Compassion and loving-kindness training
Reviewed November 2026
Research strength — Moderate for reducing depression, anxiety, stress and self-criticism against waitlist controls, Preliminary against active psychological treatments and on real-world prosocial behaviour.
Traditional convergence — Very strong (systematic training of goodwill and compassion toward oneself, strangers and difficult people appears independently in Buddhist, Yogic, Confucian and Christian traditions for over two thousand years).
In short: structured practices that train goodwill and self-compassion produce moderate improvements in depression, anxiety, stress and self-criticism across dozens of randomised trials.1 They also alter how people respond when they make mistakes or witness someone else being treated unfairly.2 The catch sits in how those trials are controlled: gains shrink sharply when measured against active comparison groups rather than waitlists,3 and a critical meta-analysis of prosocial meditation outcomes found compassion gains disappeared when studies used active controls or independent researchers without the meditation teacher as an author.4 For people with trauma or severe shame, warmth exercises can also trigger threat and distress rather than immediate comfort.5
What it is
Deliberately cultivating an attitude of goodwill toward ease and happiness (loving-kindness) and care toward suffering (compassion) — directed toward yourself, loved ones, strangers and difficult people.6
In practice and research, it appears in three main forms:
- Loving-kindness and compassion meditation — silently repeating simple wishes of safety, ease and freedom from distress while bringing different people to mind in sequence.6
- Self-compassion training — noticing your own struggle without harsh self-attack, treating difficulty as shared human experience, and speaking to yourself as you would to a friend.1
- Compassion-focused therapy — a clinical psychotherapy combining imagery, soothing breathing and cognitive work to quiet shame and self-criticism.7
It is a skill of intention and perspective rather than trying to force warm feelings on demand.6
What it does for you
It lowers depression, anxiety, stress and self-criticism moderately
Three independent meta-analyses converge on the same middle range against control groups:
| Meta-analysis (trials) | Depression | Anxiety | Stress / distress | Self-criticism |
|---|---|---|---|---|
| Compassion-based interventions (21 RCTs, 1,285 people)8 | d = 0.64 | d = 0.49 | d = 0.47 | — |
| Self-compassion interventions (27 RCTs)1 | g = 0.66 | g = 0.57 | g = 0.67 | g = 0.56 |
| Compassion-focused therapy7 | g = 0.49 (k = 23) | — | — | g = 0.40 (k = 17) |
In the 27-trial self-compassion analysis, larger average gains also appeared on rumination (g = 1.37) and eating behaviour (g = 1.76), and improvements in depression continued to grow during one-to-three-month follow-ups.1
Loving-kindness raises daily positive emotions
Across 24 empirical studies (1,759 participants), loving-kindness meditation produced medium increases in daily positive emotions in both waitlist-controlled trials and non-randomised studies, and small-to-large increases in immediate positive emotion during practice.9 Interventions focused specifically on loving-kindness produced larger positive-emotion gains (medium) than those focused only on suffering-focused compassion (small).9
Across 22 randomised trials of kindness-based meditation broadly, positive emotions also rose significantly against progressive muscle relaxation (g = 0.42).6
Short-term compassion training increases costly helping in the lab
In a controlled trial, healthy adults who completed two weeks of 30-minute daily compassion training gave more of their own money to compensate an unfairly treated victim in a subsequent game than an active control group trained in cognitive reappraisal.2 That behavioural shift tracked with changes in brain activity when viewing images of human suffering.2
How much you need to do
- Eight-week group programmes (weekly classes plus daily home practice, such as Mindful Self-Compassion or Compassion Cultivation Training) are the format behind most clinical and community trial data.8
- Two weeks at 30 minutes a day of guided audio was enough to shift neural responses and laboratory helping behaviour in healthy adults.2
- More hours did not guarantee larger gains across positive-emotion trials: neither total programme length in weeks nor weekly practice time significantly predicted effect size across studies reporting duration.9
So: 15 to 30 minutes of guided practice most days for two to eight weeks is the range tested in trials. Because duration has not shown a clear dose-response curve, steady consistency matters more than marathon sittings.9
Practices that rest on it
Holding Love Steady · Loving-Kindness · Giving and Receiving · Compassionate Presence
Where the evidence stops
Gains narrow considerably against active controls. Across 56 randomised trials of self-compassion interventions, effects were significantly smaller whenever interventions were compared against active comparison treatments rather than passive waitlists, and most trials still lacked active controls and carried high risk of bias.3 In the 22-trial review of kindness-based meditation, results against active controls were inconclusive across several outcomes.6
Prosocial claims depend heavily on study quality and investigator involvement. A systematic review and meta-analysis specifically testing whether meditation makes people kinder toward others found moderate gains in compassion and empathy, no significant effect on aggression, social connectedness or prejudice, and one sharp warning: compassion increases appeared only in waitlist-controlled studies and only when the meditation teacher was a co-author on the paper.4
Publication bias is present in the clinical literature. Both the 27-trial self-compassion review1 and the compassion-focused therapy meta-analysis7 flagged publication bias — meaning unpublished null trials likely pull real-world averages lower than the published figures above.
