Self-Beliefs - The three selves people confuse

What the science says

Consensus

We talk about "confidence" as if it were one thing. It is at least three, and the differences are the whole point of this article.

‍

The first is self-concept: your mental picture of who you are. It is descriptive, the answer to "who am I?", built from organised packets of self-knowledge that psychologists call self-schemas, which steer what you notice about yourself and quietly resist information that does not fit (Markus, 1977). "I am a shy person", "I am good with numbers", "I am a parent" are all self-concept.

‍

The second is self-esteem: how worthy you feel overall. It is evaluative, the answer to "do I like myself?", a global sense of your own value (Rosenberg, 1965). It is the one most people mean by confidence, and, as we will see, the one most oversold.

‍

The third is self-efficacy: your belief that you can carry out a particular task. It is predictive and narrow, the answer to "can I actually do this?", and it is specific to the job in hand (Bandura, 1977). You can have high self-efficacy for public speaking and low self-efficacy for parallel parking, regardless of how you feel about yourself in general.

‍

The reason to keep them apart is that they do not do the same work. Self-efficacy is the one that earns its keep: across a large body of evidence it reliably predicts how much effort people put in, how long they persist, and how well they perform (Stajkovic and Luthans, 1998). Better still, it can be built, and Bandura identified how, in order of power. The strongest source is mastery experience, actually succeeding at something, ideally a manageable version first, so a person accumulates evidence that they can. Next is vicarious experience, watching someone like you pull it off. Then verbal persuasion, encouragement, which helps a little. And finally managing your physiological state, since a racing heart read as "I am not up to this" undercuts the belief. The practical headline: efficacy is grown mostly through small real wins, not pep talks. It is also why guided, hands-on practice is such a powerful way to teach: learning a skill gradually with a mentor who helps you notice and make sense of each small win feeds several of those sources at once, real successes, a model to watch, and honest encouragement. It is the logic behind how teaching hospitals train doctors, through graduated, supervised responsibility rather than lectures alone.

Controversies

The most useful correction in this article concerns self-esteem, because it was the subject of one of psychology's great overreaches. Through the 1980s and 1990s, a movement held that raising people's self-esteem would lift school results, reduce crime and drug use, and generally fix social problems, and it reshaped parenting and schooling around making children feel good about themselves. Then a comprehensive review looked at the actual evidence (Baumeister et al., 2003). The verdict was deflating: self-esteem does correlate with good outcomes, but it mostly does not cause them. Where the two move together, it is usually because real success raises self-esteem, or because some third factor lifts both. High self-esteem feels pleasant and comes with a bit more initiative and resilience, but it does not manufacture achievement, and programmes built to inflate it did not deliver. Competence tends to produce the worth, not the other way round.

‍

A second popular belief turns out to be backwards. It is widely assumed that low self-esteem causes aggression, that bullies and violent people secretly hate themselves. The laboratory evidence says otherwise: simple self-esteem was unrelated to aggression, and what actually predicted it was threatened egotism, a high but fragile or inflated self-regard, especially narcissism, lashing out when challenged (Bushman and Baumeister, 1998). So "more self-esteem" is not a safety measure; brittle, defended high self-regard is part of the problem.

‍

Even self-efficacy, the hero of the piece, is not unconditionally good. When researchers tracked the same people over repeated attempts and nudged their confidence up, the extra confidence sometimes lowered the next performance, because it tipped into overconfidence and slackened effort (Vancouver et al., 2002). Belief in your capability helps when it is roughly accurate; inflated past that, it can make you coast or take sloppy risks. Two more cautions complete the picture. Self-esteem that is staked on external wins, looks, status, approval, is a standing vulnerability rather than a strength, because it must be constantly defended (Crocker and Wolfe, 2001). And the whole framework carries a cultural fingerprint: the prizing of high, independent self-esteem is far more a feature of individualist Western settings than a human universal, since many cultures build the self around connection and fitting in rather than standing out (Markus and Kitayama, 1991).

Limitations

Almost all of this rests on self-report, and the constructs blur at the edges and risk circular measurement (does the scale capture self-efficacy, or just optimism?). The self-esteem story in particular is a lesson in causation: correlations were read as causes for two decades. And the bulk of the research comes from Western, individualist samples, where the very idea that high self-esteem is the goal is itself a cultural assumption.

Open questions

Is self-esteem ever a genuine cause of anything, or always a readout of how things are going? What builds self-efficacy in a way that lasts rather than fading after the intervention? And how much of this picture survives outside the independent-self cultures it was mostly studied in?

So what

The usable core: separate the three, and when you want someone (including yourself) to actually do something, build the narrow, task-specific belief that they can, mainly through real wins, rather than trying to top up their general good feeling about themselves.

For companies

The practical mistake is trying to "build confidence" in the abstract. The lever that moves behaviour is self-efficacy, and it is built, not talked into being. For staff learning a new system or a new role, design early mastery experiences, a manageable first task they will genuinely succeed at, before the hard one, and pair it with models, visible examples of people like them doing it well. For customers, the same logic makes onboarding work: a product feels usable when the first interaction produces a small, real win, which is worth far more than reassuring copy. In messaging, the honest and effective move is to genuinely show people they can succeed at the specific thing you are asking of them, the same efficacy lever that makes a clear call to action land (L3-04) and a warning land rather than frighten (L3-05). What does not work is generic "you've got this" when the person has no evidence they have.

