More on IQ, WM, and achievement

[Gwern Branwen]

"Sorry, Strivers: Talent Matters"
https://www.nytimes.com/2011/11/20/opinion/sunday/sorry-strivers-talent-matters.html
(Also on Hacker News: http://news.ycombinator.com/item?id=3258655 )

> Research in recent decades has shown that a big part of the answer is simply practice — and a lot of it. In a pioneering study, the Florida State University psychologist K. Anders Ericsson and his colleagues asked violin students at a music academy to estimate the amount of time they had devoted to practice since they started playing. By age 20, the students whom the faculty nominated as the “best” players had accumulated an average of over 10,000 hours, compared with just under 8,000 hours for the “good” players and not even 5,000 hours for the least skilled.

In case people missed the Ericsson stuff the first time:
http://www.gwern.net/docs/1993-ericsson-deliberatepractice.pdf and and
some of the papers in
http://dl.dropbox.com/u/5317066/cambridge-expertise.pdf

> Exhibit A is a landmark study of intellectually precocious youths directed by the Vanderbilt University researchers David Lubinski and Camilla Benbow. They and their colleagues tracked the educational and occupational accomplishments of more than 2,000 people who as part of a youth talent search scored in the top 1 percent on the SAT by the age of 13. (Scores on the SAT correlate so highly with I.Q. that the psychologist Howard Gardner described it as a “thinly disguised” intelligence test.) The remarkable finding of their study is that, compared with the participants who were “only” in the 99.1 percentile for intellectual ability at age 12, those who were in the 99.9 percentile — the profoundly gifted — were between *three and five times more likely* to go on to earn a doctorate, secure a patent, publish an article in a scientific journal or publish a literary work. A high level of intellectual ability gives you an enormous real-world advantage.

One of the articles in http://www.hoagiesgifted.org/longitudinal.htm
(Looks like some interesting linkzs there as well.)

> In our own recent research, we have discovered that “working memory capacity,” a core component of intellectual ability, predicts success in a wide variety of complex activities. In one study, we assessed the practice habits of pianists and then gauged their working memory capacity, which is measured by having a person try to remember information (like a list of random digits) while performing another task. We then had the pianists sight read pieces of music without preparation.
>
> Not surprisingly, there was a strong positive correlation between practice habits and sight-reading performance. In fact, the total amount of practice the pianists had accumulated in their piano careers accounted for nearly half of the performance differences across participants. But working memory capacity made a statistically significant contribution as well (about 7 percent, a medium-size effect). In other words, if you took two pianists with the same amount of practice, but different levels of working memory capacity, it’s likely that the one higher in working memory capacity would have performed considerably better on the sight-reading task.

Pianist study: http://news.msu.edu/media/documents/2011/10/5b176194-ba9a-498d-87c3-c51bc0b1c66b.pdf

Abstract:

> It is clear from decades of research that, to a very large degree, success in music, games, sports, science, and other complex
domains reflects knowledge and skills acquired through experience.
However, it is equally clear that basic abilities, which are
known to be substantially heritable, also contribute to performance
differences in many domains, even among highly skilled
performers. As we discuss here, our research shows that working memory
capacity predicts performance in complex tasks even
in individuals with high levels of domain-specific experience and
knowledge. We discuss implications of our findings for the
understanding of individual differences in skill and identify
challenges for future research.

> Results from our own research are illustrative. In one project (Hambrick, Salthouse, & Meinz, 1999), we had over 800 participants attempt to solve crossword puzzles, and complete tests both of fluid abilities (often referred to as Gf), including reasoning ability and perceptual speed, and of crystallized abilities (often referred to as Gc), including vocabulary, cultural knowledge, and esoteric vocabulary—that is, knowledge of words rarely encountered outside of crossword puzzles, such as aril (a seed covering) and etui (a needle case). Across four studies, there were strong effects of crystallized abilities on puzzle performance, whereas effects of fluid abilities were near zero. Indeed, in one study, the number of clues solved in a difficult New York Times puzzle correlated .87 (very highly) with esoteric vocabulary but near zero with reasoning ability.
>
> In another study (Hambrick, Meinz, Pink, Pettibone, & Oswald, 2010), we investigated the impact of domain knowledge on learning. The participants were approximately 500 undergraduate students, and the study took place in two sessions. In the first session, the participants completed tests of fluid abilities and crystallized abilities, intellectual openness, and interest in and knowledge of politics. Then, after approximately 2 months—and shortly after the 2004 U.S. presidential election—they returned to the lab and took tests to assess knowledge of events surrounding the campaigns and elections that took place after the first session. The major finding was simply that preexisting knowledge of politics was far and away the strongest predictor of knowledge acquired about the campaign.

> However, it is just as clear that basic cognitive abilities are important. General intelligence—which is substantially heritable (Plomin, DeFries, McClearn, & McGuffin, 2008)—is widely regarded as the single best predictor of a number of real-world outcomes, including job per- formance (Schmidt & Hunter, 2004) and academic achievement (Kuncel & Hezlett, 2007), and correlates positively with skill in domains such as chess (Grabner, Stern, & Neubauer, 2007) and music (Ruthsatz, Detterman, Griscom, & Cirullo, 2008). Lubinski, Benbow, and colleagues (see Robertson, Smeets, Lubinski, & Benbow, 2010) have even documented that individual differences within the top one percent of cognitive ability predict individual differences in scientific achievement. For example, children who scored in the 99.9th percentile on the math section of the SAT by age 13 were found to be eighteen times more likely to go on to earn a PhD in a Science, Technology, Engineering, and Mathematics (STEM) discipline than children who “only” scored in the 99.1th percentile. In his bestselling book Outliers, alluding to the idea that intelligence is a “threshold” variable (Torrance, 1962), Malcolm Gladwell (2008) commented that “the relationship between success and IQ works only up to a point. Once someone has reached an IQ of somewhere around 120, having additional IQ points doesn’t seem to translate into any measurable real-world advantage.” In his own bestselling book, The Social Animal, David Brooks (2011) expressed the same idea: “A person with a 150 IQ is in theory much smarter than a person with a 120 1Q, but those additional 30 points produce little measurable benefit when it comes to lifetime success” (p. 165). Malcolm Gladwell and David Brooks are simply wrong. A high level of intellectual ability puts a person at a measurable advantage in science—and the higher the better.

