On this page
"Learning by doing" is the phrase most often linked with John Dewey, but what he wrote was sharper than the slogan. For Dewey, doing something only becomes learning when you connect what you did to what happened as a result. Modern research supports a version of this, with an important condition: guidance. Here is what Dewey said, what studies found, and a five-step routine for any science experiment or problem.
Who was John Dewey?
The Stanford Encyclopedia of Philosophy describes Dewey (1859-1952) as one of the founders of American pragmatism and "arguably the most prominent American intellectual for the first half of the twentieth century." At the University of Chicago he founded the Laboratory School, "a site to test psychological and educational theories," with his wife Alice as principal from 1896 to 1904. He then moved to Columbia University. His books on education include How We Think (1910) and Democracy and Education (1916).
What did Dewey mean by experience?
In Democracy and Education, Dewey says experience has two parts. "On the active hand, experience is trying." On the passive side, "it is undergoing": we act on something, and then we live with the consequences. Activity alone is not enough: "Mere activity does not constitute experience." His example: "It is not experience when a child merely sticks his finger into a flame; it is experience when the movement is connected with the pain which he undergoes in consequence."
He also argued that school methods which are "permanently successful" "give the pupils something to do, not something to learn; and the doing is of such a nature as to demand thinking, or the intentional noting of connections; learning naturally results." And in one well-known line: "An ounce of experience is better than a ton of theory simply because it is only in experience that any theory has vital and verifiable significance."
What are Dewey's five steps of reflective thinking?
In How We Think, Dewey breaks reflective thought into "five logically distinct steps":
a felt difficulty;
its location and definition;
suggestion of possible solution;
development by reasoning of the bearings of the suggestion;
further observation and experiment leading to its acceptance or rejection.
This is close to what a scientist does, and it gives you a routine you can use in a lab or on a hard problem.
What does research say about learning by doing?
The evidence favours doing with guidance, not doing alone.
Unguided discovery underperforms. A meta-analysis by Alfieri and colleagues (2011), covering 164 studies, found explicit instruction beat unassisted discovery under most conditions (d = −0.38). But "enhanced" discovery, with feedback, worked examples, scaffolding or prompts to explain, beat other forms of instruction (d = 0.30).
Beginners need structure. Kirschner, Sweller and Clark argued in Educational Psychologist (2006) that minimally guided instruction is "less effective and less efficient" than strongly guided instruction, and that the advantage of guidance recedes only when learners have high prior knowledge. Our post on John Sweller's cognitive load theory explains why.
Inquiry works best when teachers lead it. A 2012 meta-analysis by Furtak and colleagues of 37 studies of inquiry-based science teaching found an overall mean effect size of 0.50, with teacher-led activities about 0.40 larger than student-led ones.
Read together, these fit Dewey's own condition: the doing has to "demand thinking." Hands-on work without a question and without feedback is the "mere activity" he warned against.
Worked example: Dewey's five steps on a pendulum
Felt difficulty. A heavy and a light pendulum seem to swing at the same rate. Is that right?
Define it. The question is whether the period T depends on mass, length or swing size.
Suggest an answer. For small swings, T = 2π√(L/g), so T depends on length and not on mass.
Reason out the consequences. If that is right, doubling the length should multiply T by √2 ≈ 1.41, and doubling the mass should change nothing.
Test it. Time 10 swings for lengths of 0.5 m and 1.0 m and divide to find T for each. Is the ratio close to 1.41? Repeat with a heavier bob at the same length.
Then add Dewey's reflection step in writing: what did you predict, what happened, and why? If your ratio came out at 1.6, ask what you did differently, such as a wide swing or a timing error. That written comparison is the "intentional noting of connections."
Students who like building things will also enjoy Seymour Papert's work on learning by making, and for the guided side of practice see worked examples versus solving from scratch.
If you do not have a lab, the Fermi Simulation Lab has interactive simulations for ideas that are easier to see than to read. Write down your prediction before you change a setting, then check it. Try the pendulum routine this week and keep the written prediction next to your result.
