Deliberate Practice: How Education Fails to Produce Expertise

Photo: iStockphoto

Thanks to recent, hugely popular books about the development of expertise, the term deliberate practice is coming into common usage as the kind of practice that produces expertise.

Deliberate practice requires careful reflection on what worked and what didn’t work. A budding concert pianist may practice a particularly troublesome passage listening for places where his fingers do not flow smoothly. A chess student may spend hours analyzing one move of a world-championship chess match trying to see what the grandmasters saw. This kind of practice demands time for reflection and intense concentration, so intense that it is difficult to sustain for longer than 3 hours per day.

As I have learned more about deliberate practice, I often think about its lessons for the educational system. And they are not happy ones.

In the grade-school years, deliberate practice is already hard to find. My strongest memory from fifth-grade mathematics is pages and pages of tedious three-digit-by-three-digit multiplication problems. Day after day! It is, alas, the kind of rote practice that I have done for chess: simply playing lots of games.

In a classic paper, “The Role of Deliberate Practice in Chess Expertise,” Neil Charness and colleagues studied the effect on chess rating of different types of chess practice, including total hours of serious study (i.e., deliberate practice) and total hours of tournament play (their Table 3). The effect of deliberate practice far outweighed the effect of tournament play. Not surprisingly, my chess skill today is only slightly higher than it was nearly 40 years ago when I learned to play chess.

In college, at least in science subjects, the situation is hardly better. As an undergraduate physics student, one often takes four or five science or mathematics courses per semester. Each course assigns a set of problems every week, each set taking about 6 to 8 hours. Thus, solving lots of physics problems is how one spends most weekday evenings, many sessions lasting late into the night; in my time the physics majors had keys to the physics library, where we would order pizza to fuel the weary problem-solving neurons, and I remember seeing many sunrises from the windows of the library.

Those problem-set hours total almost a whole other working week laid on top of the other academic tasks of attending lectures and reading notes. In college as in grade school, where is the time for deliberate practice?

[The notes were mostly a copy, with my own mistakes added, of what the professor wrote on the board, which was usually a copy of the textbook (with perhaps fewer mistakes added). The Gutenberg method of teaching, which takes account of the invention of the printing press, is little used.]

Is it any wonder that, just as I can play chess but have little insight into how the game works, science and engineering students graduate able to “work hard” but cannot solve problems expertly and creatively? What would an educational system look like that took seriously the principles of deliberate practice?

Leave A Comment

Comments are moderated and generally will be posted if they are on-topic and not abusive.



View All Comments »
  1. Kristjan says:

    Some “deliberate practice”: I am studying maths at the University of Manchester (UK) and in most courses we have a “mid semester take home test” which normally consists of a single really hard problem which some times takes days to understand and solve. And there is also a dissertation.

    But I agree, we need more of it, in every level and in most of the areas.

    Thumb up 4 Thumb down 0
  2. Megan says:

    My son’s school uses Rocket Math in grades 2 and up–starting with addition learn 2 new math facts (plus the reverse) at each level. To go up a level you solve 40 problems in a minute and when you get “stuck” you go back a few levels to review what you haven’t mastered.

    I think there is a time and place for it. Math facts and spelling would be and appropriate place for deliberate practice. But it takes a lot of time, time that could be spent on critical thinking, reasoning and creating. Not to mention it would be difficult to do when you have a 1:25 teacher student ratio.

    Thumb up 4 Thumb down 5
    • Crissa says:

      Speed of recall does not equal mastery of the subject. Masters recall quickly – but so do crammers.

      Besides, you didn’t even bother to identify what is ‘practice’ vs ‘not practice’ in your example.

      Thumb up 1 Thumb down 1
  3. Al says:

    When thinking about deliberate practice,be careful to think about learned skills (arithmetic), conceptual knowledge (physics concepts), and applying conceptual knowledge (doing physics problems,using mathmatict skill) as different areas.

    Often times this distinction is lost in public schools. Especially when it comes to social promotion. Many current problems stem from physically mature young adults who have the brain power to grasp more advanced concepts,but lack the skills (reading, writing,and arithmetic) to display that knowledge on a standardized test.

    Well-loved. Like or Dislike: Thumb up 8 Thumb down 1
    • Scott says:

      It’s important to note that as teacher being in the education sector for years, teachers do NOT like social promotion.