This is very relevant to issues of *transfer* and playing games, and
what we've already seen about chess:

> It seems, then, that individual differences in performance on many complex tasks arise from both acquired characteristics and basic abilities. With this generalization as our starting point, a major goal of our collaborative research, which we began as graduate students working with Timothy Salthouse, has been to investigate the interplay between these two types of factors. We have been especially interested in the question of whether various forms of domain knowledge moderate the impact of basic cognitive abilities on performance. More concretely, as illustrated in Figure 1, we have asked whether relationships between basic abilities and complex task performance are weaker at high levels of domain knowledge (red line) than at lower levels (blue line), as evidenced by Ability × Knowledge interactions.
>
> This possibility, which we refer to as the circumvention-of-limits hypothesis, has been mentioned by a number of theorists. Ackerman (1988) proposed that general intelligence is important in the initial stage of skill acquisition, when it is necessary to hold steps of executing a skill in the focus of attention, but then drops out as a predictor of performance as knowledge is proceduralized. Similarly, Ericsson and colleagues have argued that the acquisition of knowledge and skills through deliberate practice—engagement in activities specifically designed to improve performance in a domain (Ericsson, Krampe, & Tesch-Römer, 1993)—enables circumvention of performance limitations associated with basic abilities: “The effects of extended deliberate practice are more far-reaching than is commonly believed. Performers can acquire skills that circumvent basic limits on working memory capacity. . .” (Ericsson & Charness, 1994, p. 725).

> In one study (Hambrick & Engle, 2002), we had participants complete tests of baseball knowledge and complex-span tasks and then perform a complex memory task in which they listened to fictitious radio broadcasts of baseball games and tried to remember both the major events of the games and information about the players. Not surprisingly, domain knowledge (of baseball) had a very strong positive effect on memory performance. However, there was also a positive effect of working-memory capacity and no evidence for a Working-Memory Capacity × Domain Knowledge interaction. Working-memory capacity was as important as a predictor of memory performance at high levels of domain knowledge as it was at low levels. Similarly, in a study of skill in Texas Hold’em poker (Meinz et al., 2011), we found positive effects of both domain knowledge (of poker) and working-memory capacity on performance of a hand evaluation task, in which the goal was to evaluate the likelihood of drawing a card that would enable a win, and a game memory task, in which the goal was to remember hands in a game of Hold’em. But once again, there was no evidence for Working-Memory Capacity × Domain Knowledge interactions.

> However, we have wondered whether deliberate practice is *sufficient* to account for individual differences in skilled performance—or just necessary....To answer this question, we designed a study to find out whether working-memory capacity would predict piano sight-reading ability (i.e., the ability to play pieces with no preparation) even among individuals with thousands of hours of deliberate practice. Fifty-seven pianists representing a wide range of cumulative deliberate practice—from 260 to over 31,000 hours—performed a battery of complex-span tasks to assess working-memory capacity along with a sight-reading task. Not surprisingly, we found that deliberate practice was a powerful predictor of sight-reading performance. In fact, it accounted for nearly 50% of the variance. However, we also found that working-memory capacity was a positive predictor of performance above and beyond deliberate practice. Furthermore, as illustrated in Figure 2, there was no evidence for a Deliberate Practice × Working-Memory Capacity interaction—and thus no evidence that high levels of deliberate practice reduced the effect of working-memory capacity on performance. This was true even when we used deliberate practice *devoted specifically to sight-reading* as a predictor variable.
>
> We speculated that working-memory capacity limits the number of notes the player can look ahead in the piece of music he or she is playing—a factor that has been shown to correlate very strongly with sight-reading skill. More generally, we argued that although deliberate practice may well be necessary to reach a very high level of skill, it is not always sufficient. Campitelli and Gobet (2011, this issue) summarize evidence from their own research that leads to this same conclusion. The general conclusion that we draw from this and the other studies reviewed here is that a high level of domain knowledge doesn’t guarantee circumvention of limits associated with basic abilities. Basic abilities matter for novice performance, and sometimes they matter for expert performance.

--
gwern
http://www.gwern.net

[Pontus Granström]

I do not find this surprising, and it's also a good motivator for continuing brain training since it leads to real world benefits.

--
You received this message because you are subscribed to the Google Groups "Dual N-Back, Brain Training & Intelligence" group.
To post to this group, send email to brain-t...@googlegroups.com.
To unsubscribe from this group, send email to brain-trainin...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/brain-training?hl=en.

[polar]

Well, "its only one, suggestive social-science study" :) - so I
believe gwern, you will think of it as critically as you do with
Jaeggi 2008. I read this study some time ago, and it is interesting,
but too much dedicated to praise iq, "talent" and their heritability.
One book, altough not contrary finding, but counterbalancing one,
comes to my mind now (its a great read :) http://polarpage.sk/performanceAddiction.pdf

On 21 lis, 01:15, Gwern Branwen <gwe...@gmail.com> wrote:
> "Sorry, Strivers: Talent Matters"https://www.nytimes.com/2011/11/20/opinion/sunday/sorry-strivers-tale...

