Basic characteristics like strength, wisdom, and charisma have a numerical value found by rolling three six-sided dice, and then adding any particular modifiers.

james said...

https://www.topendsports.com/testing/tests/ball-throw.htm

I found a 2d10 useful in picking door prize winners recently.

I'm not sure how you read a d1, and a d0 sounds very zen.

Narr said...

Dice? I have four, six, eight, ten, twelve, and twenty-sided dice--can't really wargame seriously without a diverse selection of randomizers.

I D&D'd before it was cool.

Cousin Eddie

Assistant Village Idiot said...

One could theoretically have a three-sided die, a cylinder. It would take some work to get the dimensions correct so that it came up 1/3 of the time on each, though. A coin is a 2-sided die.

Narr said...

"Coin = 2-sided die." Excellent point.

If visualizing your 3-sided die is a test of IQ, I am failing.

Cousin Eddie

Assistant Village Idiot said...

It's not an IQ test. You are making it too hard. If you imagine a cylinder about the proportion of a squat can - maybe something between a tuna can and tomato sauce - if you threw it on a table it might land on either of the flat sides, or it might roll freely. Rolling freely can be thought of as a third side. To get the proportions right so it didn't mostly roll, such as a soda can, or always land on a flat side, like the tuna can, would take a lot of trial-and-error. I estimate it would be something like a cylinder with equal height and diameter.

Thomas Doubting said...

You could just make the cylinder a triangle, so there would be 3 long rectangular sides and 2 small triangular ends. If you rolled it on the long side, it wouldn't land on the ends.

I've also used coins with binary values when dice weren't available:
penny = 0/1
nickel = 0/2
dime = 0/4
quarter = 0/8
You can mix them as needed, and if you don't want a 0 result, add 1 after you've counted them up.

So:
penny + nickel + 1 = a d4
penny + 2 nickels + 1 = d6
penny + nickel + dime + 1 = d8
etc.

A d3 then would be 2 pennies + 1.

Of course, you can just do d6/2 and round up.

Does this post make me look like a geek? :)

Zachriel said...

stevo: Any test is an IQ test and the ones who do well have higher IQ. Makes sense to me.

Any test that you can study for is not an IQ test, which is supposed to test for innate ability.

Assistant Village Idiot: But there are people who take those tests repeatedly and don't get any better.

Sure, so IQ would be the measure of how much progress someone makes through equivalent preparation, not the results of the test itself. Even then, there can be confounding factors, which could include a lifetime of exposure to similar tests as well as simple interest in such problems.

Nor does it appear a single value can capture the wide variety of how intelligence is expressed. Some people have good language skills, but poor maths skills, or the reverse. Or take someone, such as Louis Armstrong, an undoubted musical genius who transformed musical tradition. He may not have scored well in either a language or maths skills.

Cranberry said...

So, I'm lazy. Rather than trying to create a 3-sided die, I'd use 6-sided die, but only count either the odd or even numbers. If I were only counting odd numbers, the even number would be 0. Thus, with a normal set of 2 dice, your possible values are 2,3,4,5,6,7,8,9,10,11,12

With odd numbers, you can roll: 0,1,2,3,4,5,6,8,10

With even, 0,2,4,6,8,10

This has the advantage of giving a wider range of outcomes than a 3 sided die, which could only give 2,3,4,6.

Plus, you avoid the physical problems of trying to reinvent the wheel (die) so to speak, or chasing wandering dice around the room.

Cranberry said...

In general, when people object to IQ scores, I think they're objecting to the idea that IQ scores determine life outcomes. Certainly, there is a great deal of research that finds that in general, having a high IQ is beneficial. On the other hand, I can think of people with very high IQs who were not successful. Temperament, character, and personality play large roles in success.

The problem in internet discussions of IQ arises because the people arguing for and against IQ inevitably think they themselves possess high IQ.

Right now, wide-scale IQ tests are not administered to the public. There are tests which are aking to IQ tests, such as the SAT. The SAT has been progressively, purposefully made to be less of an IQ test over the years. The beauty of the original standardized tests is that they discovered the highly intelligent children who were not teachers' pets.

Conscientiousness, diligence, obedience, orderliness, friendliness--these do not correspond with high intelligence. They do, however, correspond with teacher nominations to gifted programs.

