High IQ: Not as good for you as you thought

A continuation of our "greatest hits" from past Cognitive Daily postings:
[originally posted on December 14, 2005]

IQ has been the subject of hundreds, if not thousands of research studies. Scholars have studied the link between IQ and race, gender, socioeconomic status, even music. Discussions about the relationship between IQ and race and the heritability of IQ (perhaps most notably Steven Jay Gould's Mismeasure of Man) often rise to a fever pitch. Yet for all the interest in the study of IQ, there has been comparatively little research on other influences on performance in school.

Angela Duckworth and Martin Seligman estimate that for every ten articles on intelligence and academic achievement, there has been fewer than one about self-discipline. Even so, the small body of research on self-discipline suggests that it has a significant impact on achievement. Walter Mischel and colleagues found in the 1980s that 4-year-olds' ability to delay gratification (for example, to wait a few minutes for two cookies instead of taking one cookie right away) was predictive of academic achievement a decade later. Others have found links between personality and college grades, and self-discipline and Phi Beta Kappa awards. Still, most research on self-discipline has achieved inconsistent results, possibly due to the difficulty of measuring self-discipline. Could a more robust measure of self-discipline demonstrate that it's more relevant to academic performance than IQ?

To address this question, Duckworth and Seligman conducted a two-year study of eighth graders, combining several measures of self-discipline for a more reliable measure, and also assessing IQ, achievement test scores, grades, and several other measures of academic performance. Using this better measure of self-discipline, they found that self-discipline was a significantly better predictor of academic performance 7 months later than IQ.

How did they arrive at this result? They studied a group of 8th-graders at the beginning of the school year. They used five different measures of self-discipline: the Eysenck Junior Impulsiveness scale (a 23-question survey about impulsive behavior), the Brief Self-Control Scale (13 questions measuring thoughts, emotions, impulses, and performance), two questionnaires in which parents and teachers rated the student's self-discipline, and a version of Mischel's delay of gratification task. Students were given an envelope containing $1, and were told they could spend it immediately or bring it back in a week for a $2 reward. The students were also given an IQ test (OLSAT7, level G).

At the end of the school year, students were surveyed again and several measures of academic performance were taken. The data included final GPA (grade point average), a spring achievement test, whether they had been admitted to the high school of their choice, and number of hours they spent on homework. All except two measures correlated more strongly to self-discipline than to IQ. Scores on spring achievement tests were correlated both to self-discipline and IQ, but there wasn't a significant difference. Duckworth and Seligman suggest that this could be partially due to the fact that achievement tests are similar in format to IQ tests. The other area where there was no significant difference was in school absenses.

Most impressive was the whopping .67 correlation between self-discipline and final GPA, compared to a .32 correlation for IQ. This graph dramatically shows the difference between the two measures:


Both IQ and self-discipline are correlated with GPA, but self-discipline is a much more important contributor: those with low self-discipline have substantially lower grades than those with low IQs, and high-discipline students have much better grades than high-IQ students. Even after adjusting for the student's grades during the first marking period of the year, students with higher self-discipline still had higher grades at the end of the year. The same could not be said for IQ. Further, the study found no correlation between IQ and self-discipline—these two traits varied independently.

This is not to say this study will end the debate on IQ and heredity. The study says nothing about whether self-discipline is heritable. Further, the self-discipline might be correlated differently with achievement for different populations; this study covered only eighth graders in a relatively privileged school. Perhaps self-discipline has a different role at other ages, or in more diverse populations (though the study group was quite ethnically diverse—52% White, 31% Black, 12% Asian, and 4% Latino). Perhaps the most important question which remains is how best to teach children self-discipline—or whether it can be taught at all.

Duckworth, A.L., & Seligman, M.E.P. (2005). Self-discipline outdoes IQ in predicting academic performance of adolescents. Psychological Science, 16(12), 939-944.

