The virtue of excellence

Friday, November 12, 2010

Practice

Sebastian Marshall is on my side with feedback and practice. Read it. He (and I) disagree with the WSJ article here. FWIW, There is a major factor that none of the folks are paying attention to.

Returning to the Aretaevian theory of learning/expertise...

90% of learning is motivation (active/eager engagement)
9% is practice
1% is other stuff -- Talent, etc.

When everyone has similar motivation and practice, then talent takes over...and until everyone has reached the same levels of practice and motivation, talent is halfway to irrelevant...at least for folks in the first 4-sigma above average.

What is missing from the discussion is a quality-of-practice issue...Practice is part quantity and part quality. Shooting 1 million baskets does not make you a basketball star...but it is very likely to make you able to hit the basket when you shoot (without pressure, from at least most of the places on the court where you practiced). On the other hand, if you're throwing those baskets up underhanded, around the back, and goofing about, even 1M baskets shot won't necessarily make you a good shooter. What makes you a good shooter is the adjustment on shot 2 after shot 1 misses...and the attempt to recreate shot 2 with shot 3 if shot 2 goes in.

On the other hand...your chances of being a basketball expert if you haven't taken the couple hundred thousand shots required is near zero. Practice is necessary, even if not sufficient.

Similarly, Arithmetic expertise is not gained by just getting a math worksheet with 300 problems, and writing something down every day for 10 years (300+ 365*10 ~= 1M) Arithmetic expertise is gained by working on improvements in speed, and dealing with the feedback. If you write 7*8=54, and never get it checked, you're making yourself LESS good at math.

On the other hand, if you don't have addition/multiplication math facts in your head for instant recall, you can barely learn fractions because you have too many steps to think about. Proper fractions analysis is: What is 1/2 + 2/3 ... well, thats 3/6+4/6 = 7/6 = 1 1/6...if you pause to ask what's 2*3 or 3+4...you cannot do fractions...and you certainly can't use fractions for anything that takes cognitive effort.

On the gripping hand, if you do your 1M problems of arithmetic...that still doesn't make you good at Algebra (though you're MUCH MUCH better prepared).

Practice: Quantity, Topic, and Feedback are all huge. Insufficient attention to any of those and you won't get good at whatever you're trying to learn or become good at.

Hypothesis...IQ allows you to better self-manage your next steps...but almost never as well as a coach could.

10 comments:

Jehu said...

Aretae,
One SD of IQ is roughly a factor of two in learning speed. For instance, typical new material requires 7-8 repetitions for an average (IQ 100) child to learn. Children around IQ 115 take about half that, and children around IQ 130, 1/4 that. This is fairly uncontroversial educational research and it is congruent with my own experience. If you put this into linguistic terms rather than mathematical ones, this maps to difficult, average, and easy.

Of course, not all repetitions are created equal---some don't even 'count' as such if they don't actually engage the student's attention. My experience is that you can teach a person to do something that the average person in the field can do who is one SD smarter than they are if and only if they are willing to 'sweat blood' (e.g., teaching a person with a 100 IQ to do basic electrical engineering). You'll have a high failure rate---under normal circumstances, engineering has several courses that fail a little over half of the people that take it, but it can be done. It's not scaleable though because few people have the persistence and endurance to 'sweat blood'. This does point out one of the methods of beating people with more talent than you though...you can work harder...of course, if you're at the 'only the obsessed need apply' level, then you'll find it VERY hard indeed to 'work harder' than your competition.

drpat said...

almost never as well as a coach could

Almost never as well as a good coach could.

There are a LOT of coaches/teachers/instructors who don't actually help a given student to learn.

Either because:
1. The teacher is just going through the motions
2. The teacher is aiming at the wrong level, eg. running through multiplication tables with someone who is ready for fractions OR explaining fractions to someone who still needs to think about 2x3=?
3. The teacher doesn't actually know.


Looking back at when I've not been able to get something, it's usually been reason 2. Reason 2 that is caused by reason 1. Or reason 2 that is caused by reason 1 that is caused by reason 3.

drpat said...

forgot to activate the email notification

Jehu said...

DrPat,
Other reason:
The teacher isn't within the IQ envelope where effective communication takes place of the student. Research indicates that past a couple standard deviations that very little gets done. This is one of the hidden advantages of homeschoolers---because the spread between parent and child is generally far smaller than that of child and unrelated teacher, communication just works better. It's possible to extend your reach a little bit through heavy reliance on religious and/or sporting context, but it's a hard problem.

Freedom said...

The thing about natural talent is that it kicks in early in the equation. That 1% is huge at creating the motivation at low levels of practice. Also, if you're looking at achieving minimum competence, rather than mastery, natural talent can count for a lot.

drpat said...

Is natural talent important at creating early motivation? Maybe, but I'd point to other factors that can work just as well and may well dominate:

1. Experience in something transferable. Two people start skiing for the first time. Both have similar natural talent, but one has trained as a gymnast and in martial arts. He already has the leg strength, balance and body awareness that most skiers take months to develop. He also has the impact tolerance, falling skills and pain tolerance that means he is willing to take much greater risks than normal. Naturally he ends up doing black diamond runs on his 3rd day of ever seeing snow. (True story) Result: Lots of motivation.
2. An external motivation that creates lots of enthusiasm, for long enough for the student to start succeeding. The initial success then leads to internal motivation.

In both cases, the natural talent may well be average, but the initial jump means they end up better than most people ever do.

Getting back to my earlier comment about poor instruction. I left out a major cause, which is having other students who are at a different level in the same class. That almost requires inappropriate instruction for some.

drpat said...

I've noticed I'm discussing physical skills, not mental ones. Which could be why IQ is featuring less in my analysis.

I think it's because I have more experience with both teaching and learning physical skills. Mental skills I'm restricted to the learning and applying modes.

Aretae said...

Jehu,

1. I've posted research that matches your 2x/stdv number. Further, I've taught folks from the 4-sigma level down to the negative-one sigma level. The numbers I saw said between 1.6x and 1.8x, but that's degree of precision, not disagreement. My experience doesn't disagree. And I took that into account.

2. My experience says taht you can teach someone to do darn near anything, regardless of the average person in the field's qualifications, if you're willing to put the time in. Sweating blood isn't necessary. Doubling the time to learn will substitute just fine.

Aretae said...

Dr. Pat.

I spend my days now as a corporate trainer teaching programming. Level is ALWAYS the problem from the teacher's point of view.

When I'm teaching Java, and I've got a guy with 5 years of C#, I know precisely what to teach. When I've got a guy with 6 months of COBOL, I know what to teach. When I've got a guy with 20 years of C, I know what to teach....but when all 3 are in the same class, I'm screwed. There's simply not a darn thing I can do and win.

IQ envelopes can be beaten by teacher experience. I have personally spent a lot of time (>2K hours) teaching below average IQs...for at least a 5-sigma difference. First time I tried, I was completely befuddled. Later, I got better.

Aretae said...

Freedom, DrPat,

Talent does kick in early, but 90% of the time that I've seen talent kick in early, it's actually Dr. Pat's situation, not Freedom's.

Kids who are artistic in kindergarten have often been drawing for 3 years, and so artsy folks are not necessarily more talented, but do have 3 years more experience. And 3 years and 3 weeks experience is about 150x as good as 3 weeks experience.

I'm sure it's not ALL experience. But damn...the amount that is transferable skill is awfully high.