The Confidence You Want for Your Child Comes From Challenge

Confidence isn’t something children need before they try hard things. It’s something they build because they try hard things—with agency, support, and meaningful work.

Action comes before confidence.

This is a core theme we see daily in Discovery Studio: learners take action, hit friction, iterate, and then discover they can do more than they thought.

Why “action comes before confidence” is backed by research

In psychology, a close cousin of confidence is self-efficacy—the belief that “I can do this.” Decades of research on self-efficacy shows that the strongest source of this belief is mastery experiences: successfully working through something difficult and seeing progress. In other words, confidence is built after learners take action, struggle, and improve—not before. That’s why “action comes before confidence” isn’t just a slogan—it’s consistent with how confidence actually forms. (Bandura, 1997)

The hidden risk of “easy”

When learning stays consistently easy, learners don’t just remain comfortable—they often become bored. And boredom isn’t neutral. Research on academic boredom has found meaningful links to lower engagement and higher burnout, suggesting that under-challenge can quietly erode motivation over time. (Study on academic boredom, engagement & burnout)

Discovery Studio: challenge designed on purpose

Discovery Studio is built around meaningful, project-based work—work that requires learners to build, test, revise, and explain their thinking. Two pillars have shaped our year: the Machine Learning Quest and Science Thursdays: Mission to Mars. Writer’s Workshop and Civilization run alongside these projects, adding daily cognitive “resistance training.”

Discovery Studio is built on collaboration, iteration, and real work that matters.

The Machine Learning Quest: building an AI application (and meeting real frustration)

In this quest, learners aren’t simply “learning about AI.” They’re building an AI application— something that takes inputs from the real world and produces an output that can be used in a real scenario. It starts with excitement: a bold idea, a vision for what the app will do, and a sense that “this should work.”

Then reality arrives in one word: data. Learners quickly learn that machine learning isn’t magic—it’s pattern recognition based on the examples the system is given. If the examples are limited, messy, inconsistent, or skewed, the model becomes limited, messy, inconsistent, or skewed. A model might work brilliantly in one moment and fail in the next. It might recognize one learner’s voice, face, or pose—but struggle with another. It might do fine under one lighting condition and fall apart when the background changes.

This is where the productive frustration shows up. Learners hit moments like: “Why did this work yesterday but not today?” “Why is it misclassifying this?” “Why did accuracy get worse after I added more data?” These aren’t signs of failure—they’re signs the learner is now doing real machine learning.

And then comes the big realization: bias can happen “by accident.” If training data represents only a narrow slice of the world (one person, one background, one angle, one tone of voice), the model starts to treat that slice as “normal.” Learners aren’t just told that data bias exists—they experience it and learn how to respond: add variations, broaden the dataset, label more carefully, test across conditions, and revise assumptions.

We don’t rescue learners out of that frustration. We coach them through it. They slow down, collect better data, test systematically, and iterate. And when the model improves, something bigger than technical skill emerges: earned confidence. “I didn’t know how to do this… I struggled… I kept going… and now I can.” That’s a mastery experience—the strongest builder of self-efficacy in the research. (Bandura, 1997)

Deep work often looks like focus, iteration, and peers helping peers through a stuck moment.

Science Thursdays: Mission to Mars (project-based learning with real constraints)

Science Thursdays are built around Mission to Mars—a sequence of engineering and science challenges where learners design, test, and refine solutions the way real teams do. This isn’t “read about Mars.” It’s build for Mars: prototype, test, collect observations, and revise.

Each mission comes with constraints: limited materials, specific criteria, and trade-offs. Learners plan before building, then discover (often quickly) that first attempts rarely work. Sometimes a design looks brilliant… until testing reveals the weak point. Other times it works—but not reliably—forcing learners to decide whether to refine the current approach or rethink the design entirely.

This is where project-based learning becomes more than “hands-on.” It becomes habits of mind: persistence, collaboration, evidence-based decision making, and iteration. Learners practice disagreeing respectfully, revising plans, and letting go of solutions they were attached to—because the data says it’s not working yet.

Over time, learners start to normalize iteration. They move from “This didn’t work” to “Good—now we know what to try next.” And that shift is a confidence builder, because it changes identity: I am someone who can work through hard problems.

We build curiosity and courage through real experiences—then bring that energy back into challenging projects.

Why challenge builds confidence (and resilience) long-term

Learners build the confidence that lasts by accumulating real proof: “I can start before I feel ready. I can get stuck and keep going. I can improve through effort.” Research links self-efficacy to resilience-related outcomes— how learners respond under stress, setbacks, and uncertainty. (Self-efficacy & resilience)

Just as importantly, when learners have agency and feel supported, motivation rises. Self-Determination Theory research shows that autonomy-supportive, need-supportive environments are associated with stronger motivation and engagement outcomes. (Need-supportive teaching & student motivation)

We don’t wait for learners to feel confident before giving them hard things to do.

We give them hard things to do—so confidence can emerge through action.

Final takeaway

If we want confident learners, we can’t build a world where everything feels easy. We build a studio where learners take action, face friction, and learn how to persist. That’s where confidence comes from. That’s what “action comes before confidence” looks like in real life.

Research & sources

1. Bandura, A. (1997). Self-Efficacy: The Exercise of Control. Link
2. Academic boredom, engagement, and burnout (evidence linking boredom with burnout and lower engagement). Link
3. Need-supportive/autonomy-supportive teaching and student motivation outcomes (Self-Determination Theory). Link
4. Self-efficacy and resilience-related outcomes. Link