More like this

How about studies of the correlation between IQ and metrics of social and occupational functioning? I'd bet something similar would emerge -- people who score high on IQ tests are probably more mercurial and flighty and impulsive than those with merely "average" levels of self-discipline.

this study covered only eighth graders in a relatively privileged school.

I was thinking this during the whole post, and then here's confirmation. In statistics, this is called "restriction of range" for some variable, here IQ (because the school is privileged, indicating that students of low or average IQ are unlikely to be found in the expected proportions). The less a variable varies, the less power it has to account for some outcome.

Let's say you look at the NBA (I'm not sure if this is true for them or not; just suppose for sake of illustration). And you find that height doesn't vary as much among them as among a representative sample of the population. You'd find that out by estimating variance or SD of height in the biased (NBA) and representative samples.

That means that NBA players are more tightly bunched around their mean -- say that 70% were within 1 inch of 6'7. In a representative sample, 70% of adult American males tend to be within *3* inches of 5'9 or 5'10. Looking at height differences among NBA players to see if they predict differences in performance is like splitting hairs. It seems naive to expect that just because player A is 6'7.25 and B is 6'7, that A would out-perform B. A real study would probably show that height accounted for some non-trivial but wimpy amount of variance in performance.

That's roughly what kind of height differences you'd see in a biased sample. In a representative sample -- say, those in a high school gym class, unselected in any way. Since height now varies much more, it can account for more variance in outcomes, and you'd surely find that being taller than average conferred a great advantage in basketball.

Discipline clearly matters too, but anytime someone doesn't find that IQ strongly predicted academic performance, it's almost always due to this statistical mistake. They only look at a group whose IQ doesn't vary that much.

Kemibe -- personality traits and IQ do not correlate one way or the other, for any personality trait. Actually, Openness to Experience correlates weakly with "verbal" or "crystallized" IQ -- knowledge that's accumulated through experience and stored in memory. That makes sense: more Open people are going to be more voracious in accumulating facts, vocab words, etc.

But it doesn't correlate with "performance" or "fluid" IQ -- ability to reason through a pattern-recognition problem (like, they give you 3 pictures where some pattern is going on, and you have to figure that out and say what the next picture in the sequence would look like).

So let's open another can of worms and talk about the heritability of "SQ".

This would explain why my husband, the "smartest" person I know, is also the worst student when it comes to formal education!

A huge problem with these kind of studies is that people assume GPA has some intrinsic value. I knew plenty of people at MIT who were incredibly smart, but had poor GPA's because they were easily distractable by extracurricular activities. I believe IQ tests-- and for that matter SAT test-- measure something valuable-- but are rather poor at predicting "success in school" because school involves lots of "toeing the line" and follow directions and busy work.

I have found during my lifetime of observing people that most persons who have a high IQ and for whom learning comes very easily don't apply themselves. In other words, they don't have self-discipline and float on "being smart", and are easily bored. Thus they are often not high achievers because even if you have a high IQ achieving excellence in anything takes HARD WORK. I've also found that people with high IQ's are not very practical and tend to be idealists and dreamers, which sometimes creates roadblocks for them. My x-husband's family members are all like that, high IQ's, low achievement and no self-discipline. It's actually a bit sad because of the wasted brain power.

On the other hand, when a high IQ person does have self-discipline, the sky's the limit. They are the great achievers and geniuses in our society. But the two have to go hand in hand, and often those with nothing but pure determination and self-discipline achieve more than those with a high IQ.

By roseindigo (not verified) on 07 Jul 2007 #permalink

Tom B, aren't you just committing the anecdotal version of bad sampling and reasoning that Agnostic describes above.

By Moonenite (not verified) on 07 Jul 2007 #permalink

"Tom B, aren't you just committing the anecdotal version of bad sampling and reasoning that Agnostic describes above. "

No doubt, but I wasn't claiming my observations were Science, either. My point was that maybe "success" means different things to different people.