> (Also on Hacker News:http://news.ycombinator.com/item?id=3258655)
>
> > Research in recent decades has shown that a big part of the answer is simply practice — and a lot of it. In a pioneering study, the Florida State University psychologist K. Anders Ericsson and his colleagues asked violin students at a music academy to estimate the amount of time they had devoted to practice since they started playing. By age 20, the students whom the faculty nominated as the “best” players had accumulated an average of over 10,000 hours, compared with just under 8,000 hours for the “good” players and not even 5,000 hours for the least skilled.
>

> In case people missed the Ericsson stuff the first time:http://www.gwern.net/docs/1993-ericsson-deliberatepractice.pdfand and
> some of the papers inhttp://dl.dropbox.com/u/5317066/cambridge-expertise.pdf

>
> > Exhibit A is a landmark study of intellectually precocious youths directed by the Vanderbilt University researchers David Lubinski and Camilla Benbow. They and their colleagues tracked the educational and occupational accomplishments of more than 2,000 people who as part of a youth talent search scored in the top 1 percent on the SAT by the age of 13. (Scores on the SAT correlate so highly with I.Q. that the psychologist Howard Gardner described it as a “thinly disguised” intelligence test.) The remarkable finding of their study is that, compared with the participants who were “only” in the 99.1 percentile for intellectual ability at age 12, those who were in the 99.9 percentile — the profoundly gifted — were between *three and five times more likely* to go on to earn a doctorate, secure a patent, publish an article in a scientific journal or publish a literary work. A high level of intellectual ability gives you an enormous real-world advantage.
>

> One of the articles inhttp://www.hoagiesgifted.org/longitudinal.htm

> (Looks like some interesting linkzs there as well.)
>
> > In our own recent research, we have discovered that “working memory capacity,” a core component of intellectual ability, predicts success in a wide variety of complex activities. In one study, we assessed the practice habits of pianists and then gauged their working memory capacity, which is measured by having a person try to remember information (like a list of random digits) while performing another task. We then had the pianists sight read pieces of music without preparation.
>
> > Not surprisingly, there was a strong positive correlation between practice habits and sight-reading performance. In fact, the total amount of practice the pianists had accumulated in their piano careers accounted for nearly half of the performance differences across participants. But working memory capacity made a statistically significant contribution as well (about 7 percent, a medium-size effect). In other words, if you took two pianists with the same amount of practice, but different levels of working memory capacity, it’s likely that the one higher in working memory capacity would have performed considerably better on the sight-reading task.
>

> Pianist study:http://news.msu.edu/media/documents/2011/10/5b176194-ba9a-498d-87c3-c...

>
> Abstract:
>
> > It is clear from decades of research that, to a very large degree, success in music, games, sports, science, and other complex
>
> domains reflects knowledge and skills acquired through experience.
> However, it is equally clear that basic abilities, which are
> known to be substantially heritable, also contribute to performance
> differences in many domains, even among highly skilled
> performers. As we discuss here, our research shows that working memory
> capacity predicts performance in complex tasks even
> in individuals with high levels of domain-specific experience and
> knowledge. We discuss implications of our findings for the
> understanding of individual differences in skill and identify
> challenges for future research.
>
> >  Results from our own research are illustrative. In one project (Hambrick, Salthouse, & Meinz, 1999), we had over 800 participants attempt to solve crossword puzzles, and complete tests both of fluid abilities (often referred to as Gf), including reasoning ability and perceptual speed, and of crystallized abilities (often referred to as Gc), including vocabulary, cultural knowledge, and esoteric vocabulary—that is, knowledge of words rarely encountered outside of crossword puzzles, such as aril (a seed covering) and etui (a needle case). Across four studies, there were strong effects of crystallized abilities on puzzle performance, whereas effects of fluid abilities were near zero. Indeed, in one study, the number of clues solved in a difficult New York Times puzzle correlated .87 (very highly) with esoteric vocabulary but near zero with reasoning ability.
>
> > In another study (Hambrick, Meinz, Pink, Pettibone, & Oswald, 2010), we investigated the impact of domain knowledge on learning. The participants were approximately 500 undergraduate students, and the study took place in two sessions. In the first session, the participants completed tests of fluid abilities and crystallized abilities, intellectual openness, and interest in and knowledge of politics. Then, after approximately 2 months—and shortly after the 2004 U.S. presidential election—they returned to the lab and took tests to assess knowledge of events surrounding the campaigns and elections that took place after the first session. The major finding was simply that preexisting knowledge of politics was far and away the strongest predictor of knowledge acquired about the campaign.
> > However, it is just as clear that basic cognitive abilities are important. General intelligence—which is substantially heritable (Plomin, DeFries, McClearn, & McGuffin, 2008)—is widely regarded as the single best predictor of a number of real-world outcomes, including job per- formance (Schmidt & Hunter, 2004) and academic achievement (Kuncel & Hezlett, 2007), and correlates positively with skill in domains such as chess (Grabner, Stern, & Neubauer, 2007) and music (Ruthsatz, Detterman, Griscom, & Cirullo, 2008). Lubinski, Benbow, and colleagues (see Robertson, Smeets, Lubinski, & Benbow, 2010) have even documented that individual differences within the top one percent of cognitive ability predict individual differences in scientific achievement. For example, children who scored in the 99.9th percentile on the math section of the SAT by age 13 were found to be eighteen times more likely to go on to earn a PhD in a Science, Technology, Engineering, and Mathematics (STEM) discipline than children who “only” scored in the 99.1th percentile. In his bestselling book Outliers, alluding to the idea that intelligence is a “threshold” variable (Torrance, 1962), Malcolm Gladwell (2008) commented that “the relationship between success and IQ works only up to a point. Once someone has reached an IQ of somewhere around 120, having additional IQ points doesn’t seem to translate into any measurable real-world advantage.” In his own bestselling book, The Social Animal, David Brooks (2011) expressed the same idea: “A person with a 150 IQ is in theory much smarter than a person with a 120 1Q, but those additional 30 points produce little measurable benefit when it comes to lifetime success” (p. 165). Malcolm Gladwell and David Brooks are simply wrong. A high level of intellectual ability puts a person at a measurable advantage in science—and the higher the better.
>
> This is very relevant to issues of *transfer* and playing games, and
> what we've already seen about chess:
>
>
>
> > It seems, then, that individual differences in performance on many complex tasks arise from both acquired characteristics and basic abilities. With this generalization as our starting point, a major goal of our collaborative research, which we began as graduate students working with Timothy Salthouse, has been to investigate the interplay between these two types of factors. We have been especially interested in the question of whether various forms of domain knowledge moderate the impact of basic cognitive abilities on performance. More concretely, as illustrated in Figure 1, we have asked whether relationships between basic abilities and complex task performance are weaker at high levels of domain knowledge (red line) than at lower levels (blue line), as evidenced by Ability × Knowledge interactions.
>
> > This possibility, which we refer to as the circumvention-of-limits hypothesis, has been mentioned by a number of theorists. Ackerman (1988) proposed that general intelligence is important in the initial stage of skill acquisition, when it is necessary to hold steps of executing a skill in the focus of attention, but then drops out as a predictor of performance as knowledge is proceduralized. Similarly, Ericsson and colleagues have argued that the acquisition of knowledge and skills through deliberate practice—engagement in activities specifically designed to improve performance in a domain (Ericsson, Krampe, & Tesch-Römer, 1993)—enables circumvention of performance limitations associated with basic abilities: “The effects of extended deliberate practice are more far-reaching than is commonly believed. Performers can acquire skills that circumvent basic limits on working memory capacity. . .” (Ericsson & Charness, 1994, p. 725).
> > In one study (Hambrick & Engle, 2002), we had participants complete tests of baseball knowledge and complex-span tasks and then perform a complex memory task in which they listened to fictitious radio broadcasts of baseball games and tried to remember both the major events of the games and information about the players. Not surprisingly, domain knowledge (of baseball) had a very strong positive effect on memory performance. However, there was also a positive effect of
>