For political parties and education

The self-esteem movement is a cautionary tale worth learning from directly: pouring resources into making people feel good about themselves, detached from real competence, did not produce the achievement it promised (Baumeister et al., 2003). The durable version is to build the skills and the small wins that earn the confidence. At the collective level, shared efficacy, a credible sense that "we can actually do this", mobilises in a way that abstract pride does not, and it has to rest on a believable plan rather than slogans. Barack Obama's 2008 "Yes We Can" is a clear illustration, precisely because the slogan alone was never the point: it put collective efficacy at its centre, change as something voters were capable of achieving together rather than waiting to receive, and it mobilised so widely because that rousing line was backed by a serious organising and turnout operation that made the "we can" credible. The belief and the means to act on it reinforced each other.

For government

Behaviour-change programmes work better when they target self-efficacy than when they aim at mood or esteem: health, training, and financial-capability interventions land when they give people a concrete, achievable first step and the sense that they can take the next one (Bandura, 1977). The broader lesson from the self-esteem decades is humbling and useful: an appealing idea ("raise self-esteem and outcomes follow") drove real policy before anyone had shown it was causal, so the standing question for any feel-good intervention is whether it builds a capability or just a feeling.

How to use this

Three habits. First, separate the three selves: notice whether you are dealing with how someone describes themselves, how they feel about themselves, or whether they believe they can do the specific thing, because only the last reliably drives action. Second, build efficacy the way it is actually built, through small real successes and visible models, not pep talks, and keep it roughly honest, since overconfidence has its own costs. Third, stop chasing self-esteem as a target; it tends to follow real competence, and brittle, propped-up self-regard causes more trouble than it prevents.

‍

Case studies

• The self-esteem bubble (Baumeister et al., 2003). For two decades, schools and parenting absorbed the idea that boosting children's self-esteem would lift their grades and behaviour. A thorough review of the evidence found the causal arrow mostly ran the other way, or through third factors: self-esteem is more a by-product of doing well than an engine of it, and deliberate efforts to inflate it did not produce the promised gains. It is one of the clearest cases of an attractive idea outrunning its evidence and shaping policy before the science was in. DOI 10.1111/1529-1006.01431
• Who actually turns aggressive (Bushman and Baumeister, 1998). Testing the common belief that low self-esteem drives violence, researchers found no such link. What predicted aggression was threatened egotism, people with high but fragile or narcissistic self-regard, retaliating when it was punctured. The finding overturns a piece of folk wisdom and sharpens the point that "more self-esteem" is not a fix, and that the shape of self-regard matters more than its level. DOI 10.1037/0022-3514.75.1.219

‍

References

• Bandura, A. (1977) 'Self-efficacy: toward a unifying theory of behavioral change', Psychological Review, 84(2), pp. 191–215. Available at: https://doi.org/10.1037/0033-295X.84.2.191 (Accessed: 18 June 2026).
• Baumeister, R.F. et al. (2003) 'Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles?', Psychological Science in the Public Interest, 4(1), pp. 1–44. Available at: https://doi.org/10.1111/1529-1006.01431 (Accessed: 18 June 2026).
• Bushman, B.J. and Baumeister, R.F. (1998) 'Threatened egotism, narcissism, self-esteem, and direct and displaced aggression: does self-love or self-hate lead to violence?', Journal of Personality and Social Psychology, 75(1), pp. 219–229. Available at: https://doi.org/10.1037/0022-3514.75.1.219 (Accessed: 18 June 2026).
• Crocker, J. and Wolfe, C. (2001) 'Contingencies of self-worth', Psychological Review, 108(3), pp. 593–623. Available at: https://doi.org/10.1037/0033-295X.108.3.593 (Accessed: 18 June 2026).
• Markus, H.R. (1977) 'Self-schemata and processing information about the self', Journal of Personality and Social Psychology, 35(2), pp. 63–78. Available at: https://doi.org/10.1037/0022-3514.35.2.63 (Accessed: 18 June 2026).
• Markus, H.R. and Kitayama, S. (1991) 'Culture and the self: implications for cognition, emotion, and motivation', Psychological Review, 98(2), pp. 224–253. Available at: https://doi.org/10.1037/0033-295X.98.2.224 (Accessed: 18 June 2026).
• Rosenberg, M. (1965) Society and the Adolescent Self-Image. Princeton: Princeton University Press.
• Stajkovic, A.D. and Luthans, F. (1998) 'Self-efficacy and work-related performance: a meta-analysis', Psychological Bulletin, 124(2), pp. 240–261. Available at: https://doi.org/10.1037/0033-2909.124.2.240 (Accessed: 18 June 2026).
• Vancouver, J.B., Thompson, C.M. and Tischner, E.C. (2002) 'Two studies examining the negative effect of self-efficacy on performance', Journal of Applied Psychology, 87(3), pp. 506–516. Available at: https://doi.org/10.1037/0021-9010.87.3.506 (Accessed: 18 June 2026).

‍