Does learning by doing actually work?

A lot of people talk about learning by doing. How does learning by doing actually improve learning outcomes? Wait... does it improve learning outcomes?

00:00 Introduction
00:29 The practice case
1:19 The indirect learning case
1:36 Worked Examples vs. Discovery
2:30 Tell & Practice vs. Invent & Tell
3:35 Volcanoes and Hearts
5:00 What the heck is my point?

For videos on deliberate practice, see:

For more on interleaved practice, see:

For a discussion on spaced practice and interleaved practice, see:

Footage sources:

Integration-by-parts worksheet from:

Check out their content if you like what you see.

References:

For a study comparing "Tell & Practice" and "Invent & Tell" see (This is the one with the clowns, etc.):

Practicing Versus Inventing With Contrasting Cases: The

For one of the classic pieces comparing "worked examples" to "pure discovery," see:

[KathySierraVideo]

For motor skills, learning by doing is the obvious choice. I think the issue is whether we learn best through direct instruction vs. exploration and self-organization.
The argument is typically, “we know what happens when the learning is left up to the learner… and it’s not good!”
This is where the ideas of non-linear pedagogy are (I think) a big part of the future of learning.
It support self-organization and discovery, but using *constraints* to narrow the search space strategically.
And “Differential Learning” is another non-linear pedagogy approach, but not as strategic as Constraints-Led Approach.
Both of these are primarily motor learning, but they are just as valid for cognitive skills. We don’t yet have examples of people applying it in non-movement domains, but I’m hopeful 😁

[jamiececilielange]

When I'm just listening to a teacher speaking, I struggle to stay present and focus, but when I interact with the knowledge in the form of exercises, I stay present and understand it a lot better.

I play piano and sometimes I don't learn much by learning a song, but more by practicing a specific skill like timing, notereading and playing by ear.

[wilaustu]

It sounds like the principle is that you want a strong bond between experience and mental representation. In the learning process, sometimes it makes sense to start with the experience, and other times it makes sense to start with the mental representation.

[naturalglory]

Currently self studying math. I have tried to apply the discovery method of learning. When learning to derive a formula ill get a general guide(a previous formula or a definition or even the first step) then Ill try to work through the steps myself. Then when I get stuck Ill peek at the steps. Then go back to trying it myself, then when I get stuck again ill take another peek etc etc until I get to the end.
The problem I've wondered is, if this is even worth the effort? This video was great, just wish you had given more specifics about when to use one vs the other.
I'm also curious about how this relates to solving online programming puzzles. To give background, coding puzzles are really popular ways to exercise your programming knowledge (algorithms, data structures etc.) Some of these problems are pretty hard. Some people advise trying to solve the problem and if you cant after an hour look at the solution. This makes sense bc by looking at the solution youll be able to learn something to fill a gap in your knowledge.
The thing is if I hadn't looked at the solution and just kept trying to solve the problem(even if it took several days) - would it lead to some sort of deeper understanding? I can imagine a situation where one leaves the problem unsolved but thinks about it regularly until they are able to connect the dots between a new/old concept. Wouldn't this create a better opportunity for learning because they push themselves to apply their knowledge in this edge case? Obviously this method can be thrown out the window for school work but when the goal is deep understanding...

Another tidbit - could there be a sweet spot in your problem solving timing that can lead to more benefits(motivation to keep solving problems, better retention, or even solving the problem itself)? For example, stop working on a problem when frustrated or keep working until your sick of it? Or just dip your toes into the frustration and then stop. Personally I've experienced being so fed up with the problem that even when I solved it, I got very little pleasure from it.

This has been on my mind recently and I'm so grateful to have found a channel like yours that not only makes entertaining videos but extremely informative ones!

[Jakob]

Building my 3d printers myself from kits is always a cool experience, and just by doing that I've learned several design principles I was later able to use in my own projects

[colleesu]

In developing an eLearning course on American Government, I spent a lot of time with the SMEs and a graphic designer developing the flow of how a bill becomes a law. When we were done, we had a bit of eye candy the students could look at, but not process. I realized in that moment I should have created an activity where groups of students create a graphic and share it with the other class participants to discuss and debrief. They'd learn more by creating the graph and explaining it than by passing their eyes over it.

[ty_]

Learning by doing works if the teacher or self studying student is able to facilitate the session at the end to bring them back to the learning outcomes, because if not students might focus on the wrong aspects of the activity and not really learn, or they might come to the "wrong conclusions"

[cooledcannon]

I would say learning by doing is best when practice or muscle memory or real life experience is important and learning by doing is worst when understanding and analysis is important.

[GustavoSilva-nyjc]

1:44 Okay, now that's an avengers level threat to my focus

[GustavoSilva-nyjc]

1:50 That's mind blowing, never thought about trying to make students reinvent the wheel, can you imagine how much
"chemical experiments and explosions" it took to make certain discoveries??? Might be laughable and boderline impossible to expect people, especially inexperienced brats to that on a single afternoon, it's the of thing that happen in films to show how genius someone is (and people buy that!! Remember Will Hunting?!). But this gives me some anxiety, because your channel supercharged my learning almost more than any other source and i had great discoveries over the years that exploded my productivity, "wouldnt be better to just search more??? What if i never heard of superforecasters, cargo cult and victorian doctors?! I feel like an idiot testing so little info that might be proven irrelevant like when i discovered the stoics, nietszche and jung were outdated as heck" but it needs to be done... cause my brain is dumb and requires testing.

[mr.winter]

I think learning by doing is essentially a way of forcing discipline on your learning style. As far as I understand it, information is best encoded when one relates it to other information and tries to find key concepts that connect it to a broader context. A lot of good learning by doing essentially forces you to do this. Taking the clown example from the video, when learning about fractions like this one is unable to just memorize the rules for how to do calculations with fractions and instead one is forced to understand why fractions are a thing, what they are used for and what they represent in reality.
If one already implements these more advanced learning techniques there is little point in doing the practical activity, as it only wastes time. If one would have thought about how the way a heart is built relates to the concept of arteries and veins anyway, forcing them to think about it by investing time into building a model is just going to waste their time.
So if what I wrote is true, learning by doing is essentially like a test of the way one thinks. If doing an activity is not significantly harder than thinking about it, that does imply that one already thinks of it in a very broad way, but if doing it is significantly harder, that’s probably a sign that one has only learned isolated facts. If the attempt to apply facts adds difficulty, that is probably because it adds the need for relating the facts one has so far only memorized.
P.S. I strongly encourage you to continue with this channel as I think the topic is important and that you really add a lot of value to this platform.

[randomguy]

The baking soda volcano is not to teach about volcanos but about exothermic reactions

[englishwithanes]

I would like to see a video about self-learning its pros and cons..
Is it true we can learn anything by ourselves and if so to what extent ??

I mean..you can definitely learn basic maths or some germany grammar or programming by yourself..
But what about things like medecine or other more academic fields..you can't become a doctor unless you go to college right ??

what about soft skills ??
What about swimming dancing or other motor skills..you need a coach right ??

[englishwithanes]

I was actually wondering whether you have a researchgate or google scholar account, are you a researcher?

Also, are you writing a book?