The Science of Self-Compassion: Is It Really More Stable Than Self-Esteem?

When a project goes wrong, a job application is rejected, or our best effort falls short, the first thing that often comes to mind is not a reasoned analysis but harsh self-criticism.

For decades, “boosting self-esteem” was often treated as the standard answer for mental health. In recent years, research on “self-compassion,” advanced by psychologist Kristin Neff and others, has gained traction. It is often promoted as a “more stable, less easily shaken source of self-worth” than self-esteem.

But if we stop being hard on ourselves, will we become complacent? Is self-compassion really better than self-esteem across the board?

Clarifying the Concepts: Evaluating Yourself vs Responding to Pain

To understand the scientific distinction, we need to start with the operational definitions: what questions do these concepts actually ask, and how are they measured?

• Self-esteem answers an evaluative question: “How good am I, and am I worthy?” It rests on positive judgments of our abilities, appearance, or achievements.
• Self-compassion answers a question about our response: “How do I treat myself when I am hurting, make a mistake, or fall short?”

In 2003, Neff described self-compassion in terms of three pairs of opposing dimensions:

1. Self-kindness vs self-judgment: Offering ourselves understanding and support during setbacks instead of attacking and shaming ourselves.
2. Common humanity vs isolation: Seeing imperfection and suffering as part of the human experience instead of believing “I’m the only one this inadequate.”
3. Mindfulness vs over-identification: Recognizing pain and emotion as they are, without suppressing or denying them and without being completely consumed by negative feelings.

These three pairs also matter when interpreting the scale: the total self-compassion score includes less self-judgment, isolation, and over-identification. So when it is associated with less anxiety and depression, we cannot immediately conclude that the entire association comes from “being kinder to ourselves.” A meta-analysis also found that the negative dimensions were more strongly associated with psychological symptoms than the positive dimensions.

Self-esteem usually operates on the premise that “I’m doing well.” Self-compassion has a crucial advantage: it can start working even before a positive evaluation has been earned.

The term “self-compassion” also needs clarification. It does not mean self-pity, nor does it require covering up mistakes with positive words. Here, “compassion” means recognizing pain, refusing to add to it through shame, and considering how to reduce harm. Reducing it to “being a little nicer to yourself” misses both common humanity and the need to face problems honestly.

The Evidence for “Greater Stability”: Less Fluctuation in Self-Worth

Popular psychology often reduces “self-compassion is more stable than self-esteem” to “self-compassion scores never go down.” But longitudinal studies that follow the same people over time do not support that reading.

A long-term study that followed 3,180 people for 29 years found that overall self-esteem itself has considerable long-term trait stability. It is not something that shatters at the slightest touch.

So where does the evidence for “greater stability” come from? The classic example is Neff and Vonk’s (2009) study of a Dutch community sample over eight months. The researchers measured participants’ momentary sense of self-worth (state self-esteem) nine times and found that:

• After controlling for overall self-esteem, people with greater self-compassion showed significantly less fluctuation in momentary self-worth (partial r = −.23; “significantly” is a statistical term here, meaning the difference was unlikely to be due to sampling error). In other words, the study observed less variation in self-evaluation across repeated measurements, not immediate emotional recovery after every setback.
• People with greater self-compassion were less likely to base their self-worth on external conditions, such as appearance, others’ approval, or immediate success and failure (contingencies of self-worth).
• When things went poorly, they showed markedly lower tendencies toward social comparison (ranking themselves against others), rumination (repeatedly replaying setbacks and negative feelings), and defensive anger.

But this was not a direct comparison of the stability of two traits. The researchers repeatedly measured state self-esteem; they measured self-compassion and overall self-esteem only in the final wave. They did not repeatedly measure “state self-compassion” over the eight months, and the final scale scores cannot establish that self-compassion caused the lower fluctuation observed earlier.

The central conclusion supported by the evidence is: “People with greater self-compassion experience less fluctuation in self-worth,” not “self-compassion itself is a more stable trait than self-esteem.”

