On Failure

The definition of insanity [1] is doing the same thing over and over again and expecting a different result.

I hate that saying. I much prefer:

If at first you don't succeed, try, try again.

Why? Because things rarely work the first time, and because “the same thing” is almost never exactly the same thing.

In research we try to account for all the variables, and to hold as much constant as possible. But we can't control everything, and we don't even necessarily recognize all the variables: different time of day, different batch of materials, different weather, different container, different attitude, different anything. And any one of those different things may cause the experiment to fail.

And it may fail again, and again, and again before all the right conditions are met.

Experimental scientists [2] fail a lot. We fail all the time, even most of the time. We learn from our failures, dust ourselves off, and “try, try again.”

Think about that saying:

If at first you don't succeed, try, try again.

It's not “if at first you don't succeed, give up,” or “if at first you don't succeed, you're stupid,” or even “if at first you don't succeed, try again.”

Try, try again. And then try again after that. Once isn't going to be enough.

Don't trust the scientist who has never failed; they're probably lying. Failure is a part of research; it's the challenges and obstacles you puzzle over, work around and overcome. It's the majority of the problems in “problem-solving.”

But it can be demoralizing to fail.

It's pretty easy to conflate “the experiment failed” with “I failed,” and “I failed” becomes “I am a failure” all too quickly, as the tally of failed experiments grows.

Here is the gateway to self-doubt, hanging wide open. Don't walk through; it'll snap closed behind you, and it's hard to climb out.

Laboratory courses, even those founded on “discovery-based” learning, do not teach students to fail. Failure of an experiment is a failure of the student. If classroom demonstrations always have known outcomes and go smoothly, if teaching labs punish failed experiments, aren't the students trained to fear being wrong and trying things that might not work? Of course, it's helpful to illustrate a concept or demonstrate a trend with a successful experiment, but what is the price of this success? A student who fears failure and then begins a research project is unprepared for the “try, try again” part.

Doing something over and over and looking for a different result is not the definition of insanity, it's the definition of experiment.[3]

[1] Sometimes “stupidity” is used instead of “insanity.”
[2] I make no claims about the rest.
[3] experimental : based on untested ideas or techniques and not yet established or finalized