Shouldn't we be asking the Marine Corps how to cultivate self-discipline and applying it as a component of kids' (and adults) education?

It would also be interesting to break-it down into learnable skills, such as goal-setting, emotional self-regulation, working memory...and stop thinking of both IQ and self-discipline as fixed traits.


I don't know if Marine Corps discipline would help, especially with high-IQ individuals (are you familiar with Swofford's Jarhead, or the movie made from same? I don't think slamming people's heads against chalkboards is healthy, plus intelligent people will just find more creative ways of rebelling, a la the one Marine who shot the Bedouin's camels). I think a better approach would be for the child to have a mentor at a critical age to guide and inspire him/her.

By David Group (not verified) on 08 Jul 2007 #permalink

This isn't particularly surprising to me: I consider myself to be quite bright and have consistently tested quite well on several measures of 'IQ'. However, I know I'm as lazy as sin and that is what is to blame for my lack luster performance at Uni and work.

Of course this is only a correlational study and can be classed as no more than an interesting statements about eighth-grade kids.

Now, if they had administered some sort of educational intervention to increase self-discipline and measured GPA in both high and low IQ individuals it would have made for a more interesting study. Although the write up seems to suggest that self-discipline is a stable innate trait not subject to variation across time (in a similar manner to IQ), so one wonders if such an intervention would actually work

By alephsmith (not verified) on 08 Jul 2007 #permalink

This research sounds like a variant of the work by Lewis Terman.

That research focused on asking the question whether a high (academic) IQ predicts high achievement in the real world. After conducting longitudinal research on 1500 children with IQs greater than 140 ("genius" intelligence), the research revealed that participants could be classified into two groups: the (successful) "A" group had an average income 5 times greater than the general population (in 1955); 2/3 graduated from college; many had academic and professional degrees and became doctors, lawyers, professors (etc.)--although there were no "creative" geniuses. The (unsuccessful) "C" group were earning slightly above the average income; only a handful had obtained professional degrees; many were having issues in their personal lives, were less healthy, had higher rates of alcoholism, and were three times as likely to divorce as the "A" group.

The main conclusion was that high (academic) IQ is not the primary determinant of success. Instead, emotional intelligence (EQ) appears to play the primary role, defined by personality characteristics such as goal orientation, perseverance, and self-confidence (Terman & Oden,1959).

The research reported here seems to arrive at a similar conclusion, vis-a-vis the stronger correlation between self-discipline and GPA and a weaker correlation between IQ and GPA.

By Tony Jeremiah (not verified) on 08 Jul 2007 #permalink

Hi David, I confess I am not an expert in how Marine Corps (or similar) are trained. My only two data points are:

1) 2 former Marine Corps, not active, who explained to me in detail a number of problem-solving and critical-thinking challenges they were exposed to since Day 1 of training. I am not a military or violence fan by any means, but I was pretty impressed by their descriptions.

2) the cognitive simulations that pilots use to improve their attention allocation strategies in high workload attention demands.

I fully agree that we should find better ways to train self-discipline without aggressive paraphernalia. But we probably can learn something from what they do-which I am sure is more refined that slamming people's heads.

This study is of low value as the time frame is way too short, and it only looks at very limited and conventional standards of academic achievment. If they follow these kids over several decades to see how their overall lives turn out, it could be a very useful study. My bet is that self-discipline correlates to a fair degree of conventional success, but that it doesn't correlate strongly to outstanding original achievement by itself.

By Obdulantist (not verified) on 09 Jul 2007 #permalink

"the cognitive simulations that pilots use to improve their attention allocation strategies in high workload attention demands."

Sounds interesting. Is there a link somewhere you could post?

I wouldn't have thought the marines covered much beyond how to shoot at stuff.

An eighth-grader who can hold on to an envelope for a week and then remember to submit it for a reward of $1 may be self-disciplined. But he is also an eager-to-please conformist. A high Conformism Quota is important to achieve good school grades regardless of intelligence and actual learning.