Moving Beyond the Myth: Two Systems That Complement Each Other

A meta-analysis of 76 studies and more than 35,000 people (Muris & Otgaar, 2023) found that self-compassion and self-esteem are strongly correlated (r = .65). Even when both are considered together, each independently predicts mental health. The simple claim that self-compassion is superior to self-esteem across the board does not hold up. The researchers also note that part of the correlation comes from overlap in concept and measurement: the two main scales both include reverse-scored negative items, such as self-criticism, isolation, and rumination.

A four-year study of adolescents further found that healthy self-esteem often provides a foundation for developing self-compassion. When self-esteem takes a hit, self-compassion can buffer the negative effects of low self-esteem on mental health.

Many people worry: “If I stop scolding myself, will I lose the motivation to improve?”

Four experiments by Breines and Chen (2012) tested this concern directly. Compared with a self-esteem boost group or a control group, participants guided to respond with self-compassion after an exam failure or a moral transgression showed stronger motivation to improve: they were more willing to spend time studying and more likely to take action to make amends.

Self-compassion is not an excuse for failure. It may reduce the additional pain of self-shaming, leaving more room to face reality and act.

These four experiments, however, primarily measured short-term motivation to improve. They cannot directly establish effects on long-term performance. Whether practice can improve one’s mental state is a separate question. A meta-analysis of randomized controlled trials (Han & Kim, 2023) found small to moderate improvements in depressive symptoms, anxiety, and stress at the end of self-compassion-related interventions. Follow-up findings were more limited, and the studies had a high overall risk of bias.

These programs often combine mindfulness practice, guidance, and support. Their full effects cannot be credited to the single instruction “be kind to yourself,” much less can that instruction guarantee that the five steps below will produce equivalent results.

A Five-Step Practice for Setbacks

When a project goes wrong or you face a setback, consider these five practices:

1. Name the specific event accurately: Replace a sweeping judgment of your character with the facts. Instead of “I’m an incompetent engineer,” say, “An unexpected bug appeared during this deployment.” Compassion begins with seeing reality clearly, not avoiding responsibility.
2. Stop the shaming, keep the standards: Turn the inner rebuke “How could you be so stupid?” into constructive questions: “Which part of the decision did I miss? What safeguard should I put in place next time?” The goal is to get better information for debugging, not to punish yourself.
3. Remember our common humanity: Remind yourself that mistakes, setbacks, and limits to our abilities are part of everyone’s growth. You are not the only person who has messed something up.
4. Take the smallest workable action: Once you have steadied your emotions, ask: “What is the first step I can take to make this right?” Whether it is proactively sharing an update, writing a fix, asking a colleague for help, or getting a good night’s sleep first, choose one thing you can do.
5. Allow an emotional pause without forcing positivity: For some people, practicing self-compassion can bring up intense sadness or a defensive reaction, sometimes called “backdraft.” There is no need to force yourself to “feel good.” You can shorten or pause the practice. If it brings up overwhelming traumatic memories or symptoms, seek appropriate mental health support.

The Limits of Self-Compassion

Cultural context also shapes how self-compassion is understood and practiced. A comparative study of samples from the United States, Thailand, and Taiwan found that average self-compassion scale scores were highest in Thailand, lowest in Taiwan, with the United States in between.

This is not grounds for the sweeping conclusion that “Taiwanese people are less able to treat themselves kindly.” The study compared scale responses from specific samples, and understandings of self-criticism, humility, and “compassion” may differ across settings. At a minimum, it reminds us not to assume that readers across Asia will understand or practice the concept in the same way simply because it has Asian roots.

• Compassion for yourself is not the same as compassion for others: A meta-analysis (Chio et al., 2026) found only a weak correlation between self-compassion and empathic concern for others (r = .19). Caring for yourself and fulfilling your responsibilities to others both require practice; neither replaces the other.
• It cannot replace systemic improvements: Self-compassion can ease personal self-criticism under intense pressure, unreasonable performance reviews, or workplace bullying. But it cannot justify an unreasonable environment. Companies cannot use “teaching employees self-care” to obscure structural problems such as excessive working hours and inadequate resources.