> ...
>
> číst dál »

[Aman Idle]

Does anybody remember the site to that mental something game that was in german please? I would appreciate it, if someone could let me know. :)

[Mike]

Thanks Gwern, very interesting.

This completely confirms my intuition about my poor sight reading. it requires a high WM and concentration. I always hated it since I was a kid. I know there's a problem with my brain lol (it's not WM, but focus in my case).

(doing KhanAcademy math exercises (I'm almost done, about 10 more to go) in the last 2 weeks really made me notice my inattention mistakes and my ADD again.)

here's some lengthy brainstorming. feel free to read and answer.

==

From what I feel, and from the articles you provide, I think we could say that:

• deliberate practice, (and practice of sight reading) improves sight reading up to a point--there would be a limit on what practice can do. the WM component of the task stays the same for an given individual.
• so the effect of practice on sight reading performance (for this particular task, sight reading--which I always felt required an especially high WM and concentration) probably resembles a logarithmic curve with a limit, to which we add a constant, WM capacity (and concentration), to get the "measure of performance (sight reading grade)".

I think to get a better model of the task, we could imagine WM manipulating chunks of musical structures notes. As the pianist improves (learns new pieces, techniques and chords) he recognizes bigger and more complex chunks (recognizing intervals, then chords voicings, typical chord sequences, typical segments of melodies, etc.--instead of reading note by note) and can read ahead more with the same WM capacity, giving him more comfort. Since musical pieces always have novelty in them, deliberate practice (learning chunks, eg voicings, chords sequences, melodies) can only improve performance up to a point (although: for jazz chunks are very high --> practice helps more there, for J.S. Bach chunks are lower --> you need more WM there). I think this describes better the interaction between WM and LTM/experience on that task, in computational terms.

I think what's crucial here is that the following model:

Ackerman (1988) proposed that general intelligence is important in the initial stage of skill acquisition, when it is necessary to hold steps of executing a skill in the focus of attention, but then drops out as a predictor of performance as knowledge is proceduralized.

can be true on some tasks, but less for others (depending on the amount of novelty involved). some skills can be proceduralized more than others (eg: jazz improvisation, driving a car VS sight reading some J.S. Bach). I think what is crucial is that people with very high IQs succeed at those harder tasks that involve more novelty processing--the other tasks -> anyone can become excellent at them, so they wouldn't be called achievements. high IQs (and especially those with high WM in my opinion) can go where others can't in dealing with high novelty environment, that's all. the fact that high IQs really is linked to achievement doesn't mean that skills aren't proceduralizable, it only means that achievement doesn't mean being good at easily proceduralizable tasks. achievement usually requires proceduralizing a lot of skills and coordinating them together in ways that few people can do. so a high IQ will always be needed to:

• proceduralize more skills per minute throughout your life
• proceduralize skills that most people can't proceduralize because they require a lot of previous proceduralization and a lot of concentration

summary: obviously higher IQ people will achieve more and faster. but it doesn't mean that all skills aren't proceduralizable --> a lot of them are. eventually a low IQ can match performance of a high IQ on some very hard tasks (if the task doesn't involve too much raw WM like sight reading some J.S. Bach) with enough learning and repetition. only it will take too long for them to catch up with higher IQs. Also some tasks just require some raw WM (sight reading hard pieces), they are less "chunkable" or "reducible to proceduralization" than other skills. maybe sight reading was not the best pick to prove such proceduralization of skills, and a supposed lesser influence of IQ on performance over time.

what's interesting is that on KhanAcademy data graphs (exercise progress over time), there are different patterns of achievement: different types of giftedness. some kids skyrocket then plateau, while some others take a longer time to start a steady increase, as if they had a sudden revelation--they figured out something--and they finally catch up and even out do to the initially more gifted kids. why?

http://www.khanacademy.org/images/screenshot-tour/class-small.png 

my theory about this is that:

• the kids that skyrocket then plateau have a high WM and deal easily with novelty.
• the kids who take a longer time before starting a steadier increase don't deal well with novelty. their brain chemistry is different or altered, --> it's in between the gifted and non-gifted, BUT they seem to find a way to overcome their initial limitation on some aspects of the tasks--they find a trick. I think what happens is that they get good at learning "lessons" from exercises, ideas that can later be applied to other exercises. I think this is the most plausible explanation--they store and cross-link reusable chunks of knowledge more than the super high WM guys do at first (even though they can too, they just don't need to at first).

what else could explain these different curves of improvement? 

• Do you think it's possible that some kids excel at novel tasks but fail to really store "lessons" that can be applied elsewhere?
• Do you think it's possible that some kids are not good at dealing with novel problems, but good at identifying and reusing some "tricks" and "lessons" learned?

