What the data actually shows
Bandura's self-efficacy theory identifies four main sources of confidence, and they are not equally powerful. The strongest is mastery experiences — actually performing a task successfully. Direct experience of coping and succeeding is the most durable input, because it is the most credible: you are not telling yourself you can do it, you are showing yourself you already did. This is why competence and confidence usually rise together, with competence leading.
The second source is vicarious experience — watching people similar to you succeed. Seeing someone comparable manage the thing raises your own belief that you could too, which is part of why relatable role models and peers matter more than distant, exceptional ones. Third is social persuasion — credible encouragement from others — which can help on the margin but is weaker and fades quickly if it isn't backed by real experience. Fourth is your physiological and emotional state: how you read your own racing heart, nerves, or fatigue. Reinterpreting those signals as readiness rather than as proof of impending failure can shift confidence, but it is a supporting lever, not the main one.
The ordering is the practical heart of the theory. Because mastery dominates, the most reliable way to become more confident at something is to do it, in manageable steps, and let success accumulate. Affirmation-style self-belief, sitting in the social-persuasion or self-talk category, is one of the weaker sources — and research on positive affirmations finds the effects are generally small, and can even backfire for people with low self-esteem, who tend not to believe the statements they're repeating.
Why this feels different from how it actually is
Confidence feels like a trait — something some people simply have and others don't — because we mostly encounter it as a finished product. We see a confident colleague presenting fluently and infer an inborn quality, without seeing the dozens of earlier, clumsier attempts that built it. The mastery history is invisible, so the confidence looks innate.
It also feels like it should come first because that's how the most visible advice frames it: believe in yourself, then you'll be able to act. That sequence is intuitive and appealing, but it inverts the more reliable order the research describes. The result is a common trap — people wait at the starting line for a feeling that, in practice, only arrives once they've crossed it a few times.
And nerves get misread. Because we expect confident people to feel calm, we treat our own pre-performance anxiety as evidence that we're not ready or not cut out for it. In reality, even experienced performers feel the physiological arousal; the difference is largely in how that arousal is interpreted, not in whether it's present.
What the research says to do about it
The clearest implication of self-efficacy theory is to build confidence through graded mastery: break the thing into steps small enough that you can succeed, then increase the difficulty as each success accumulates. Each completed step is direct evidence that you can cope, which is the input the theory says matters most. Acting before you feel ready, at a manageable scale, is the mechanism — not a personality flaw.
Use vicarious experience deliberately by seeking out relatable models — people roughly like you who have done the thing — rather than distant, exceptional ones. Watching a comparable peer succeed raises your own efficacy belief more than watching a virtuoso, who is easy to dismiss as a different kind of person. Mentors and supportive communities work partly through this channel.
Reinterpret your physiological state rather than trying to eliminate it. Research on reappraising arousal suggests that reading a racing heart and nervous energy as readiness or excitement, rather than as a warning sign, can improve performance and the confidence that follows. The goal isn't to feel nothing; it's to stop treating normal nerves as proof you'll fail.
What the research says does not help
Affirmation-only approaches — repeating 'I am confident' or 'I will succeed' without any backing experience — are among the weakest levers, and the research on positive self-statements finds the effects are small and sometimes negative. For people who don't already believe the statement, affirmations can highlight the gap between the claim and their self-image and leave them feeling slightly worse, not better.
Waiting to feel confident before you begin reliably backfires, because the experience that would build confidence is the very thing being postponed. Confidence is largely downstream of action; deferring action to wait for it keeps you stuck at the input that matters least and starves the input that matters most.
Trying to suppress nerves entirely is both difficult and unnecessary. Physiological arousal before a challenge is normal even for highly capable people, and treating its mere presence as a problem can become its own source of doubt. The evidence favours reinterpreting the arousal over fighting to make it disappear.
Real numbers in context
Confidence research is about mechanisms more than tidy percentages, so it resists a single headline statistic — and any source offering a precise 'X% of confidence comes from doing' is overstating the evidence. What Bandura's framework supports is an ordering: mastery experiences are the strongest and most durable source of self-efficacy, vicarious experience and social persuasion are progressively weaker, and managing your physiological state is a supporting lever. The practical figure worth remembering is not a number but the rank order.
The affirmation evidence is the clearest cautionary data point. Studies on repeating positive self-statements find generally small effects, with some research suggesting they can backfire for people with low self-esteem. That doesn't make encouragement worthless — credible social persuasion has a real, if modest, place — but it does mean 'just believe in yourself' sits near the bottom of the list of things that actually build confidence, not at the top where popular advice puts it.