It seems like I’m always pushing my kids slightly outside their comfort zone, and trying to get them to do something new. So recently, I figured it might be good for me to model some of this myself by taking up a new sport.
The experience has given me a lot more empathy for what my kids must go through. Because I had completely forgotten what it feels like to be a beginner, with the constant confuzzlement and overwhelm that comes with learning something totally foreign.
And a lot of it has reminded me of what it felt like to study music too.
There’s a lot of technique involved, and a ton of details to keep in mind – from the fundamental basics to the subtle and nuanced. Sure, you can slow some things down and work on them in isolation, but ultimately, you have to be able to put everything together into a fluid sequence.
Unsurprisingly, all my old learning habits from music kicked in too. Like my tendency to want to do things perfectly, at the highest level, right from the beginning. After all, “practice doesn’t make perfect; perfect practice makes perfect,” right?
But then I see some of the other newbies going through the same skills with somewhat less attention to detail, clearly not as worried about messing up little things here and there. And not only do their movements look more fluid, but they seem to retain more of what we’re learning from one week to the next too.
So at the end of the day, which is the most efficient way to learn? Is my “perfect practice” approach the right way? Or is it better to take the “imperfect” approach in the early stages of learning?
Learning how to throw
A team of researchers at the The University of Hong Kong recruited 216 third to fifth grade students1 to participate in a 5-week training program.
Everyone started out with a test of throwing accuracy and form, where they were asked to throw 10 beanbags as accurately as possible at a life-size picture of their PE teacher holding a softball inside a baseball mitt. They stood 5 meters away, and were to aim for the softball inside the catcher’s mitt.
Over the course of the next few weeks, the students received a weekly 15-minute practice session, where they practiced throwing beanbags at a square target taped on the wall, again from 5 meters away.
One group of students (the error-reduced or “imperfect practice” group) threw towards a large, easy target area in week one (2.4m x 2.4m), a medium-sized target in week two (1.1m x 1.1m), and a small target in week three (.45m x .45m).
Another group of students (the error-strewn or “perfect practice” group) went in the reverse order, throwing at the small difficult target in week one, the medium target in week two, and the large target in week three.
And then a week later, they were tested on their throwing accuracy and form once again, just like they were at the beginning of the study.
Two measures of improvement
The students’ improvement from week 1 to week 5 were measured in two ways.
To measure accuracy, the researchers calculated the distance from where the beanbag hit the wall, to the center of the target.
And to evaluate throwing form and technique, one of the researchers (a physiotherapist), and three trained evaluators viewed videotape of each student’s throws to establish a movement form score from 0-8, based on the movement of their a) arm/hand, b) hips/shoulders, c) feet, and d) trunk.
So which sequence of practice led to better performance on the test?
Easy to challenging vs. challenging to easy
In terms of accuracy, the error-reduced group (i.e. where they progressed from easy to more challenging standards over the course of training) made significant strides in accuracy from the beginning of training to the end. On the contrary, there wasn’t a significant improvement in performance for the error-strewn (i.e. perfect practice) group.
This was reflected in their throwing form too, as children in the error-reduced practice group showed greater improvements in throwing form than those in the error-strewn group.
There are several reasons why starting off with more forgiving standards and progressively raising the bar may have resulted in greater learning and performance.
One possibility is that by making the task easier to succeed at in the early stages of learning, it freed the students up to just throw. And avoid that tendency to get overly analytical or concerned about “perfect” execution, which can lock us up and constrain the explorative process of figuring out the best way to perform a skill.
It’s possible that confidence may have played a role too. After all, there’s nothing that squashes confidence like being totally inept at something on Day 1. So having more early successes may have kept students engaged in learning and trying to improve on each week’s performance. Which seems reasonable, given that the students who had the worst pre-test throwing scores seemed to benefit most from the error-reduced training.
At some point, it’s essential to be a nerd, delve into the details, and develop a deep understanding of the underlying mechanics of a skill. So in the long run, maybe things balance out a bit.
But as I reflect back on most learning experiences I’ve had, my teachers have generally tried to get me to master the gist of a new skill first, even if the results weren’t perfect, and then over time, address the finer details and nuances of the skill.
I was rather resistant to this approach, thinking it was being too permissive of sloppiness or imperfection. But studies like this one suggest that perhaps they were right. That instead of trying to achieve an expert level of perfection right from the beginning, perhaps it’s more effective to lower the bar, get the basics down, and then refine, refine, refine to a progressively higher standard once the basic foundations of a skill are locked in.
In my opinion, the research doesn’t really help the article, because of the 2nd group (error-strewn)… It is better to keep the square small, not giving the bigger and bigger one. Just like for musicians who practice for perfection from day 1: it doesn’t mean they are practicing less perfectly later, are they?
I know the difficult-to-easy progression for the error-strewn group seems a little odd, and isn’t very realistic, but I think the results may have been similar even if they had kept the small target the same across training for that group.