I think it's a plausible explanation. this also corresponds to what I always noticed: that the kids who were better at musical sight reading were not necessarily good at understanding chords sequences and voicings and understanding how melodies work in relation with chords (although I'm sure they could have extracted that knowledge/understanding but they didn't need to). they definitely had a different way of learning than me. they obviously had a very high WM but just didn't make the effort to identify lessons from the pieces they read and learned. on the other hand, maybe because I had a lower WM (or maybe it's just a correlation), I learned very early to "compress" a piece, identify tricks (chords, typical melodies etc.) and understand what the structure was, and reuse those tricks.

So learning styles would exist. (?). could some people learn slower but then retain a lessons better? 

if yes, what could cause this? could there be 2 different intelligence systems in the brain, that are relatively independent? or could it be that some variables of brain chemistry eventually tweak how a gifted brain (high g factor) behaves on the long term?

I don't know, but it's interesting to notice that:

• being on coffee constantly is a short term booster -> it would make you excel at new tasks, but your neuronal growth factors are lessened so potentially less learning and cross linking of info on the long term: you would meet a plateau on khan academy.
• being sleepy constantly (and having sleepy brain chemistry all the time) would make you less good at dealing with novelty but better at learning and linking new experiences and retrieving tricks, lessons, summaries--reusable lessons.

maybe some people are more in one category or another, and this creates learning styles.

please comment on this. what else could cause these different patterns of progression on KhanAcademy? 

• (is it just a coincidence? or social factors, like the more gifted kids let themselves be catched up to in order to make friends and not be isolated too much?)
• or maybe you think that you have sleepy or hyper high iqs and that it doesn't affect their learning and their iq at all.
• also I read that certain conditions or learning deficits affect only a part of the brain/learning process (eg asperger, add, adhd, low LI, etc.). so if gifted, these people would cope differently and have different learning styles, it's possible.

maybe that for a gifted person you should try to optimize your brain chemistry, and/or force yourself to follow particular study methods to boost proceduralization. maybe once you have have the essential genetic factors for high IQ, having a perfectly balanced brain chemistry and style is the difference between highly gifted and high achievers (150 iq?), and just mediocre gifted people (130 iq?).

also there are different types of high achievers: entrepreneurs, university teachers, etc. and they don't seem to have the same personality types and same baseline neurotransmitter chemistry. maybe just picking the right field that matches your brain chemistry is essential.

what would be very interesting would be to follow the progression of kids/people on khanacademy or similar exercises, while knowing their iqs and knowing their brain chemistry or personality.

==

I feel like I improved a bit at sight reading after DNB, I should try to really practice it systematically now. I wonder if I could become a good sight reader after all those months of DNB training. I never practiced sight reading itself, only memorization of pieces, composing and improvising.

I have strong clues that I have ADD inattentive or SCT (sluggish cognitive tempo, or something similar, although I look really normal and intelligent), yet I could reach relatively high levels on DNB (reach 9-10 usually under 20 sessions, max 13, with default settings in BW). I know that chunking and other strategies come into play in that game, but it still is an indicator that I have at least a normal WM. I think my brain problem is that I can't keep focus (which leads to inattention mistakes, frequent daydreaming, and eventually creativity), which is a different component than WM capacity per se, (it can affect outcomes on WM tests, obviously, but it's not the same component as raw storage size). I'm almost sure it's a jump in focus that is the problem, a randomness that I can't control very well. (ie inattention mistakes, that happen every few seconds, eg 7*4=11, mistakes that happen more often than for most people, and it stresses me out in some situations, social settings and conversations, sight reading, etc.--eg I learned to filter out what I say, when I was a kid I could occasionally say far fetched things in a conversation, mildly embarrassing, eg appear clumsy, forgetful or missing important sentences someone said in a conversation, while still appearing generally normal, intelligent and creative).

any comments welcome.

[J.]

@Mike

I think it may come down to the fact that DNB is primarily a
recognition and not a retrieval task. The information may be in your
brain for prompting, but if nothing cues it, it doesn't come out,
making it appear like a WM deficit/poor attention. Sight reading
obviously isn't a matter of recognition, it's a matter of retrieval.
Many of the other aspects of music you mentioned however, seem to be
more recognition based, which might explain why they were so easy for
you. Just some thoughts.

On Nov 22, 12:54 pm, Mike <mikebk...@gmail.com> wrote:
> Thanks Gwern, very interesting.
>
> This completely confirms my intuition about my poor sight reading. it
> requires a high WM and concentration. I always hated it since I was a kid.
> I know there's a problem with my brain lol (it's not WM, but focus in my
> case).
> (doing KhanAcademy math exercises (I'm almost done, about 10 more to go) in
> the last 2 weeks really made me notice my inattention mistakes and my ADD
> again.)
>
> here's some lengthy brainstorming. feel free to read and answer.
>
> ==
> From what I feel, and from the articles you provide, I think we could say
> that:
>

>    - deliberate practice, (and practice of sight reading) improves sight

>    reading up to a point--there would be a limit on what practice can do. the
>    WM component of the task stays the same for an given individual.