You don't need to be particularly intelligent to do well at school - it's not like it's that hard. High IQ + Attention Deficit Disorder really doesn't help...

Hello Tom B,

Here you can read an interview with Daniel Gopher, one of the key scientists behind such training


Hart S. G and Battiste V. (1992), Flight test of a video game trainer. Proceedings of the Human Factors Society 26th Meeting (pp. 1291-1295).

Shebilske, W. L., Volz, R. A., Gildea, K. M., Workman, J. W., Nanjanath, M., Cao, S., & Whetzel, J. (2005). Revised Space Fortress: A validation study. Behavior Research Methods, 37, 591-601.

Volz, R.A., Johnson, J.C., Cao, S., Nanjanath, M., Whetzel, J., Ioerger, T.R., Raman, B., Shebilske, W.L., and Xu, Dianxiang (2005). Fine-Grained data acquisition and agent oriented tools for distributed training protocol research: Revised Space Fortress. Down Load Technical Supplement, Psychonomic Society Web-based Archive (see 37,591-601).


I've known quite a few people with very high IQs, and most of them were messed up pretty badly by school. When someone is forced, for the first 20 years of his life, to wait out the clock in boring rooms while being "taught" things that he alread knows, the end result is usually not a highly motivated individual. Usually, you get someone who is good at amusing himself and at waiting out the clock.

I think that there is some truth to what the study is showing. I am a year ten student, and have been to various schools over the last few years. Each one has shown (to some extent) the results above. People with higher amounts of self-discipline do excell. I have seen this at every school I have been to, with every school age group. It's even obvious in my family. We all have high IQ's, so there's nothing there that could change things, but when you look at self discipline, it all changes. My older sister, for example, sits up late at night, gets all her work done, and works through her weekends. I'm a bit lazier. I go to bed early, don't finish assignments, and use my weekends for rest. And, suprisingly enough, she's getting better marks.

i am guessing that to understand a particular subject well, one needs to have a certain amount of IQ AND a certain amount of decipline. i.e. a smarter person will prolly grasp something sooner than a person who is not so smart. so the time spent in learning something is probably going to be less for the smarter person. but dicipline does has its advantages. maybe a better measure of success should be the product of dicipline and IQ rather than just one of them alone.

If high IQ kids learn in elementary school that they don't have to work very hard to do well, why are we surprised when, in middle school the really smart ones haven't learned to work very hard.

By Laura Lynn Walsh (not verified) on 19 Jul 2007 #permalink

Hey, I took an IQ test on IQ Test, and my score was 141, which means that I am a genius. In reality I am a genius. So I can conclude that all Tests which are meant to judge your IQ aren't wrong.

"That research focused on asking the question whether a high (academic) IQ predicts high achievement in the real world. After conducting longitudinal research on 1500 children with IQs greater than 140 ("genius" intelligence), the research revealed that participants could be classified into two groups: the (successful) "A" group had an average income 5 times greater than the general population (in 1955);....."

This is a study of IQ compared to earnings - it's a gigantic leap to therefore directly link IQ to 'success' as success can be measured in a multitude of ways. Our ubiquitous social preponderance of linking money to success may help explain why there aren't enough skilled laborers around! Time to grow up as a society.

Yes, I have a good IQ, yes I got a good degree, yes I was bored half of the time at University, and yet I gained a minuscule percentage of my life success because of it.

By Bob Chippens (not verified) on 29 Sep 2008 #permalink

Has it crossed anyone's mind as to whether GPA is a good measure of a person's value? Maybe being told you're a "C" or "D" student your whole life is a self-fulfilling prophecy. I cannot imagine that a variety of children can be grouped together for 8 hours a day and rated on a variety of activities and that the average rating is the child's value to society.

And does no one recall that IQ tests were created decades ago to determine if a person were mentally handicapped? So having a "high IQ" as measured by these test, just means you're good at doing what mentally retarded people can't do?! Congratulations! You're a champion speed-walker!

