Psychologist Albert Bandura's self-efficacy research found that genuine belief in one's own ability is built primarily through "performance accomplishments" โ€” actually succeeding at something real โ€” far more than through encouragement or praise. This has an uncomfortable implication for a lot of well-meaning parenting and product design: telling a child "you're so good at this" or handing them an easy win to feel good does much less for lasting confidence than a well-calibrated hard problem, solved.

What Bandura actually found

Bandura's 1977 paper identified four sources people draw on to form beliefs about their own capability, and ranked them by strength:

  1. Performance accomplishments (the strongest) โ€” actually doing the thing successfully, especially when it required real effort and wasn't guaranteed.
  2. Vicarious experience โ€” watching someone similar to you succeed.
  3. Verbal persuasion โ€” being told you can do it, by someone you trust.
  4. Emotional/physiological state (the weakest, and can cut either way) โ€” how calm or anxious you feel going in.

The ordering matters. Praise and encouragement โ€” verbal persuasion โ€” sit near the bottom. They can help someone attempt a task they'd otherwise avoid, but on their own they don't produce durable belief in capability. What does is the accumulated evidence of having actually done hard things and succeeded.

Why this cuts against a lot of instinctive parenting and app design

It's natural to want to protect a struggling child by softening the task or piling on encouragement โ€” "you can do it!", "great effort!", an easier follow-up question so they leave the session on a high note. Bandura's model suggests this often doesn't do what it's meant to do. A child who's told they're capable, without evidence, tends to know the difference. A child handed an easy win after a hard failure gets a boost that fades fast, because the win wasn't real evidence of anything.

The stronger move โ€” harder to do, less immediately comfortable โ€” is holding a task at genuine difficulty and letting the child solve it themselves, even slowly, even with mistakes along the way. The resulting "I actually did that" is the accomplishment Bandura's research says sticks.

Where step size comes back in

This is why step size (see step size explained) and self-efficacy are the same problem viewed from different angles. A task pitched too easy generates a "win" with no real evidentiary weight โ€” the student already knew they could do it. A task pitched too hard generates a failure, which is negative evidence, undermining self-efficacy rather than building it. Only a task at the edge of current ability โ€” hard enough to be a genuine test, achievable enough to actually be won โ€” produces the kind of accomplishment Bandura's research is describing.

This is also why effort matters specifically, not just correctness. A problem solved instantly, with no visible struggle, provides weaker evidence of capability than one that took real, sustained effort to crack โ€” see why effort earns points at aitutors.me for how this plays out in practice.

What this means for the "great job!" instinct

None of this means praise is useless โ€” verbal persuasion is still one of Bandura's four sources, and it can matter, especially for a child hesitant to even attempt something. But it works best as a nudge toward attempting a real challenge, not as a substitute for one. "I think you can get this one โ€” give it a go" (persuasion, pointed at a genuine task) does more long-term work than "well done!" attached to something that was never actually hard.

A practical test for parents

Ask: did my child's last "win" require anything of them? If a task was solved instantly with no hesitation, it likely didn't move the needle on self-efficacy, regardless of how you responded to it. If it took real, sustained effort โ€” even with some wrong turns along the way โ€” that's the accomplishment doing the actual work, whether or not you say anything about it at all.

FAQ

What is self-efficacy?

Albert Bandura's term for belief in one's own capability to succeed at a specific task. It's task-specific, not general self-esteem โ€” high in one subject, low in another, at the same time.

Does praising my child build their confidence?

Bandura's research found praise ("verbal persuasion") is the weakest of four sources of self-efficacy. The strongest is "performance accomplishments" โ€” actually succeeding at something that required genuine effort.

Why does an easy task not build real confidence?

Because self-efficacy comes from evidence about capability, and an easy success provides weak evidence โ€” the student already knew they could do it. A task that required real effort provides much stronger evidence.

Duke Harewood ยท founder, aitutors.me ยท Updated 11 Jul 2026.