No, People Do Not Learn Best by Failing

I was walking with my wife and she told me that a colleague’s child, a middle school student, was running a science experiment in which he “watered” plants with a range of liquids, in addition to a “control” of water. The experiment’s purpose was to see if some liquids benefited the plants more than water. The liquids chosen included milk, orange juice, and other common things that humans drink.

(Water had the best result.)

My comment to my wife was that it sounded like the teacher had not taught the foundation of experiments. A scientific experiment is not usually a trial of random things. Rather, an experiment is usually based on an insight, such as “Aha! — I think that X happens because of Y!” That is called a hypothesis, and an experiment is designed to test if a hypothesis is true.

That’s not what the student did in this case, and it makes me think that the teacher did not explain the scientific process; or if they did, they are not coaching the student in how to design a hypothesis and experiment. Instead, they are letting the student “just try shit”. That’s not learning, because if, say, orange juice had worked better than water, there would be hypothesis that was proven — nothing would be learned in terms of how things work. Instead, a mystery would be uncovered: “Why does orange juice work better than water?” There would be no starting point. Orange juice is a complex mixture of cells and substances.

There is a class of scientific experiments that test random things. This is done, for example, to find chemicals that boost or limit growth of certain kinds of cells, such as cancer cells. However, the chemicals are not chosen at random, but are carefully extracted and purified from organisms. Thus, when a chemical effect is found, there is a starting point, because the chemical has been isolated.

In the Agile community, there is a widespread narrative that people learn by trying things and failing. That is an enormous distortion of how learning occurs. Yes, people do learn when they fail, but often they learn the wrong lesson.

For example, suppose your car’s radiator hose bursts. Assume that you know nothing about cars, so you try wrapping silver tape around the leak. But you need to refill the fluid, so you make a guess: “Let me try milk”. That fails. Did you learn?

Of course not. You did not learn because you had no working theory on which to base your attempt. If you had a theory, informed by what others had explained to you, that a radiator contains radiator fluid, your experiment would have been better constructed: You would have examined radiator fluids in a store and picked one: that’s your experiment. Did you pick the right one? You will find out, but at least you are not making wild guesses.

The same idea applies in our everyday work. For example, suppose that you lead a team and a complex issue comes up. You need to have quality discussions with your team to solve the issue. However, if you have never learned anything about leadership, you will be trying things at random. Depending on your personality, you might do one of the following:

1. Figure it out yourself and tell everyone what to do.
2. Challenge the team to figure it out and walk away.
3. Engage in an active discussion and make a final decision.
4. Facilitate a discussion and go with what the team decides.
5. Ask the team members to write down their ideas and send them to you: you will pick the best one.
6. Some combination of the above.

On the other hand, if you have studied the subject of team leadership, you will view the situation through a more nuanced lens. You will know that groups have better discussions if the participants first think independently and then convene. You will also know that a group of people tend to be unable to “connect the dots” on complex issues. And you will know that people often do not know what they don’t know, and often assume they know more than they know.

All these things will inform the approach that you try. Thus, you might approach the situation differently. You will still experiment, but your experiment will be informed by theory. It will be a smarter and more effective experiment, and you will likely learn more than if you had not started with the theoretical knowledge.

Thus, failure is not the thing to celebrate. Learning is. And you learn more if you are thoughtful and informed. You learn more if you are not too cautious, but also not reckless. You learn more if you try things carefully and with intention, using an approach designed to narrow in on an area that needs to be understood better.

Failure is never something to celebrate. Learning is.