>    - so the effect of practice on sight reading performance (for this

>    particular task, sight reading--which I always felt required an especially
>    high WM and concentration) probably resembles a logarithmic curve with a
>    limit, to which we add a constant, WM capacity (and concentration), to get
>    the "measure of performance (sight reading grade)".
>

> I think to get a better model of the task, *we could imagine WM

> manipulating chunks of musical structures notes. As the pianist improves
> (learns new pieces, techniques and chords) he recognizes bigger and more

> complex chunks (**recognizing intervals, then **chords voicings, typical

> chord sequences, typical segments of melodies, etc.--instead of reading

> note by note)* and can read ahead more with the same WM capacity, giving

> him more comfort. Since musical pieces always have novelty in them,
> deliberate practice (learning chunks, eg voicings, chords sequences,
> melodies) can only improve performance up to a point (although: for jazz
> chunks are very high --> practice helps more there, for J.S. Bach chunks
> are lower --> you need more WM there). I think this describes better the
> interaction between WM and LTM/experience on that task, in computational
> terms.
>
> I think what's crucial here is that the following model:
>
> Ackerman (1988) proposed that general intelligence is important in the
> initial stage of skill acquisition, when it is necessary to hold steps of
> executing a skill in the focus of attention, but then drops out as a
> predictor of performance as knowledge is proceduralized.
>

> can be *true on some tasks, but less for others (depending on the amount of
> novelty involved)*. some skills can be proceduralized more than others (eg:

> jazz improvisation, driving a car VS sight reading some J.S. Bach). I think
> what is crucial is that people with very high IQs succeed at those harder
> tasks that involve more novelty processing--the other tasks -> anyone can
> become excellent at them, so they wouldn't be called achievements. high IQs
> (and especially those with high WM in my opinion) can go where others can't

> in dealing with high novelty environment, that's all. *the fact that high

> IQs really is linked to achievement doesn't mean that skills aren't
> proceduralizable, it only means that achievement doesn't mean being good at

> easily proceduralizable tasks*. achievement usually requires

> proceduralizing a lot of skills and coordinating them together in ways that
> few people can do. so a high IQ will always be needed to:
>

>    - proceduralize more skills per minute throughout your life
>    - proceduralize skills that most people can't proceduralize because they

>    require a lot of previous proceduralization and a lot of concentration
>

> *summary*: obviously higher IQ people will achieve more and faster. but it

> doesn't mean that all skills aren't proceduralizable --> a lot of them are.

> eventually a low IQ can match performance of a high IQ on *some *very hard

> tasks (if the task doesn't involve too much raw WM like sight reading some
> J.S. Bach) with enough learning and repetition. only it will take too long

> for them to catch up with higher IQs. *Also some tasks just require some

> raw WM (sight reading hard pieces), they are less "chunkable" or "reducible

> to proceduralization" than other skills.* maybe sight reading was not the
> best pick to prove such proceduralization of skills, and a supposed *lesser
> influence of IQ on performance* over time.

>
> what's interesting is that on KhanAcademy data graphs (exercise progress
> over time), there are different patterns of achievement: different types of

> giftedness. *some kids skyrocket then plateau, while some others take a

> longer time to start a steady increase, as if they had a sudden
> revelation--they figured out something--and they finally catch up and even

> out do to the initially more gifted kids*. why?http://www.khanacademy.org/images/screenshot-tour/class-small.png

> my theory about this is that:
>

>    - the kids that skyrocket then plateau have a high WM and deal easily
>    with novelty.
>    - the kids who take a longer time before starting a steadier increase

>    don't deal well with novelty. their brain chemistry is different or
>    altered, --> it's in between the gifted and non-gifted, BUT they seem to
>    find a way to overcome their initial limitation on some aspects of the

>    tasks--they find a trick. *I think what happens is that they get good at

>    learning "lessons" from exercises, ideas that can later be applied to other

>    exercises*. I think this is the most plausible explanation--they store

>    and cross-link reusable chunks of knowledge more than the super high WM
>    guys do at first (even though they can too, they just don't need to at
>    first).
>
> what else could explain these different curves of improvement?
>

>    - Do you think it's possible that some kids excel at novel tasks but

>    fail to really store "lessons" that can be applied elsewhere?

>    - Do you think it's possible that some kids are not good at dealing with

>    novel problems, but good at identifying and reusing some "tricks" and
>    "lessons" learned?
>
> I think it's a plausible explanation. this also corresponds to what I

> always noticed: *that the kids who were better at musical sight reading

> were not necessarily good at understanding chords sequences and voicings
> and understanding how melodies work in relation with chords (although I'm
> sure they could have extracted that knowledge/understanding but they didn't

> need to)*. they definitely had a different way of learning than me. they

> obviously had a very high WM but just didn't make the effort to identify
> lessons from the pieces they read and learned. on the other hand, maybe
> because I had a lower WM (or maybe it's just a correlation), I learned very
> early to "compress" a piece, identify tricks (chords, typical melodies
> etc.) and understand what the structure was, and reuse those tricks.
>

> So learning styles would exist. (?). *could some people learn slower but
> then retain a lessons better?*

> if yes, what could cause this? could there be 2 different intelligence
> systems in the brain, that are relatively independent? or could it be that
> some variables of brain chemistry eventually tweak how a gifted brain (high
> g factor) behaves on the long term?
> I don't know, but it's interesting to notice that:
>

>    - *being on coffee constantly* is a short term booster -> it would make

>    you excel at new tasks, but your neuronal growth factors are lessened so
>    potentially less learning and cross linking of info on the long term: you
>    would meet a plateau on khan academy.

>    - *being sleepy constantly* (and having sleepy brain chemistry all the

>    time) would make you less good at dealing with novelty but better at
>    learning and linking new experiences and retrieving tricks, lessons,
>    summaries--reusable lessons.
>
> maybe some people are more in one category or another, and this creates
> learning styles.
> please comment on this. what else could cause these different patterns of
> progression on KhanAcademy?
>

>    - (is it just a coincidence? or social factors, like the more gifted

>    kids let themselves be catched up to in order to make friends and not be
>    isolated too much?)

>    - or maybe you think that you have sleepy or hyper high iqs and that it

>    doesn't affect their learning and their iq at all.

>    - also I read that certain conditions or learning deficits affect only a

> problem is that I can't keep focus (which leads ...
>
> read more »

[eurekasabiduria]

kaj kion vi pensas pri la laboro de sinjoro lazlo polgar

http://en.wikipedia.org/wiki/L%C3%A1szl%C3%B3_Polg%C3%A1r

?