This approach to learning is called “errorless learning,” and despite what the term “errorless” implies, doesn’t mean that individuals make zero errors during learning. It just means the task/environment is adapted so that they make fewer errors during learning (like making the target bigger, or shortening the distance of the throw, etc.).
The idea being, if individuals achieve more success early on, they don’t have to do as much conscious, explicit learning to correct their performance. Which not only seems to accelerate the initial learning curve, but results in a skill that’s less dependent on conscious control and more capable of running on autopilot (evidenced by more stable performance when having to count backwards from 100 by 2’s, for example). There are problems with implicit learning, of course, but I could see how it might not be such a bad thing in the early going.
Which also starts getting into research on working memory, and how we get a little paralyzed if we try to give ourselves too many technical instructions all at the same time (like not only monitoring our sound, but trying to consciously control the thumb, finger pressure, elbow, shoulder on our left side, and bow speed, pressure, direction, etc. too).
Hope that helps to clarify!
Hi Dr. Kageyama,
I’m somewhat fascinated to hear that there are problems with implicit learning. Not that I find it hard to believe, but that I hadn’t considered what the disadvantages might be. Do you have readings to suggest (or want to explain the matter yourself)?
I think this might be interesting to write about at some point, but my experience with implicit vs. explicit learning was that implicit was great up until a certain point. But then, it was hugely beneficial to understand the nuances and details and mechanics of a skill, so that I could not only troubleshoot more effectively but execute more consistently too.
Vanessa Hoang points out one important flaw in applying the results of this experiment to music. Here’s another:
Performing music is a series of continues tasks that unfold over time. Conversely, this experiment measured a series of discrete tasks separated by time. Thus, the results of the experiment don’t apply to performing music.
This is because the cognitive process of performing music is highly impacted by cognitive dissonance. In my 25+ years of teaching basic musicianship skills (i.e. often called “aural skills” or “sight-singing”), I have learned that cognitive dissonance is one of the key factors that inhibits progress and success in developing musicianship skills.
Cognitive dissonance arises when a student makes a mistake yet keeps going. Part of the brain gets distracted/stuck on analyzing that mistake, while the student is continuing. This usually causes another mistake a few notes later; and the cognitive dissonance piles up until it knocks the train off the track and forces the student to stop. All too often, the student works on fixing the last mistake, when in fact they should work on fixing the first mistake (which is what actually caused the problem).
The goal of practicing is to learn to *avoid* cognitive dissonance as much as possible, i.e. by doing everything as close to perfect as possible.
In simple terms: There are two goals inherent in practicing: 1) developing accuracy and 2) developing fluency. Fluency will naturally develop (almost on its own) as one practices for accuracy. However, the converse is not true: If one is practicing for fluency too early in the practicing process, then accuracy will never develop (because the student is too willing to accept little inaccuracies).
In my experience, this is the biggest error that students make in their practice routines, i.e. they practice for fluency much too early (and too often); this allows cognitive dissonance to creep in, which then thwarts their progress (and causes frustration).
Of course, there will always be errors that occur in any perfomence. Thus, learning to *ignore* cognitive dissonance (e.g. learning how to play through mistakes without getting rattled) is also an important skill to develop. However, that’s a very different skill. In other words, putting one’s mind into “practice mode” is different than putting one’s mind into “performance mode.” And there are ways to practice making that transition from one brain mode to the other. But in my experience, practicing with the brain in performance mode (i.e. learning to ignore cognitive dissonance) should come much later in the process than practicing with the brain in practice mode. Fluency will develop from practing for accuracy; but accuracy won’t develop if all the student does is practice for fluency.
Your attitude toward errors may be the biggest factor creating cognitive dissonance. If one teacher reinforces that mistakes are fine sometimes, and another teacher reinforces that all mistakes are to be avoided, their students will react differently to mistakes. You may not be saying that all mistakes are to be avoided, but students pick up on subtle cues.
This is one reason that trained classical musicians often have difficulty transitioning to jazz improvisation. They are so highly conditioned to there being one right answer that they lock up when confronted with infinite possibilities.
Hi Jon: I agree with you. I’ve worked in both jazz and classical music, and with amateurs, college students, professionals, etc. Each of those situations has a unique set of factors, such as expectations, notions of “correctness,” peer pressures, etc., which influence the types of pedagogical approaches that are effective.
That said… I’ve noticed that asking a student to take a step back and practice towards accuracy is often a key to unlocking their stumbling blocks (and frustration) towards improvement, even with students who are working in jazz, or students who are simply amateurs but want to improve their proficiency.
I worked in conservatory settings for 20+ years; but I now work mostly with amateur musicians (in a non-auditioned community chorus, giving piano lessons, etc.). Many of them ask “how to become better”; so the desire for improvement is similar, but the social circumstances are different.
Musicians (of all sorts) are bombarded with social cues that reinforce the musicianship is a behavioral skill, and that the “only” way to improve/succeed is to emulate that approach. They put undue pressure on themselves by emulating the musicians who make it appear so easy/fluent; they expect to develop similar fluency before they’re proficient enough to achieve it, and they dis themselves whenevr they can’t do things as fluently as their role models.
