Self-compassion: what it is and what the evidence says

Short answer

Self-compassion means treating yourself with the kindness you would offer a friend when things go wrong. Across 20 pooled samples it showed a large negative association with depression, anxiety and stress [C1]. Trials of compassion-based training report small to medium reductions in self-criticism and distress, though most compare with waitlist rather than active controls [C5][C6].

What it is and how it shows up

Self-compassion is a way of relating to your own pain: noticing it, recognising that struggle is part of being human, and offering yourself warmth instead of punishment. The most-used instrument, the Self-Compassion Scale, was built around three key components of self-kindness, common humanity and mindfulness, while its other half of the items are negative indicators that reflect the opposite of those components [C4].

In everyday life, self-compassion is what changes the tone after a mistake at work, a rejected application or a hard conversation: instead of the internal commentary that spirals into "I always mess this up", the response leans toward "this was hard, and hard things happen to people". This is what separates it from self-esteem, which depends on evaluating yourself positively; self-compassion does not require any verdict on your worth, only a stance of warmth toward your own difficulty.

Researchers also treat self-compassion as the counterpart of self-criticism, the process of harsh negative self-evaluation. High self-criticism is associated with psychopathology and poorer therapy outcomes, and self-compassion interventions were developed to target it directly [C6].

Self-criticism and self-compassion are studied as two sides of the same relationship. Harsh self-evaluation is consistently associated with psychopathology, and self-compassion interventions were developed to target it directly; pooled trials report reductions in self-criticism after training [C6]. In the eating-related evidence, self-criticism showed a medium positive link with disordered eating, while self-compassion showed preliminary negative relationships [C9].

What research shows by approach

The sections below group the available studies by the question they examined. As elsewhere on this site, the picture is deliberately specific: effects describe the populations and formats actually studied, and nothing here generalises automatically to any individual reader.

Links to mental health and well-being

The anchor finding comes from a meta-analysis of 20 samples from 14 studies, all using the Neff Self-Compassion Scale: self-compassion showed a large negative association with psychopathology, r = -0.54 [C1]. A correlation of this kind cannot settle direction: low self-compassion may deepen symptoms, symptoms may erode self-compassion, or both [C1].

A separate meta-analysis of 79 samples with 16,416 participants examined well-being specifically and found a correlation of r = .47, with the relationship stronger for cognitive and psychological well-being than for affective well-being [C2]. A subsample of experimental studies pointed in the same causal direction [C2].

The measurement debate matters for reading these numbers. The strength of the pooled association depends on how the self-compassion scale is scored: when positive and negative items are examined separately, the negative indicators track mental health problems more strongly than the positive ones, and part of the pooled correlation may reflect this overlap [C4].

Self-criticism and a measurement debate

Two findings nuance the field. A meta-analysis of interventions found a medium reduction in self-criticism after self-compassion training, g = 0.51, with greater reductions for longer programmes and against passive controls rather than active ones [C6]. At the same time, a meta-analysis of 18 studies on the scale itself found that the negative indicators were linked to mental health problems more strongly than the positive ones, and suggested that a total self-compassion score can inflate the association with symptoms [C4]. In plain terms: part of the correlation between self-compassion and symptoms may reflect harsh self-criticism being measured inside the same questionnaire [C4].

Training programmes and compassion-focused therapy

The intervention evidence is broad but uneven. A meta-analysis of fifty-six randomised controlled trials found that self-compassion focused interventions had small to medium effects on depressive symptoms, anxiety and stress right after the programme, and small effects on depressive symptoms and stress at follow-up [C5]. The overall risk of bias across those trials was high, and few of them compared self-compassion training with an active treatment [C5].

For clinical populations specifically, a review of compassion focused therapy included fifteen studies from 2013 to 2022 and reported improvements in compassion-based outcomes and clinical symptomology, with a range of small to large effect sizes across self-compassion, self-criticism, depression and eating disorders [C7]. The methodological quality of many of these studies was unclear, and long-term effects remain underexplored [C7].