Assistant Village Idiot said...

@ Cranberry - Regarding the assumption of high IQ: when I used to answer on Quora, the question would often be asked "I just got a 127 on an IQ test. Is that a good score?" I enjoyed answering "In the real world it is a very good score, better than 95% of other humans. On the internet, it seems to be well below average." The point applies with EQ as well. When people bring it up, they seem to assume they've got a fair bit of it themselves.

I had not really figured out that the SAT was changing itself intentionally in that direction, but I think you are right. When they added the writing section (which some suspected darkly was to disadvantage foreign Asian students who were scoring better than Americans) it was certainly a move away from aptitude rather than achievement. To move intentionally away from aptitude would have to involve increasing the cultural bias rather than decreasing it, wouldn't you think? The research that the tests are not culturally biased is long-standing, but maybe the data has changed over the years and they have become at least slightly biased over time. They went for a particular bias early on when they eliminated math questions that females did more poorly on, operating on the assumption that there must therefore be bias.

Gifted programs - yup. Exactly that. This generally favors girls, but it does particularly disadvantage a particular type of girl who does not show those expected feminine traits. They get a backlash.

Zachriel said...

"The SAT and IQ test measure different things, and are difficult to compare. The most notable difficulty in comparing the SAT to an IQ test is that you can improve your SAT score over time; research has shown that most students get a higher score the second time taking the exam....

"The answer to why you can’t calculate your IQ from your SAT score is simple: the SAT is designed to test the facts, concepts, and skills you have acquired over your academic career. To test this, the SAT will present problems and situations that require you to rely on the information already in your possession. Alternately, IQ tests evaluate your ability to formulate answers with only the information provided to you. More simply, an IQ test examines not what you’ve learned, but your capacity for learning."


https://blog.collegevine.com/do-sat-scores-correlate-to-iq/

Sam L. said...

I was before D&D. I've heard of it, but never knew anyone who played it...that I knew of. And now, in my Fortress of Solitude...

Thomas Doubting said...

Cranberry: Rather than trying to create a 3-sided die, I'd use 6-sided die, but only count either the odd or even numbers. If I were only counting odd numbers, the even number would be 0.

I'm curious why you would want to do that.

In D&D (& similar games), you want a d3 to give 3 equally possible results: 1, 2, or 3.

The simplest way to do that would be to roll a d6, divide the result by 2 and round up (so 1 & 2 => 1, 3 & 4 => 2, and 5 & 6 => 3).

You would want 2d3 to give results from 2-6.

If you need a wide range of possibilities, you would use a d20, giving a range from 1 to 20.

So I'm curious what your method is accomplishing.

Assistant Village Idiot said...

I am going to guess that it might be the center-clustering aspect, though I am not sure. To get a score from 3-18, you could get a hypothetical 16-sided die and add two to the result. But the idea is to cluster the results in the center, so that a score of 18 represents 1-in-216 rather than 1-in-16.

While I was very pleased at my own cleverness at conceiving of a cylinder as a three-sided die, it occurs to me that the purely Newtonian physics of this might be unstable if one stated adding in other dice, so that throwing three of these at a time might not yield results that are identical to 3 dice rolled independently and not touching each other. They might interfere with each other in unpredictable ways. It is exactly the sort of thing that math people get fascinated with and devote way too much time to.

Thomas Doubting said...

Sure, anytime you use 2+ dice, you're going to get a bell curve. So I'm curious why he (or she?) wants to throw 2d3 and get possible outcomes higher than 6, or lower than 2.

I could use my d6/2 method and get a bell curve by throwing 2 or more of them.

Just curious.

You have a good point about the cylinder d3. I didn't think of that. With my long triangular d3, I guess you could hold up however many you needed vertically a foot or so above the table and drop them.

Narr said...

Fascinating discussion of dice, but afield from my own recreational needs--and intellectual strengths.

My son (who tests good, like me and his mom) was in the public system's gifted program. I can only say that it was more challenging than what he and his friends dubbed the "Tarditional" program, and the even more tarditional courses that I took in the 60s.

Cousin Eddie

james said...