Long-term durability is thinly tested. Outside the one-to-three-month follow-up subset in self-compassion trials,1 longer-term effects of compassion-focused therapy7 and loving-kindness meditation remain largely unestablished, and most trials rely on small samples and self-report scales.8
Unknowns: which specific element inside multi-week programmes — soothing breathing, imagery, cognitive reframing or group support — accounts for most of the benefit;7 whether daily life helping changes outside laboratory economic games;4 and why loving-kindness outperformed compassion practice on positive emotions.9
And one thing to actively disbelieve: that repeating goodwill phrases will automatically flood you with warm feelings, or that feeling blank or irritated means you are doing it wrong. Practice often feels mechanical at first, and clinical gains do not require an immediate emotional rush.6
Who should be careful
Fear of compassion and "backdraft". For people with histories of neglect, trauma or intense self-criticism, trying to generate kindness toward themselves or receive it from others can register as threatening rather than comforting.5 A meta-analysis of 19 studies (4,723 participants) found strong associations between fears of self-compassion (r = .49) and fears of receiving compassion (r = .48) on one side and shame, self-criticism and depression on the other — even stronger in clinical samples.5 Exposure to kindness-based meditation can initially feel challenging or distressing for some beginners.6
The adjustments below are clinical and practical judgements rather than trial-tested safety rules, and are marked as such.
Start outside yourself if self-directed kindness triggers distress. If wishing yourself well brings up shame, tightness or harsh rejection, skip yourself at first and begin with a pet, a neutral acquaintance or someone easy to wish well.
Do not start with people who harmed you. Moving straight to wishing goodwill toward an abusive or actively harmful person can feel invalidating or re-traumatising. Leave difficult relationships out of early practice, or work with a qualified therapist where trauma is involved.
Severe depression and clinical self-hatred. Self-guided audio is not a substitute for psychotherapy when shame and self-criticism are driving clinical depression or self-harm; use structured clinical care first.
Traditional convergence
Very strong. Four independent traditions concluded over two millennia ago that goodwill and compassion are not mere personality traits you either have or lack, but dispositions that can be trained systematically — and all four tied mental training directly to daily conduct.
In early Buddhism, the Karaṇīya Mettā Sutta instructs practitioners to cultivate boundless goodwill (mettā) in all directions "even as a mother protects with her life her child, her only child," embedded directly inside upright speech and conduct; Buddhaghosa's Visuddhimagga later formalised this into a step-by-step sequence moving from oneself to a dear friend, a neutral person and a hostile person.10 In Hindu yoga, Patañjali's Yoga Sūtras prescribe four deliberate attitudes to settle a disturbed mind — friendliness (maitrī) toward the happy, compassion (karuṇā) toward the suffering, appreciative joy (muditā) toward the virtuous and equanimity (upekṣā) toward those acting harmfully.10 In China, Mencius argued from the spontaneous alarm any bystander feels seeing a child about to fall into a well that compassion is already present as an innate sprout, meant to be expanded outward from one's own family to strangers.10 And in the Christian Gospels, Jesus directs listeners to "love your enemies, do good to those who hate you, bless those who curse you, pray for those who mistreat you," treating deliberate goodwill toward adversaries as the decisive step beyond ordinary tribal affection.10
You can test the practical shift directly without adopting any doctrine: spend ten minutes wishing one neutral stranger — someone you passed in a shop or on the street today — simple safety and ease, and notice how it changes your tension and judgement during the rest of the day.
