Is Your Teenager Learning How to Think for Themselves?

Ask a teenager why they believe something and you may get an immediate, forceful answer. That’s not necessarily a problem. Confidence can be useful, and a teenager with views is usually more interesting company than one without. But a quieter question tells you more: What would change your mind?

If the answer is “nothing,” the trouble is probably not a lack of intelligence. It may be a lack of practice letting evidence move a decision. And “thinking for yourself” is a skill that needs that practice.

Opinion is cheap and judgment has obligations

An opinion is a position. Judgment is a position that comes with obligations: to know something about the subject, to say what would count against it, and to own the result when it turns out wrong. Some questions are matters of taste, and taste is allowed. Some are real disagreements about values, and reasonable people stay split. But a great many of the questions a teenager faces, such as which design to build or which source to trust, have better and worse answers, and the quality of the reasoning shows up eventually.

Confidence and accuracy are only loosely related. The psychologist Philip Tetlock spent about two decades collecting forecasts from experts on political and economic questions. As a group, the experts were only slightly more accurate than chance, and simple extrapolation formulas did a little better than they did. The people who forecast best were those who held big ideas loosely and revised them when the facts changed. Sounding sure and being right turn out to be separate accomplishments.

The hopeful half came later. In forecasting tournaments that Tetlock ran with colleagues, a training module of less than an hour on basic probabilistic reasoning improved forecasters’ accuracy by roughly 6 to 11 percent over a control group. Judgment, in other words, can improve with training and practice.

Judgment starts with knowing things

One popular answer to this problem is to teach “critical thinking” as a general skill. The cognitive scientist Daniel Willingham has argued for years that this doesn’t work the way we hope. “There is not a set of critical thinking skills,” he writes, “that can be acquired and deployed regardless of context.” Thinking is intertwined with content. Told to “consider both sides” of an unfamiliar topic, a student can’t do it well, because they don’t know enough to tell which considerations matter.

This matters for parents because it changes what we should look for. A teenager who argues fluently about everything may be rehearsing a style. A teenager who knows enough about one thing to be surprised by it, and to notice when a claim doesn’t fit, is closer to the real article.

A case in which evidence won

In 1901 Wilbur and Orville Wright had a problem. Their new glider flew badly, and the numbers they had trusted didn’t predict why. Those numbers came from two sources: tables of lift and drag compiled by the German glider pioneer Otto Lilienthal, and a pressure coefficient, 0.005, published by John Smeaton in the eighteenth century and accepted for well over a century. By Orville’s later recollection, Wilbur was discouraged enough to say that man wouldn’t fly for a thousand years.

The brothers could have kept arguing about whose authority to believe. Instead they built a test. First a test rig built on a bicycle wheel, which suggested something in the published figures was wrong. Then a wind tunnel with a 16-inch test section, in which they tested model wings, by one detailed account 38 of them, each at 45 angles, and recorded what each did. By mid-December they had their answer: Lilienthal’s tables were largely right. Smeaton’s coefficient was wrong. They discarded one authority and kept the other, and they redesigned their wings accordingly.

Two things are worth noticing. They weren’t skeptics of everything; most of what they had inherited survived. And their conclusion changed what they built. Judgment was a loop between a belief, a test that could embarrass it, and a decision about what to do next.

A good decision can still turn out badly

The former professional poker player Annie Duke calls the opposite mistake “resulting”: treating the outcome of a decision as a reliable measure of how good the decision was. Her main example is the final play of Super Bowl XLIX. Seattle, trailing by four with 26 seconds left, was on the one-yard line with one timeout. Coach Pete Carroll called a pass, it was intercepted, and the call was widely mocked. Duke asks us to imagine the same pass caught for a touchdown. The headlines would have praised the call as brilliant.

Whether the call was actually sound is still argued, and that’s the useful part. The argument has to be about the reasoning: the clock, the timeouts, the down and distance, what the alternatives would have risked. It can’t be settled by pointing at the interception. Luck and hidden information stand between a decision and its result. A student who learns this can say, “That went badly, and my reasoning still looks defensible,” without it being an excuse. They can also say, “That went well, and I got lucky,” which is rarer and worth more.

None of this licenses carelessness. The point of separating process from outcome is to inspect the process, honestly, after every result.

Feedback that improves the work

Feedback is what turns decisions into learning, and it comes in different qualities. Ed Catmull, Pixar’s co-founder, describes in Creativity, Inc. a group of directors and writers called the Braintrust that watches a film in progress and gives frank notes. The group has no authority, he stresses: it can’t tell the director what to do. The notes work because they identify problems the filmmaker can’t see from inside, and because the filmmaker still owns the decision. Catmull is blunt about why the group exists: “early on, all of our movies suck.”

That’s a useful model for a family. The best feedback names what isn’t working for a listener or a user, and leaves the choice of fix with the person making the thing. Reality gives its own feedback, too: the prototype that fails or the song that loses a room.

A decision with a test attached

Imagine Dana, a sixteen-year-old on a school robotics team, who has to choose between two gear ratios for the robot’s drive: one faster across the field, one stronger when pushing against another robot. An older teammate favors speed. Dana writes down what matters (cycle time, pushing power, how hot the motors run), what she doesn’t yet know (how much pushing power the faster ratio gives up), and a rule she can be held to: if the fast ratio stalls against a weighted test sled, she’ll switch. She asks a mentor to poke at her reasoning. The fast ratio barely clears the sled, so by her own rule she keeps it.

At the next event the robot loses two pushing matches. Two losses could be bad luck, and Dana can’t tell from them whether her rule was wrong or the draw was. What the written rule gives her is something to argue with. She can see exactly what she predicted, notice that a sled test on a shop floor differs from a crowded carpeted field, and write a better rule. Without the note, hindsight would simply tell her she had known all along.

Writing a rule or prediction down before seeing the result is useful in some work and beside the point in others. It suits a test or a forecast, and it would be silly to demand for a chord choice or an open-ended exploration. A color or a cut can be a real decision, but creative judgment mostly runs through listening, trying and revising.

Why a critical-thinking unit isn’t enough

Many schools teach critical thinking, and some do it well. Willingham’s point is part of the answer: if thinking depends on content, then a unit on critical thinking won’t carry over by itself. What does seem to help is repeated decisions in a domain where the student has built real knowledge, tied to work that pushes back. That’s an old idea held to a demanding standard. It’s also a good question to put to any program: where, specifically, does the student decide something, and what happens when they’re wrong?

At the table

A family conversation about a decision goes furthest on three questions: What matters most here? What don’t we know? What evidence would change your mind?

The harder part is modeling it. Revise an opinion of your own out loud, with the reason. Teenagers learn a lot from watching an adult say “I was wrong about that, and here’s what moved me.” Be careful not to reward contrarianism, either. Disagreeing with the adults in the room isn’t independence. Changing one’s mind for good reasons, and holding firm for good reasons, is.

Judgment is one part of growing up capable. In engineering the central accomplishment may be craft. In music it may be technique and taste. Knowledge comes first in all of them. Judgment is the part that decides what to do with it.