What the training looks like varies: protocols range from brief written self-compassion exercises to multi-week courses and full compassion-focused therapy in clinical settings [C5]. The widest pooled evidence covers programmes inspired by mindful self-compassion training, and reductions in self-criticism were greater for longer formats compared with passive controls [C6]. For clinical populations, compassion-focused therapy reports improvements in compassion-based outcomes alongside symptom measures [C7].

Physical health and chronic illness

Self-compassion is not only about mood. In a large pooled sample across 94 peer-reviewed articles, it showed positive associations with physical health and health-promoting behaviour, and multi-session interventions designed to boost self-compassion predicted increased physical health and health behaviour, which the authors read as support for causal links [C3]. Single-session inductions, by contrast, did not move the outcomes [C3].

In chronic illness, a meta-analysis of 51 studies found self-compassion negatively associated with psychological distress, r = -.516, with the association largest for stress and neurological conditions [C8]. Whether raising self-compassion reduces that distress over time is a separate question the pooled studies do not answer [C8].

Not every intervention format moved the outcomes: in the pooled analysis, the effects of single-session inductions were non-significant, which suggests that brief one-off exercises are not the same as multi-session training [C3]. For people living with chronic conditions, the association between self-compassion and distress was largest for stress and for neurological conditions [C8].

Eating, body image and self-criticism

A meta-analysis of 135 studies with 42,952 participants found a medium positive link between self-criticism and disordered eating, r = .37, and preliminary negative relationships between self-compassion and disordered eating [C9]. The authors' conclusion is practical: treatments for disordered eating may need to tackle self-criticism and nurture self-compassion rather than address symptoms alone [C9].

Limitations and uncertainty

Several boundaries deserve emphasis. Most of the evidence is correlational, so the direction between self-compassion and mental health remains open [C1]. The field's main questionnaire mixes positive items with reverse-scored negative ones, and meta-analytic work suggests the total score can inflate the association with symptoms [C4]. Trials are often small, differ in format, dose and population, frequently use waitlist or passive comparisons, and the widest intervention meta-analysis rated the overall risk of bias as high [C5][C6][C7]. Effects on health behaviour and chronic illness distress come mostly from pooled associations rather than dedicated trials [C3][C8]. Across all of this, findings describe groups of participants in studies, not predictions for any individual.

When to see a professional

Relentless self-criticism, low mood or anxiety that persists for weeks, disordered eating, or distress tied to a chronic condition all warrant more than self-help exercises. Compassion-focused therapy in trials was delivered by trained therapists in structured programmes, and the clearest self-criticism findings come from guided interventions rather than improvised practice [C6][C7]. If self-criticism accompanies thoughts of self-harm, severe restriction of food, or an inability to function, the right step is an assessment by a qualified clinician or contact with local emergency services. Costs, session frequency and availability differ between countries and health systems, so practical questions are worth raising with a professional before starting.

FAQ

What is self-compassion?

A stance of warmth toward your own difficulty, built around three key components of self-kindness, common humanity and mindfulness — noticing pain rather than punishing yourself for it [C4].

Is self-compassion the same as self-esteem?

No. Self-esteem is a positive evaluation of your worth; self-compassion does not require judging yourself at all. In meta-analysis, self-compassion tracked cognitive and psychological well-being even with self-esteem tested as a potential moderator [C2]. In pooled data the relationship with well-being remained after testing self-esteem as a moderator [C2].

Can self-compassion be trained?

Yes, that is the point of the interventions. Randomised trials of self-compassion training have reduced self-criticism and depressive symptoms, anxiety and stress, with effects that are mostly small to medium and better studied against waitlist than against active treatments [C5][C6]. The pooled trials cover group courses, individual therapy and self-guided formats, with effects mostly measured right after the programme [C5].

Does self-compassion matter for physical health?

It is linked to physical health and health-promoting behaviour in pooled data, and multi-session training programmes predicted health improvements, while the association with distress in chronic illness is largest for stress and neurological conditions [C3][C8].

Self-help information does not replace a clinician; people with severe or persistent symptoms, or thoughts of self-harm, should contact a professional or their local emergency services.