What we were taught
“You’re so smart.” “You’re a natural.” “You can do anything.”
We say it because we love them, and because we were taught that confidence works like a fuel tank: fill it with enough encouragement and a child will run on it. Tell them often enough that they’re capable, and they’ll believe it.
So it’s confusing when it doesn’t work. The child who has heard “you’re so clever” a thousand times still freezes at a hard maths problem. Still says “I’m just bad at this.” Still quits.
If praise built confidence, they should have plenty of it by now. Most of us quietly conclude that we need to praise more. The research suggests we’ve been filling the wrong tank.
What the research found
In 1977, the psychologist Albert Bandura set out where our belief in our own ability actually comes from. He called it self-efficacy — the expectation that I can do this — and argued it decides whether we start something, how hard we try, and how long we keep going when it gets difficult.
Bandura named four sources of that belief:
Doing it — our own successes
Watching others like us succeed
Being told we can
How our body feels — calm or panicked
And he made a pointed claim: the more dependable the source, the more it changes us. Being told you can do something is a weak source. Actually doing it is a strong one.
That was a theory in 1977. Thirty years of research has since tested it. In 2008, Ellen Usher and Frank Pajares reviewed the studies of these four sources in schools and concluded that mastery experience — actually doing it — is typically the most influential, though how much each source matters shifts with a student’s gender, background, ability and subject. In 2017, Angela Byars-Winston and colleagues combined 61 studies of 8,965 students and found the same thing more sharply: success at the task itself had by far the strongest link to confidence of the four sources.
Here’s what surprised researchers. For twenty years, the most famous study on praise said that how you praise makes a big difference. In 1998, Claudia Mueller and Carol Dweck ran six studies with fifth graders and found that children praised for being intelligent gave up sooner after a setback, enjoyed the task less and did worse than children praised for their effort. It became one of the most quoted findings in parenting.
Then other researchers tried to repeat it. In 2019, Yue Li and Timothy Bates ran the same experiment with 624 children aged 9 to 13 in China. In their closest copy of the original, the type of praise made a small difference to performance after failure, but no difference to any of the eight measures of motivation and attitude the original study used. In two further studies, it made no difference at all. The same year, Jaap Glerum and colleagues followed the original procedure with 108 vocational students, and found that students praised for effort and students praised for intelligence reacted to setbacks in exactly the same way.
So swapping “you’re so smart” for “you worked so hard” is probably not the lever we were told it was. The wording of the praise matters far less than whether the child actually did something hard.
The largest piece of evidence is recent. In 2024, Joshua Howard, Gavin Slemp and Xiao Wang combined 637 samples covering 388,912 students into a single analysis. They asked what happens when the adults around students support — or undermine — three basic psychological needs: autonomy, competence and relatedness. Supportive teacher behaviour went with better performance, engagement and wellbeing. Undermining behaviour went with the opposite.
One honest caution: those are correlations. Supportive adults and thriving students go together, but the analysis can’t prove which causes which, and the authors say so themselves.
What hasn’t changed
The three needs come from self-determination theory, which Edward Deci and Richard Ryan began building in the 1980s. They set out the three needs in their most-cited paper in 2000, and on our timeline the theory reappears and is refined in every decade since. (Finding — the theory has held up across forty years of research.)
Read through those three needs, confidence stops looking like a feeling you can give a child. It looks like something that grows when three things are true at once:
Competence — “I actually did it.” Not “I was told I could.” A real thing, done, just beyond what they could do before.
Autonomy — “I chose it.” Autonomy isn’t independence or doing everything alone. It’s the sense that the effort is theirs: their choice, their way of tackling it.
Relatedness — “Someone was there when I fell.” Children take risks when they feel safe enough to fail in front of people who care about them.
Notice what’s missing from that list: being told you’re clever. Praise isn’t useless. Bandura counts being told as a real source, just a weak one. But no form of words — “smart”, “hard-working”, “amazing” — can do the job of a child actually doing something hard, choosing how, with someone beside them.
And there is something every adult already knows from their own life: the confidence that lasts came from the time we did something hard, not the time someone said we could. (Wisdom — long-standing observation, not a measured finding.)
What to do this week
Set up one “just-right” challenge: something your child can almost do, but not quite yet. A recipe, a puzzle, a ride without stabilisers, a tricky paragraph. Then make three small moves — one for each need:
Let them choose how to tackle it. “How do you want to start?” (Autonomy: the effort is theirs.)
Stay near, and stay warm. Be interested in what they’re trying, not just whether it works, and let them see that getting it wrong is fine. (Relatedness: it’s safe to fail in front of someone who cares.)
When they’re stuck, help them find the next step instead of fixing it. “Which part is tricky? What could you try?” And when they get there, describe what they did, not what they are: “You tried three different ways before that worked.” (Competence: they did it, and they can see that they did.)
Then watch whether, next time, they reach for something hard on their own. That is confidence growing.
Sources
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological Review. https://doi.org/10.1037/0033-295X.84.2.191 (paywalled)
Mueller, C. M., & Dweck, C. S. (1998). Praise for intelligence can undermine children’s motivation and performance. Journal of Personality and Social Psychology. https://doi.org/10.1037/0022-3514.75.1.33 (paywalled)
Howard, J. L., Slemp, G. R., & Wang, X. (2024). Need support and need thwarting: A meta-analysis of autonomy, competence, and relatedness supportive and thwarting behaviors in student populations. Personality and Social Psychology Bulletin. https://doi.org/10.1177/01461672231225364 (paywalled)
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist. https://doi.org/10.1037/0003-066X.55.1.68
Usher, E. L., & Pajares, F. (2008). Sources of self-efficacy in school: Critical review of the literature and future directions. Review of Educational Research. https://doi.org/10.3102/0034654308321456 (paywalled)
Byars-Winston, A., Diestelmann, J., Savoy, J. N., & Hoyt, W. T. (2017). Unique effects and moderators of effects of sources on self-efficacy: A model-based meta-analysis. Journal of Counseling Psychology. https://doi.org/10.1037/cou0000219 (paywalled)
Li, Y., & Bates, T. C. (2019). You can’t change your basic ability, but you work at things, and that’s how we get hard things done: Testing the role of growth mindset on response to setbacks, educational attainment, and cognitive ability. Journal of Experimental Psychology: General. https://doi.org/10.1037/xge0000669 (free to read)
Glerum, J., Loyens, S. M. M., Wijnia, L., & Rikers, R. M. J. P. (2019). The effects of praise for effort versus praise for intelligence on vocational education students. Educational Psychology. https://doi.org/10.1080/01443410.2019.1625306 (free to read)
Evidence label — Strong for the core claim: confidence grows most from actually doing something hard. A 2008 review and a 61-study meta-analysis (8,965 students) agree that mastery is the strongest source, and a 637-sample meta-analysis links need-supportive adults to engagement and wellbeing. All of it is correlational — it shows mastery and confidence travel together most closely, not a controlled experiment proving cause. The famous praise-for-intelligence finding did not replicate in two later studies (624 children; 108 students), so the piece no longer rests on it.