      Social promotion is something to results from allowing parents too much say in the educational system.

      Education is one of the only areas where the patients often tell the professionals how to work. Imagine patients telling a hospital staff how & when to do procedures… Insanity, but that’s what happens in education, then we blame the frontline providers who are working with tied hands.

      Thumb up 4 Thumb down 4
      • Josh says:

        I hope you don’t teach English.

        Seriously though, this happens in more than just teaching. It happens whenever the process isn’t obfuscated by loads of technical jargon. Graphics Designers have this problem all the time. Anyone that’s done a bake sale flyer thinks they can design a website.

        Thumb up 4 Thumb down 3
      • James says:

        OTOH, there are apparently a large number of graphic designers who can’t design usable web sites. One major news network’s site causes my browser to freeze whenever I try to view one of their articles, a major bank/mortgage processing company produces a site that only works with one particular browser, I’ve lost track of the number of sites that try to force me to install Flash to view their $%#! animations… It’s creativity in action.

        Thumb up 4 Thumb down 3
      • Dan says:

        I am not a web designer but I know that the clients very rarely insist on how the front end is going to be built, but very often insist on what it’s going to look like. Go on craigslist sometime and search for “graphic designer”. You will very quickly find dozens of requests for designers to “be the hands” for the “creative genius” of some doofus.

        The crappy websites that you’re describing are very likely the results of companies cutting corners and paying inexperienced graphic designers to do web designer/developer work.

        Additionally you should know that many site designs will take into account the percentages of users who use a certain type/types of browser. If you are in their 2% category then they don’t care about you… and yes sometimes IE is in that 2% category haha. Even if you get the browser and version of browser right, the user may have the wrong display options selected, etc. It’s a lot harder than you think to design a site that will work on every browser version of just the last 5 years… and even harder still to get it to work with every relevant user configuration.

        If you need a web designer: find a good one, pay them well, and shut your hole :)

        Thumb up 3 Thumb down 2
  4. Eugene Evon says:

    Practice makes perfect.
    This reminds me of a study I vaguely remember showing that length of time at task had a stronger correlation to proficiency than any targeted teaching or training. This is especially true for skills we all recognize. Cramming, no matter how well-designed the learning content, can’t make me a better guitar player, golfer, presenter or speech writer. Why do we not recognize it for science and engineering students?
    Kind of a big “a-ha” for someone like me, someone with an engineering background whose career now focuses on improving performance via technical learning. Perhaps it is more important to find ways that entice learners to practice rather than promise results via shortcuts.

    Thumb up 2 Thumb down 0
    • Pat says:

      @Eugene: Either you are or I am misinterpreting the crux of the article. My understanding is that it isn’t just how many times you perform a task, it’s how much time you spent contemplating the possibilities for solving a difficult task. You can play chess every day, but you won’t get significantly better unless you take time to understand why a chess master made a particular move.

      Thumb up 4 Thumb down 0
      • Anders Mikkelsen says:

        I think Pat is right. Deliberate practice means understanding / intelligent practice, not just rote practice devoid of meaning or connections (not that there isn’t a place for rote, it’s just very limited.)

        Thumb up 2 Thumb down 0
  5. DS says:

    A good example from my experience: More than two decades ago, I was preparing for the GRE (Graduate Record Exam, an entrance requirement for graduate schools in the US; a harder version of SAT, really). One of the three components was “Verbal Ability”, of which vocabulary was an important component. There were two popular books for preparing for the vocabulary part (besides simply reading a lot). One was the “Barron’s guide”, which simply gave a list of a thousand or so words (with their meanings, and sample usage). The other one was a vocabulary book by Rosenblum and Nurnberg, which took the time to demystify word structure (Latin and Greek roots, prefixes and suffixes, etc.), built maps of words related by their roots, etc., as well as great slices of the language (e.g., popular words from sports writing, vocabulary unique to newspaper headlines, etc.).

    Needless to say, the latter approach was much more deliberate, helping to build more associations, structure, context around words; from time to time, I still encounter/use words I first learned through that book, and can’t help recalling how pleasant that learning experience was.