References
APA 7th.
Footnotes
-
Ferrari, M., Hunt, C., Harrysunker, A., Abbott, M. J., Beath, A. P., & Einstein, D. A. (2019). Self-compassion interventions and psychosocial outcomes: A meta-analysis of RCTs. Mindfulness, 10(8), 1455–1473. https://doi.org/10.1007/s12671-019-01134-6 ↩ ↩2 ↩3 ↩4 ↩5 ↩6
-
Weng, H. Y., Fox, A. S., Shackman, A. J., Stodola, D. E., Caldwell, J. Z. K., Olson, M. C., Rogers, G. M., & Davidson, R. J. (2013). Compassion training alters altruism and neural responses to suffering. Psychological Science, 24(7), 1171–1180. https://doi.org/10.1177/0956797612469537 ↩ ↩2 ↩3 ↩4
-
Han, A., & Kim, T. H. (2023). Effects of self-compassion interventions on reducing depressive symptoms, anxiety, and stress: A meta-analysis. Mindfulness, 14(7), 1553–1581. https://doi.org/10.1007/s12671-023-02148-x ↩ ↩2
-
Kreplin, U., Farias, M., & Brazil, I. A. (2018). The limited prosocial effects of meditation: A systematic review and meta-analysis. Scientific Reports, 8, 2403. https://doi.org/10.1038/s41598-018-20299-z ↩ ↩2 ↩3
-
Kirby, J. N., Day, J., & Sagar, V. (2019). The "flow" of compassion: A meta-analysis of the fears of compassion scales and psychological functioning. Clinical Psychology Review, 70, 26–39. https://doi.org/10.1016/j.cpr.2019.03.001 ↩ ↩2 ↩3
-
Galante, J., Galante, I., Bekkers, M.-J., & Gallacher, J. (2014). Effect of kindness-based meditation on health and well-being: A systematic review and meta-analysis. Journal of Consulting and Clinical Psychology, 82(6), 1101–1114. https://doi.org/10.1037/a0037249 ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7
-
Petrocchi, N., Ottaviani, C., Cheli, S., Matos, M., Baldi, B., Basran, J. K., & Gilbert, P. (2024). The impact of compassion-focused therapy on positive and negative mental health outcomes: Results of a series of meta-analyses. Clinical Psychology: Science and Practice, 31(2), 230–247. https://doi.org/10.1037/cps0000193 ↩ ↩2 ↩3 ↩4 ↩5
-
Kirby, J. N., Tellegen, C. L., & Steindl, S. R. (2017). A meta-analysis of compassion-based interventions: Current state of knowledge and future directions. Behavior Therapy, 48(6), 778–792. https://doi.org/10.1016/j.beth.2017.06.003 ↩ ↩2 ↩3
-
Zeng, X., Chiu, C. P. K., Wang, R., Oei, T. P. S., & Leung, F. Y. K. (2015). The effect of loving-kindness meditation on positive emotions: A meta-analytic review. Frontiers in Psychology, 6, 1693. https://doi.org/10.3389/fpsyg.2015.01693 ↩ ↩2 ↩3 ↩4 ↩5
-
Pali Karaṇīya Mettā Sutta (Sutta Nipāta 1.8 / Khuddakapāṭha 9) and Buddhaghosa, Visuddhimagga, ch. 9 (Brahmavihāra-niddesa); Patañjali, Yoga Sūtras, 1.33; Mencius, 2A.6; Gospel of Matthew 5:43–48 and Gospel of Luke 6:27–36. ↩ ↩2 ↩3 ↩4