Self-Compassion vs Self-Esteem: What Builds Lasting Confidence

If your confidence rises after a good week and drops after a poor one, the question of self-compassion vs self-esteem is worth a closer look. The two are often treated as the same thing, or as rivals. They are neither. They describe different parts of how you relate to yourself, and the difference matters most when something goes wrong.

This article defines both terms, walks through what one small study found, is clear about how far that evidence can carry, and ends with a short experiment you can run on your next setback.

This article is educational only and is not psychological, medical or therapeutic advice.

Two things that get mistaken for each other

Self-esteem is a global evaluation of yourself: a general sense of how much you value and approve of who you are. In research it is commonly measured with the Rosenberg Self-Esteem Scale, a self-report questionnaire [1].

Self-compassion is a separate idea. It is described as a recently introduced self-construct, measured by Neff's Self-compassion Scale (2003), and it is moderately correlated with self-esteem [1]. Related, in other words, but not identical.

A plain way to hold the distinction:

• Self-esteem is a reading: how highly you currently rate yourself.
• Self-compassion is a response: how you treat yourself when the reading is poor, or when you have fallen short.

Neither definition makes the other an enemy. You can have a high opinion of yourself and still be harsh after a mistake, or hold a modest opinion and respond to errors steadily. Nothing here suggests self-esteem is bad; the point is that it is only one part of the picture.

Why confidence built only on a rating can be fragile

Here is the practical concern, offered as reasoning rather than as a research finding. A rating of yourself has to be fed by something. For many people that something is recent results and other people's opinions. If confidence is the rating, then confidence moves whenever the inputs move: a strong quarter lifts it, a critical email lowers it.

That is a practical suggestion about how the mechanism might work, not something the research cited here has tested. What the research does offer is a narrower and more cautious point: in one student sample, higher self-esteem was not a clean marker of healthy functioning. It was positively related to aggression [1]. The next section explains what that means and what it does not.

What one study found about self-esteem, self-compassion and aggression

The study is a 2009 doctoral dissertation in school psychology [1]. Here is what it did and found.

The sample and measures. A total of 181 college students at a large Northeastern US university completed five online self-report surveys: the Rosenberg Self-Esteem Scale, the Self-compassion Scale, the Narcissistic Personality Inventory, the Marlowe-Crowne Social Desirability Scale and the Aggression Questionnaire [1]. The analysis used hierarchical regression, which tests how much additional variation in an outcome each new variable accounts for [1].

The findings.

• Global self-esteem was positively related to aggression, accounting for 11.4% of the variance [1].
• When self-compassion was added, it made a unique contribution of 4.3% additional variance. The relationship was inverse: higher self-compassion predicted lower aggression [1].
• Defensive self-esteem was unexpectedly found to be related to lower levels of self-esteem [1].
• Narcissistic self-esteem and narcissism were not significant predictors [1].

The author suggested that teaching self-compassionate skills could be a useful component of comprehensive interventions aimed at reducing aggression [1].

Read plainly, the pattern is this: in this group, people who rated themselves more highly tended to report more aggression, and people who were more self-compassionate tended to report less, even after self-esteem was accounted for. That is the basis for the cautious thesis of this article. Self-compassion may be a steadier companion to confidence than self-esteem alone.

How far this evidence can carry

Not very far, and it is worth being exact about why.

• It is one dissertation. A single study, however carefully done, is a starting point rather than a settled answer.
• It is old. It was completed in 2009 [1]. Later work may have confirmed, refined or complicated it.
• The sample is narrow. 181 college students from one university in the Northeastern US, who chose to take part [1]. Your age, culture, job and circumstances may differ.
• Everything was self-reported. People described themselves on questionnaires, which captures how they see themselves rather than what they necessarily do [1].
• The design is correlational. The data show that these things move together, not that self-compassion causes lower aggression or that self-esteem causes higher aggression [1].
• The outcome was aggression, not confidence. Applying the finding to confidence is an extrapolation. It is a reasonable question to ask, but the study did not measure confidence, and nothing in it shows that self-compassion builds it.

The topic is still being explored. A 2025 article by a high-school author, published in an education-psychology lecture-notes series, takes up the mutual influence of self-esteem and self-compassion and the advantages of self-compassion [2]. It is a sign that the question is live, not a body of settled findings, and this article does not draw conclusions from it.

Everything beyond this point is practical suggestion, not evidence.

What self-compassion is not

The most common worry is that being kinder to yourself means lowering your standards, excusing mistakes or going easy on yourself. The research cited here does not test that worry either way, so treat what follows as a working distinction rather than a finding.

A useful test is to ask whether your response to a setback is accurate and steady:

| Response | What it sounds like | What it tends to do |
|---|---|---|
| Harshness | "I always do this. I'm hopeless." | Inflates the error and blurs what actually needs fixing |
| Denial | "It doesn't matter. It wasn't my fault." | Skips the information the mistake contains |
| Accurate and steady | "That went wrong for these reasons. I can fix this part." | Keeps the facts and the next step in view |

Notice that the third response keeps the standard fully intact. It simply removes the commentary about your worth. Whether that improves later performance is not something the sources here address, so the claim is limited to this: it is a different way of responding, and you can observe for yourself what it is like.

If you want a companion piece on the factual side of this, see How to Reframe a Setback Without Pretending It Is Fine.

A practical way to try it: the setback note

Treat this as an experiment, not a proven method. It is designed for a minor setback, because low stakes make it easier to notice what changes.

The exercise. After one small setback this week, such as a missed deadline, a clumsy email or a workout you skipped, write three lines:

1. What happened, in neutral facts. Only what a camera or a record would show. No adjectives about yourself.
2. What I would say to a capable colleague who had this happen. Write the actual sentence. Keep it honest; it should not pretend the mistake did not occur.
3. The next useful step, small enough to do today. Ten minutes or less if possible.

Then do the step, start your next task, and afterwards write one sentence on how steady you felt starting it.

A few ways to make it informative:

• Run it two or three times over a couple of weeks and keep the one-line steadiness notes together. A single entry tells you very little; a handful may show a pattern.
• Compare against a setback where you did not write anything, if you can recall how the next task felt.
• If line 2 is hard to write, that is itself information. Notice whether you hold yourself to a harsher rule than you would apply to a colleague.

You may notice no difference, and that is a legitimate result. Habits like this vary a great deal between people, and there is no guaranteed timeline. If you already keep a regular review, the one-line notes fit naturally into A 20-Minute Weekly Review That Keeps You on Course.

A steadier way to frame confidence

Put simply: self-esteem is a reading, and self-compassion is a response. The two can coexist, and a stable sense of confidence does not require choosing between them.

What the available evidence supports is modest. In one 2009 sample of 181 college students, higher self-esteem went with more self-reported aggression and higher self-compassion went with less, with self-compassion adding 4.3% of explained variance beyond self-esteem's 11.4% [1]. That is correlational, self-reported, dated and about aggression, not confidence. The sensible reading is a hypothesis: confidence that rests partly on how you treat yourself after a setback may hold up better than confidence that rests only on how highly you rate yourself.

You do not need to settle the research to test it. Pick one small setback this week, write the three lines, and note how steady you felt on the next task. Keep your standards where they are. Change only the way you talk to yourself about missing them, and see what you observe.