I want to bash a lot of these comments...but it seems like a waste of time.
I come from the perspective of having a good deal of military training experience (for a civillian! and probably more than certain commissioned members of the military) and doing better at standardized testing than most and having ADD.
Let me first say, that it is no joke to be ADD in the military! I've also got the non-western concept of timeliness built in, I rarely get things done on time. Am I good at problem-solving? Yeah, you give me a set of facts that have the answer built-in to them, I can give you the answer, like in tests. Beyond the machine-like obedience that all you IQ-snobs imagine that the military instills in you, they instill a concept of professionalism, of preparedness, of leadership. This means that, when you are tasked with doing something, you consider the chain of events that goes into making it happen, and thus must plan and work accordingly. This can be built-in at a young age. As far as I can tell by my experience with people, you need to hold a person to objective standards, or an intimate understanding of the inner-workings of their brain. Given that the latter is damn near impossible (I remember thinking pretty complicated things as a child, it's just that their communication skills are not as well-developed, their common-sense knowledge of the world not-all-there, that is why they ever seem to be dumb. They lack wisdom, to be sure, but they can think and understand a lot.), you should hold your child responsible for little things then work your way up. You think a 5-year old can't pour their own milk? That oh, it's too heavy, their clumsy hands will spill? Well, if you make them do it, and they know they'll get in trouble for spilling, they'll concentrate and get the job done. Yeah, you have to punish your kid, but the more you baby them, the more like a baby they will be as they grow older. If you want them to be able to manage a company as an adult, then they should be able to wake up, get ready and feed themselves in elementary school. It's not nearly impossible; you're just forcing them to concentrate and take responsibility for themselves, which will pay off later. I think I got off easy as the third of 3 kids...I can recall an instance where I spilled a cup at Pizza Hut, but for some reason, my older sister was tasked with cleaning it up. I was old enough to recognize my fault, and probably old enough to clean up most of it, but I was underestimated and lost a chance to learn responsibility. But you probably couldn't tell by looking at me. Moral of the story: hold your kids to objective standards of responsibility. Raise the bar regularly, accept their occasional snafu's, and last, but most importantly, hold them responsible for what you task them to do. If you show that authority figures are BSing to them, they will feel that way about all the rest of the figures they encounter.

As for myself...yeah, I'm working on delayed gratification with the help of stimulant medication. It certainly can work. My hope, however, is to wean myself off of it with exercise and meditation. I found that, getting off the meds after taking them a long time, I still had a lot of memory of working hard, and this got me through a few weeks. But I would get tired...I would stop listening to people after a few words, I could barely get a few words out myself without losing my train of thought. The only effective tool I've found in keeping myself disciplined was a 100% genuine belief that I was alone in the world, that I was the only person who could do my work, that achieving good grades was the only way I could get traction with the system in which we all live. But this, too, fades...
For me, it ultimately comes down to practice..do I have to use forethought and self-discipline to get myself through the day? If no, this part of my brain "atrophies" and does not get developed. However, if I study a lot and rest in combo, then I build up a lot of "charge" in the discipline part of my brain. This charge is incredibly useful for getting stuff done, but on the flipside, it gets to the point where I can't sleep if I didn't use that energy!!
So the key is to balance business and rest.

Comment 2, by Agnostic:

this study covered only eighth graders in a relatively privileged school.

I was thinking this during the whole post, and then here's confirmation. In statistics, this is called "restriction of range" for some variable, here IQ (because the school is privileged, indicating that students of low or average IQ are unlikely to be found in the expected proportions). The less a variable varies, the less power it has to account for some outcome.

In most practical situations where IQ might conceivably be of value - e.g., choosing from candidates for a job or for admission to a university - the "restriction in range" will apply, since the candidates will all be pretty similar. IQ will be a poor predictor in that case, and of little practical value.