    Well-loved. Like or Dislike: Thumb up 14 Thumb down 0
    • James says:

      But here’s a counter-example. I have never (at least since elementary school) done any kind of deliberate practice in vocabulary. All that I’ve ever done was to read widely, and perhaps look up the occasional unfamiliar word whose meaning wasn’t clear from the context, yet I managed to get quite high scores on that GRE exam.

      On the other hand, it seems to me that most of the problems given in engineering & the sciences are deliberate practice, and most E&S students graduate able to solve problems with a fair degree of expertise, and a perhaps inappropriate amount of creativity. Something which in my opinion is quite overrated, as it often seems to equate to abandoning sound engineering and producing an ungodly – but creative! – mess.

      Thumb up 0 Thumb down 0
  6. Richard says:

    This is a problem in computer programming as well; certifications that produce people able to respond with a definition when confronted by a technical term, but who have no ability to provide working software.

    A common learning method in programming is what are often called “code katas”; series of problems with a simple but non-obvious solution, gradually more difficult, that build on previous knowledge.

    Wikipedia provides examples of such problems;

    Thumb up 1 Thumb down 0
  7. Dan Grayson says:

    I think it might look like this: the class has 15 students or so, all with the same ability and background. Students work in groups of 3 or 4 during class periods; all groups work on the same assigned problem. The problem’s difficulty is such that no one knows how to solve the problem at first. The time required to solve problems increases during the semester from 20 minutes to 5 hours, each. Time outside of class is spent writing careful individual prose expositions of successful solutions, after those solutions have been thoroughly discussed by the whole class; the discussion and exposition will abstract any useful problem solving strategies employed, in case they are useful later. The professor’s role is to assign problems, to assign students to groups, and to refrain from giving hints or objections, at least until the students have been substantially frustrated; there is no lecturing. The grade will be based solely on the quality of the written expositions.

    The first question in a fifth grade math class could be “Why is 57 times 42 equal to 42 times 57?”. The followup question would be: “What about other pairs of numbers?”

    Thumb up 3 Thumb down 0
    • Scott W says:

      If you want to have students spend more time on ‘deliberate practice’ than you need to sell it to parents first. Too many families spend the time not spent in school shuttling kids between the soccer field, dance classes, piano lessons, scouts/guides, art classes, vacations DURING the school year, etc.

      When the true EDUCATION isn’t a priority, but test scores are…. little authentic real learning results.

      Thumb up 0 Thumb down 0
  8. AaronS says:

    Some years ago, upon reading of Deliberate Practice for the first time (on Freakonomics, no less!), I was absolutely fascinated! This meant that a “regular” person could, by using the right practice method, become superior in some field. So I delved into it, going so far as to correspond with one of the researchers at Florida State on the matter. Here are some thoughts on both it and the Gutenberg Method….

    First, SOME things need to be learned by rote. Why should children have to spend part of two or three years mastering the multiplication tables, when they could instead be putting that mastery to work solving problems? Or how about how much time is spent trying to memorize the Periodic Table? Or certain historical facts? Very simply, there are things that, if we knew them, we could use them, build on them, expand on them, to reach new heights, applying creativity and scholarship. So there is a place for plain ol’ rote learning.

    There seems to be two central keys to Deliberate Practice:

    First, you need a LOT of practice (the 10,000 Hour Rule is mentioned several times in articles regarding this sort of study). You can’t become and expert overnight. So this is an endeavor that takes plenty of time.

    Second, you need to KNOW what you did wrong IMMEDIATELY, so that you can ADJUST IMMEDIATELY. That’s why getting your graded test back two days later doesn’t do much for you. You need to know IMMEDIATELY that you got an answer wrong. If you do, it apparently helps you better hardwire the correct way/answer into your brain.

    Thus, we could create TESTS that teach perhaps far more effectively than the standard teach-THEN-test method. That is, if all tests were computerized, say, then when a person answered a question incorrectly, immediate feedback would be given. The student would know immediately that their answer was incorrect.

    Then comes the cool part….

    That question would have to be answered correctly to move on…BUT…that question would also be shuffled back into the mix of remaining questions, making sure that the student had to get it right again before…giving us some assurance that the student now knows the answer for sure.