I think AVI's right: the "triangular-cross-section" d3 (with pointy ends) would tend to press the sides together in your grasp--not quite as badly as throwing a handful of pennies, but in that same sort of manner. Tossing them honestly would be a bit of a skill.
However, you could arrange for less correlation by having the surfaces be rounded instead of flat.
But I can't think of any reason to actually try to use the things--too pointy. Even a d4 is unpleasant to use.

AVI's other suggestion of using a cylinder is a bit tricky to implement--the proportions might depend on how "bouncy" the material is, and you could imagine the equivalent of "shaving the dice" in difficult-to-detect ways.

I've used d3, though. A children's game whose name eludes me used a regular 6-sided die, with the numbers 1, 1, 2, 2, 3, 3 on the sides.

Assistant Village Idiot said...

A simple answer I completely missed

Cranberry said...

My (adult) children play D&D, but I only observe. I thought it was interesting, when thinking of the outcomes of throwing dice, that 0 is rarely an option.

Is there a value in the odds of each particular outcome being equal? We expect it from the tosses from 2 standard, 6-sided dice, but as long as you know the distribution of the outcome, it shouldn't matter to the game, particularly for a game in which the outcomes only determine a particular choice, rather than the a game in which larger numbers are better.

Christopher B said...

As Thomas D pointed out, the value in tossing 2d6 as opposed to 1d10, which provides almost the same number of results, is that a particular outcome between 2 and 12 is not equally likely with 2d6 unlike each outcome between 1 and 10 with 1d10. There's only one way to toss either 2 or 12, but two ways to toss 3 and 11, three ways to toss 4 and 10, and so on with six ways to toss a 7. It gives you a way to express that particular game outcomes are the most likely while allowing for exceptionally good, or bad, outcomes at random times.

Narr said...

The wikipedia article on Craps has a nice table (with terminology, i.e. "hard eight" etc.)--of course wargamers learn all the odds early: from simple 1d6 to 2d6, and 2d10-- well, I've seen many a battle turn or ship sink because of the dread 0-0 from 2d10s!

D&D was a brief phase for me--I'm more into history than fantasy--but my son and some of his
friends (thirtysomethings) still play.

Cousin Eddie

Thomas Doubting said...

In D&D, I think a bell curve is most useful for rolling permanent characteristics where you want them to cluster around the middle but still allow for high or low scores.

E.g., when creating a new character and rolling for Charisma, which is a permanent score that can only change through game events, you want most characters to have average scores, but some with higher / lower, and a few with very high / low scores.

Is there a value in the odds of each particular outcome being equal? We expect it from the tosses from 2 standard, 6-sided dice, but as long as you know the distribution of the outcome, it shouldn't matter to the game, particularly for a game in which the outcomes only determine a particular choice, rather than the a game in which larger numbers are better.

Well, generally, dice (in RPGs) don't determine choices but rather the outcomes of choices, so typically higher is in fact better.

Also, if you have to think about distribution, it slows the game down. A player usually has a number of possible actions to choose from, and the dice are rolled after the decision, so he has to think about the possible outcomes of the dice before deciding. The less he has to think about each possible action, the faster the decisions are made and the quicker the game goes. We're here to kill dragons, not do stats.

Also, for most RPGs, the whole range of dice (d4, d6, d8, d10, d12, and d20 are typical) is used, and you might throw any number from 1 to 3 of them, so there isn't a single set of probabilities, but rather multiple sets that you'd have to be familiar with. And, of course, there are situational bonuses and penalties to account for as well.

Some RPGs do other things, though. Shadowrun only uses d6s, and you might throw as many as a dozen or more dice at a time. They add up the number of 6s rolled, and compare that total to a target number to see if the action succeeded.

I can't remember the name, but there's a game that uses 2d10 for everything. However, one die is positive and the other is negative (they're usually different colors), so you'd get results like +9 and -4, so 5, or -8 and +2, so -6.

Zachriel said...

An elf was caught unarmed by a seasoned orc with an axe. She thought quickly. "I'll spit in his eye! The dungeon master determined that to be very unlikely. So she rolled a d20. Fortune rolled in the dust. And there it was, a 20!

"Ha! Take that orc!" said the elf.

"But the orc has 2 eyes," said the dungeon master.

"Oh," said the elf. She rolled again, bating her breath, ready to dart away if she could. Another 20! The orc was temporarily blinded. She tripped the startled orc, took his axe, and ended it.