[Mike]

@Joseph

I think recognition VS retrieval is interesting. I think it's the contrary that causes my poor sight reading, I think I have a defective recognition (I suppose because of a defective attention--jump in focus every few seconds--although it could be something else), but once I recognized something I can retrieve it and play it without problem and actually better than most people from memory (playing is retrieving... but also involves attention to an extent--so... I do make inattention mistakes while playing classical pieces more than the average--but... it's not that bad. my retrieving from LTM seems to suffer better the jumps in focus, while sight reading is very much affected by it, it can't run at all if there are jumps in focus). I think my "jump in focus" kills my sight reading (recognition) and prevents me from identifying structures reliably enough to keep a tempo, but my playing (retrieving) survives it (probably because whole complex gestural patterns can be activated and carried out at a very low WM/attention cost, that's just how the brain works). So in a way my "learning style" was the result of a natural selection of tasks going on in my brain: some tasks could survive better my "jump in focus" handicap, so I overdeveloped them--that's why I overwhelmingly rely on memorization, and compression of information to play the piano. that's why I naturally became good at composing and improvising --> it was a side effect of having more musical structures available in my gestural memory, and chunked in smaller reusable pieces and networks than most people would care to extract, classify and cross-link in their gestural memory. most classical performers don't really know how to compose a melody even after years of playing the piano (this doesn't apply to everyone though). in my case I could do that decently after only a few months of playing when I was a kid. this really shows we have a different way to approach the problem, cognitively, we don't rely on the same processes.

I thought about this recognitionVSretrieval way of seeing the problem before too.

• when I read a sheet, I have to slowly recognize the structures (recognition of structures is painful), associate them to something I know ("re-express it in my terms", with my previously learned concepts, almost like re-construct it-- like: aah this is simply this ok now I can play it easily), (retrieve from memory--no problem). I think anyone does have to recognize known patterns while sight reading, but my process is slower and more extreme, in that I have to really recognize everything and internalize it, before I can play it--while most people can not fully recognize a chord and play it just fine, last minute. I have to know and expect what's next in my terms, I can't rely on my attention/recognition AT ALL--it's defective. my recognition is not working properly, probably because of these "jumps in focus" I observe I have every few seconds.
• when I improvise, I have no problem retrieving melodic patterns, chords etc (retrieve--no problem). in fact that's the only way I can play a piece--by retrieving and re-organizing patterns from LTM. this is what forced me to know and understand how musical structures work together usually, very early on my own.

I think DNB is a bit like sight reading--it involves recognition and retrieving from WM, only the recognition is easier than for sight reading because it is simply letters (easier to recognize than musical structures). when I'm doing DNB, I feel like what I'm really working on and improving is mostly my capacity to sustain attention...avoid jumps in focus for a longer time. I can use my WM fully for longer stretches, and the jumps in focus can be kept smaller and not ruin the task. I can feel that. improving sustained focus definitely lowered my inattention mistakes rate while playing. I might also have improved my raw WM storage size, too, it's likely. 

this feeling of having trouble sustaining attention for many seconds at a time is the confirmation that a jump in focus is the root of my brain problem. bad recognition (sight reading) stems from it. good retrieving (playing and improvising musical pieces) can be carried out simply because it doesn't rely too much on it and I overdeveloped it as a coping strategy.

it's possible that the kids that take a while to catch up to the really gifted kids on khan academy class progress graphs have a condition like me, where they are not fast--something prevents them from dealing with novelty well--but once they started chunking the information in reusable parts and cross-linking it (with the parts of their gifted brain that works properly), they catch up to the more gifted and can even outdo them on some tasks/problems. at least that's what happened to me while learning the piano, maybe I'm transferring too much. maybe these kids just got a sudden burst of motivation that's all. but still, I have to state that quite a lot of people have an IQ subtest that is much lower than the rest, in other words a learning disability--and I think this can sometimes give rise to weird learning styles, where some people do more cross-linking.

I'm happy that I could identify in this brainstorming that I really have a jump in focus. it seems to get clearer every time I analyze it. at least I know what's the problem. I wonder if this habits for chunking and storing musical structures is the real and initial reason why I have this pathological drive to discuss and dissect what I read, and accumulate information. it's as if I developed these networks for playing the piano and now I apply them to other activities in my life without really being conscious of what my brain is doing. 

I wonder how it is to have had a very good sustained attention for your whole life while having a high IQ, it must change your personality a lot. you probably become more conventional--problems are mostly easy and evident, and you don't rely as much on your crooked personal theories, you just become very successful at doing what people ask from you. I think this might lead to more happiness, socially and especially early and lasting school success.

Again, any comments welcome ;)

[eurekasabiduria]

What do you think about Mr. Lazlo Pogar's work?

http://en.wikipedia.org/wiki/L%C3%A1szl%C3%B3_Polg%C3%A1r

[Gwern Branwen]

From the Wikipedia article:

> László Polgár (born 1946 in Gyöngyös), is a Hungarian chess teacher and father of the famous "Polgár sisters": Zsuzsa, Zsófia, and Judit. He authored well-known chess books such as Chess: 5334 Problems, Combinations, and Games and Reform Chess, a survey of chess variants. László is an expert on chess theory and owns over 10,000 chess books. He is interested in the proper method of rearing children, believing that "geniuses are made, not born". Before he had any children, he wrote a book entitled Bring Up Genius!, and sought a wife to help him carry out his experiment. He found one in Klara, a schoolteacher, who lived in a Hungarian-speaking enclave in Ukraine. He married her in the USSR and brought her to Hungary. He home-schooled their three daughters, primarily in chess, and all three went on to become strong players.

I think his belief that geniuses can be made is about as amusing and
well-proven as Boris Sidis's similar belief and his son William James
Sidis.

--
gwern
http://www.gwern.net

[cev]

The kind of people who are capable of carrying out these experiments
in child-rearing successfully, let alone dream them up in the first
place, are, I would imagine, rather intelligent, and thus more likely
to produce able offspring.

[γενβιρΟ]

Given what is currently known with regards to environmental enrichment
and genetics, I would say that this is a fair statement.

e.g. It is likely that 0 can be come a 1 but less so in regards to it
becoming a 2 or 3. The same for 2, it is likely that it can become a 3
but less so in regards to it becoming a 4 or a 5. Ergo, thresholds of
human potential.

