Confidence: What the Evidence Actually Supports
Confidence attracts more bad advice than almost any other topic, and a surprising amount of the popular material rests on findings that have since fallen over. These are the questions people actually arrive with, answered as honestly as the evidence allows.
Is confidence the same thing as self-esteem?
No, and the distinction is the most useful one here.
Self-esteem is a global evaluation — how you feel about yourself as a whole. Self-efficacy, the concept psychologists actually use for this, is belief in your capability to do a specific thing. You can have high self-efficacy about cooking and none about public speaking, and the two do not average into a single number.
This matters practically because global self-esteem is hard to move and specific self-efficacy is not. "Become a confident person" is an unworkable project. "Become someone who can run this kind of meeting" is a tractable one, and the second is what people usually want anyway.
There is a third option worth knowing. Research comparing self-esteem with self-compassion has found that self-compassion delivers most of the benefits attributed to self-esteem without its main drawback — that self-esteem is contingent on continuing to do well, and so collapses precisely when you fail. Self-compassion does not require you to evaluate yourself favourably at all.
Where does confidence actually come from?
Bandura's account identifies four sources, and they are not equal.
Mastery experiences — having actually done the thing before. This is by a wide margin the strongest source, and nothing else comes close.
Vicarious experience — watching someone similar to you succeed. Works best when the model is genuinely comparable; watching an expert is less useful than watching a peer.
Verbal persuasion — being told you can. Real but weak, and it degrades fast if unsupported by results.
Physiological and affective states — how your body feels, and crucially how you interpret it. The same racing heart can be read as fear or as readiness, and the interpretation is somewhat modifiable.
The ordering has a blunt implication. Most confidence advice targets the third and fourth sources, which are the two weakest, because they are the two that can be addressed without doing anything difficult. The strongest lever requires attempting the thing, badly, repeatedly, at a scale small enough to survive.
Does "fake it till you make it" work?
Partly, and less than the famous version claims.
The best-known research here suggested that holding an expansive posture for two minutes changed testosterone and cortisol levels and improved risk tolerance. The hormonal findings failed to replicate in a larger study, and one of the original authors subsequently stated publicly that she no longer believed the effect was real.
What has partially survived is smaller: posture may affect self-reported feelings of power. Whether that translates into behaviour or outcomes is contested and the effect sizes are modest.
So: standing up straight before something difficult is free and may make you feel marginally better. It is not an intervention, and treating it as one displaces effort from the thing that works.
Do affirmations help?
This depends on which kind, and one of them can backfire.
Repeating positive self-statements — "I am confident", "I am a lovable person" — has been found to make people with low self-esteem feel worse, not better. The proposed explanation is that a statement too far from your current self-belief prompts you to generate counter-evidence, so you spend the exercise rehearsing reasons it is false.
A different practice, values affirmation, has held up considerably better. Here you write briefly about a value that matters to you — fairness, family, craft — and why. It does not require you to claim anything about your abilities, and it appears to reduce defensiveness under threat.
The practical version: writing about what you care about is a better use of ten minutes than telling yourself you are great.
Is imposter syndrome a real thing?
It is a real, described pattern, though "syndrome" oversells it — it is not a clinical diagnosis. It was first characterised in the late 1970s in a study of high-achieving women, and has since been documented across genders and fields.
The pattern is specific: persistent difficulty internalising success, attributing achievements to luck, timing, charm, or error rather than to capability, and an accompanying fear of being found out.
Two things are worth knowing. It correlates with competence rather than incompetence, since it requires achievements to misattribute. And a growing line of criticism argues that framing it as an individual psychological defect obscures situations where the feeling is a reasonable response to an environment — being the only person of your background in a room does not require a syndrome to explain the discomfort.
The useful move is usually not reassurance but record-keeping. Attribution errors are hard to argue with in the abstract and much harder to sustain against a written list of things you did and the specific decisions you made in them.
Isn't Dunning-Kruger the opposite problem?
Roughly, and the popular version is not quite what the research found.
The finding is about miscalibration: across skill levels, people's self-assessments are compressed toward the middle. Low performers overestimate; high performers frequently underestimate. The popular retelling — that incompetent people are uniquely and wildly overconfident — overstates one half and drops the other.
There is also a live statistical critique arguing that much of the classic pattern can be produced by regression to the mean and the way the data are plotted, which would mean the effect is smaller than advertised.
The half most relevant here is the underestimation. If you are competent at something and feel uncertain about it, that is a common and well-documented pattern, not evidence against your competence.
So what do I do in twelve minutes?
Make a cup of the ordinary daily sort — this is not a ceremonial dose and does not want to be — and use the time on the two things with actual support behind them.
Six minutes: evidence. Write down three specific things you have done that you were not sure you could do. Not achievements — instances. What the situation was, what you actually did, what decision was yours. This is mastery experience, retrieved rather than manufactured, and it is the only source on Bandura's list you can access while sitting down.
Six minutes: values. Pick one thing you genuinely care about and write about why it matters to you. Not about your abilities. This is the affirmation format that has held up, and it works precisely because it makes no claim about competence for you to argue with.
Then close the notebook and go and do something slightly beyond you, at a small enough scale that failing is survivable. That is the part that actually moves the number, and no amount of sitting substitutes for it.
Mastery experience is the only lever that reliably moves this, and it is not available from a chair. Go and do the small version. Our forgiveness ritual is the one to read if it has already gone badly.
Nothing about the cup matters here. Plain ground cacao and a mug is the entire setup, and the setup was never the obstacle.