    Now, let’s consider what would happen if we just wanted to teach the multiplication tables. A first-grader is given a problem from the multiplication table. If the answer is incorrect, the student is made to know immediately. The right answer is either given or must be input before the student can continue. Later on, the student must answer that same question again (maybe even several times). Eventually, the student will know the multiplication tables.

    THEN the student can spend time USING the multiplication table, rather than learning it.

    You can even do it with flash cards. Pick up a card from PILE #1. If the student gets the answer right, it goes to PILE #2. If not, then it goes back into PILE #1 (and the correct answer is shown). Cards in PILE #2 must also be answered correctly AGAIN. If so, they go to #3; if not, back to #1. When cards reach #5 (say), they do not have to be answered again.

    The key is IMMEDIATE feedback and adjustment, in this case.

    It would not be difficult for Deliberate Practice software (or a game) to be developed in which ANY information is arranged in several very similar questions. For example:

    “President Lincoln was assassinated in 1965–True or False.”
    “In what year was President Lincoln assassinated? ___________”
    “Which year is closest to the year Lincoln was killed? 1776, 1838, 1863, 1875, 1900, 1922?”

    If an incorrect answer is given, not only must the original question be answered correctly at some later point, but so, too, a similar question, ensuring it’s not just memorization of the question form,etc.

    If we could identify the top, say, 1000 American History “facts” that every student should know, we could teach them using this method, allowing them to build on these facts, seek UNDERSTANDING of these facts (and not just possess the facts), and, in a nutshell, save time in the learning process, giving the student the time to delve into the information more than spend time learning it (since I am certain that deliberate practice teaches it faster and better).

    As for the Gutenberg Method, I have wondered why professors didn’t just Xerox their lecture notes! I was fortunate to have a field of study (philosophy) and teacher that invited student questions, so my note taking was well supplemented with additional info. But Prof. Joseph Boyle (back in Tampa, Florida) used the Gutenberg method, and woe be unto ye if you hadn’t read the material. He’d start asking questions from the first moment, picking you apart if you didn’t really have a clue, forcing you to know the information better and better.

    Well-loved. Like or Dislike: Thumb up 10 Thumb down 1
    • Steve says:

      I actually developed a website that is very similar to what you are looking for:

      Using my site you create a bunch of flashcards. When you are asked the flashcard, you attempt to remember the answer and give it a grade between 0 and 5 depending on how well you remember it. 0 if you forgot/don’t know it, 1 if you got it wrong but it sounds familiar, and 2-5 based on how easy it was for you to remember it.

      Then based on your rating, the software schedules the cards for you in an optimal rotation. You will keep on reviewing cards you get wrong and are hard, but for cards that are easy, it may be months before you are scheduled to review it again.

      Check it out at . Let me know what you think!

      Thumb up 2 Thumb down 0
    • Anders Mikkelsen says:

      This is very interesting. It seems that part of deliberate practice is to make sure that learners aren’t working on problems they can easily solve. Giving students tons of problems or lots of flash cards isn’t helpful. If a student can nail a type of problem every time, there’s no need to keep giving them the problem. It’s a bit like science – good scientists don’t keep repeating the same experiments once they understand an area, they move on. What needs to happen in deliberate learning is the learner needs to make sure they understand the problem – e.g. in a chess game why did the grand master do what he did? The learner has to think and go deeper, not just go through the motions with no understanding. (This ties in to all sorts philosophical issues – like ‘enlightenment,’ law being the expression of a loving god, art and invention as expression of a personality, authenticity, etc.) It seems like deliberate practice requires focusing until you figure out if you really understand something, and then moving on to the next thing you don’t understand. (I think Tiger Woods scrapping his golf swing and learning a new better swing is cited as an example.) It’s a progression and requires going deeper.

      Thumb up 0 Thumb down 0
    • James Briggs says:

      I think that people learn in different ways. My experience with learning to multiply is just the opposite. Memorizing the times tables was confusing for me because I didn’t know what I was doing. I found that the best way for people to learn to multiply was for them to learn to count with every number from one to twelve. Then learning time tables is a snap. That method is often used succesfully outside the US. What also helps is manipulatives like the Cuisenaire rods. I have had great success using them but they must be used in the spirit they are intended. The problem we have in this country is we have production line learning because we think learning is no different then building a machine. As a result the use of manipulatives in the US often leads to poor results.

      Thumb up 1 Thumb down 0