I’ve found that asking them to take a step back and focus on accuracy (i.e. to approach their practicing more cognitively rather than behaviorally, as well as making them aware of how cognitive dissonance is getting in their own way) can be a surprising revelation for them. It gives us mere mortals (i.e. the musicians who weren’t lucky enough to be born with immense innate abilty) “permission” to work to improve; and it gives us concrete steps to foster improvement.
So, I agree with you that the approach of “striving for perfection” runs counter to teaching things like jazz improvisation (i.e. in which teaching oneself to not get hung up on “mistakes” is a critical skill). However, I’ve found that even a jazzer or an amateur choral singer can improve their core musicianship skills by first focusing on accuracy before expecting themselves to rely on fluency.
“The goal of practicing is to learn to *avoid* cognitive dissonance as much as possible, i.e. by doing everything as close to perfect as possible.”
I think this is the point I have trouble with. Doing everything perfectly does help you “avoid” cognitive dissonance, but avoiding it prevents you from learning to “cope” with it when it inevitably occurs. If you do not practice accepting and recovering from mistakes, you will not be good at it. Atheletes practice doing their moves perfectly, but they also practice them in game situations, where opponents try to disrupt that perfection. I think it is the combination of both striving for perfection and just letting it rip that creates fluency. This is how you develop “flow”, where you are completely in the moment, and not worrying about the past or future.
Fo r me there is no question at all that development always should be from easy to challenging and NEVER the other way around. But there is a risk involved by this, i.e that errors will be learned. And herein lies the teachers och coach biggest responsibility, to detect and correct errors as early as possible. And if an error is detected it must be addressed immediately. The errors should be thoroughly analyzed and the students focus and attention should be brought to the correct way to perform through the challenging point or passage. If this is over-challenging the student the exercise must be simplified in speed, length, detail or structure. But never to the extent where it introduces boredom.
This begs the question of what an error is. The goal is to hit the center target, so anything else is an error. You can make the square larger around the target and declare that missing the target is no longer an error, but you have not reduced errors; you have simply redefined them. You are accepting less accuracy, yet it is more effective in developing accuracy.
The analog to your approach would be to have the students stand closer to the target, so they always hit it accurately, and then to gradually move them away. It would be interesting to see research comparing “large target to small target” versus “close to far”.
Noa has also discussed research that suggests that mistakes should not be addressed as early as possible. If a student is working on improving a skill, it is better to let him make many attempts and let him try to adjust on his own, and then to give summary feedback at the end of the session, than to give feedback after each attempt. So, correcting “as early as possible” is not really optimal.
One thing to consider when evaluating “standard practice” in music instruction, is that it generally fails. Most kids quit their instruments. It is like diet or exercise. Yes, there are a very small number of people for whom standard practice works, but it usually does not.
I find this discussion fascinating!
Jon, by “standard practice” do you mean giving feedback after each attempt?
In my own teaching experience, I have seen how correcting embouchure, rhythm, pitch, etc., too early/often in lessons gets some students so frustrated that they eventually quit, while other students like having something to aim for and the idea of “getting it right” right away motivates them. I think it’s important to remember that we want our students to enjoy playing music, not simply master the art of performing. (A colleague of mine made a wonderful comment in a pre-concert talk, in which she reminded the audience – and us fellow performers – that we are first and foremost “players” of our instruments and not merely “operators.”) The balance of practicing for fluency and practicing for accuracy will be different depending on the student’s experience and goals, but finding the balance of the two that keeps a student motivated and wanting to improve is universally important. It’s this striving for and achieving new heights which in the long term lead to satisfaction and enjoyment.
This seems to be at odds with the study of conservatory students (cited in this blog) who illustrated that slow practice without errors from the get-go were able to master a piece far more quickly than those that utilized the gradual improvement concept alluded to in this article. How can these differing results be reconciled?
Great question; I think you’re alluding to this study of piano students, correct?
I thought about this as well when reading through the study, and puzzled over it for a while. It does seem like the findings are compatible on a basic level, though it’s always tricky to interpret of course, because the two studies have many differences.
In the piano study, the pianists (who were asked to learn a small fragment from a Shostakovich piece) whose performances of the passage the following day were ranked highest, did prioritize the minimization of errors. But they did this by slowing things down to make the task easier – much like the group of kids in the beanbag study who started out by aiming at an easier, larger target area. And as the pianists’ ability to play the phrase more accurately increased, I believe they played closer and closer to the final tempo. Which is like the beanbag kids who threw at a smaller and smaller target from one week to the next.
So ultimately, there seems to be a tricky balance between cultivating excellence, but making sure the learner isn’t overwhelmed by too many details and constant “failure” either. Maybe we can achieve this balance by making sure learners don’t pick up bad habits in the foundational areas that are absolutely essential to playing one’s instrument whether one is a beginner or world-class professional, but waiting to introduce some of the higher-level details until the student is ready to take things to the next level?
I’m guessing age probably accounts for some of the difference.