[Mike]

Definitely cultural enrichment could improve your intelligence in theory, especially when you're a kid because the brain is more plastic at that time of life, and optimized for learning. but it doesn't seem to affect adulthood IQ as much as Polgar and Boris Sidis believed.

I think cultural enrichment can give a boost in childhood IQ, but probably won't give a big difference in adulthood IQ. you can stuff more information in that little brain, and it's probably going to become more intelligent than its peers on the short run, but on the long run it seems like your brain's fixed (by genetics+nutrition in the womb) info processing capacities (WM size, processing speed, ...etc?) will determine pretty accurately what amount of information you accumulate and master after say, 2 decades. 

it's like tossing a dice or a coin every minute of your life --> even if you cheat a bit during the first years of your life (get cultural enrichment, (which is great I think btw)) you will tend to reach, sooner or later, the average predicted by the feature of the dice or coin that you're tossing (your WM and processing speed for example, or your g factor). it seems that you can help the little brain to a small extent but it will still have to digest what you teach him on its own, and also observe and think on its own for most of the minutes in a day--and you can't really help it with that.

following that logic I think that DNB (+excelent nutrition +getting an optimal amount of sleep every day) would bear more promise than cultural enrichment because they affect your processing capacities, at every-single-moment-in-time. I would still suggest cultural enrichment. especially playing the piano early (5) which is really the only "cultural enrichment" that has been shown to raise adult IQ by a few points. it's interesting to notice that learning the piano resembles playing DNB in many respects. probably most tasks or games that require to calculate or process or remember things within a fixed amount of time (or a tempo) would have the potential to improve the basic factors intelligence by a little bit. the only ones I know so far that fit this description are DNB and playing an instrument (I recommend the piano).

I think adult intelligence is genetically heritable (at .7 I think in twin studies?). environment plays a role, but maybe not only cultural enrichment. maybe more nutrition in the womb. maybe the quality of your tap water will affect a kid's intelligence more than cultural enrichment. I think it's possible that providing optimal amounts of nutriments and fats to the baby in the womb and a minimal amount of toxins represents the bulk of the "environmental factors" of adult intelligence.

in the case of twin studies --> I know that often one of the twins gets more food in the womb, because it is closer to the placenta (or vice versa). this could be a major cause of variation in adult intelligence in twin studies, but I've never heard it discussed.

2011/11/24 γενβιρΟ <carsth...@hotmail.com>

[eurekasabiduria]

The Polgar woman was already genetically very intelligent?

thank you

[Mike]

probably she was, but not necessarily. 2 parents can have an offspring that is more intelligent than them, by recombination of their genes. I don't have statistics about this, but I don't think it's such a rare thing. because shuffling of genes is the product of chance, you can get, if you're lucky, all the favorable versions of intelligence genes present in both parents, combined in 1 offspring. or the contrary.

that's why you still get a lot of gifted people coming from lower classes of society. if a very large number of genes were involved in intelligence, and none of these genes had a very powerful effect on intelligence, you would get offspring that would be closer to an average between both parents in intelligence, almost every time. that is not the case--there can be wide differences in intelligence between brothers and sisters--at least from what I know, it's not uncommon. this would tend to prove that there are at least a few locus that have a very powerful effect on intelligence. some people suggested that this is the cause for the constant instability or aristocratic classes, at least in europe/north america. but would this mean that india (and maybe china too) have different intelligence genes? (people, answer this one...)

but this is all speculation. on average a child's IQ is in between both parents, or not too far, and because of homogamy (people marry people similar to them, in terms of education/experience but also IQ), this creates a lot of averaging in the end. 

--> in western countries you are 5 times more likely to be gifted if you are from the upper class, compared with the lower class. surely this number varies a lot but it gives an idea. (because lower classes are greater in number there are still more gifted people coming from the lower classes).

I would like to get more information about this though. if anyone has any.

On Thu, Nov 24, 2011 at 12:24 PM, eurekasabiduria <eurekas...@gmail.com> wrote:

The Polgar woman was already genetically very intelligent?

thank you

[Gwern Branwen]

https://en.wikipedia.org/wiki/Regression_toward_the_mean

On Thu, Nov 24, 2011 at 1:50 PM, Mike <mike...@gmail.com> wrote:
> that's why you still get a lot of gifted people coming from lower classes of
> society. if a very large number of genes were involved in intelligence, and
> none of these genes had a very powerful effect on intelligence, you would
> get offspring that would be closer to an average between both parents in
> intelligence, almost every time. that is not the case--there can be wide
> differences in intelligence between brothers and sisters--at least from what
> I know, it's not uncommon. this would tend to prove that there are at least
> a few locus that have a very powerful effect on intelligence. some people
> suggested that this is the cause for the constant instability or
> aristocratic classes, at least in europe/north america. but would this mean
> that india (and maybe china too) have different intelligence genes? (people,
> answer this one...)

May I suggest reading Clark's _Farewell to Alms_? One of his major
results is that England had pretty high social mobility, rewarding
intelligence & initiative, and also that the rich enjoyed more
reproductive success. He discusses the obvious counter-examples of
China and Japan, pointing to studies finding that while the rich there
did have similar reproductive edges, the edges were far smaller (one
example - samurai households in Tokugawa Japan frequently had to adopt
to keep their line going). His recent PDF
http://tuvalu.santafe.edu/~bowles/RulingClass.pdf discusses all this
as well, with more info. (Not sure if I've linked it before.)

> but this is all speculation. on average a child's IQ is in between both
> parents, or not too far, and because of homogamy (people marry people
> similar to them, in terms of education/experience but also IQ), this creates
> a lot of averaging in the end.
> --> in western countries you are 5 times more likely to be gifted if you are
> from the upper class, compared with the lower class. surely this number
> varies a lot but it gives an idea. (because lower classes are greater in
> number there are still more gifted people coming from the lower classes).

--
gwern
http://www.gwern.net

[eurekasabiduria]

So Mr. Polgar did not smarter?