Positive Mindset and Happiness: Gratitude, Mindfulness and Hope

Gratitude: real, but small

The largest review so far, published in 2025, pooled 145 papers with nearly 25,000 people in 28 countries. Gratitude exercises raised wellbeing by a small amount on average. Effects varied between countries, and in several (including Japan, India and the UK) they were not statistically clear.

An earlier review from 2016 asked a sharper question: compared with what? Gratitude beat doing nothing and beat neutral activities such as listing daily events. But it did no better than other psychologically active exercises. In other words, gratitude works about as well as other good exercises, not better.

For depression and anxiety specifically, a 2021 review of 27 studies (about 3,700 people) found a small effect that shrank when gratitude was compared with a real alternative activity. The authors concluded that people seeking relief from depression or anxiety should look first to approaches with stronger evidence.

Three Good Things: why it works may surprise you

In 2005, a widely cited online experiment found that "Three Good Things" increased happiness and reduced depressive symptoms for months. In 2012, researchers repeated it with a better control group, including a "positive placebo" in which people simply thought about positive things about themselves. The exercise raised happiness, but so did the placebo, and it did not beat the control on depression. The likely lesson: much of the benefit comes from regularly turning attention to positive, personal material. That is still useful. It just means there is no special magic in the exact wording.

Mindfulness: helps adults, mixed in schools

A 2021 review of 136 trials (over 11,600 adults in everyday settings such as workplaces and universities) found that mindfulness courses reduced anxiety, depression and stress compared with doing nothing, by a small to moderate amount. Against other active programmes, mindfulness was generally not better. A 2023 analysis using the raw data from 13 trials confirmed a reliable reduction in distress.

Schools are a different story. The MYRIAD trial (2022) gave mindfulness lessons in 42 UK secondary schools and compared them with 43 schools teaching usual social and emotional lessons, with about 8,400 pupils aged 11 to 14 in the analysis. After one year there was no benefit. For pupils already at risk of mental health problems, scores were slightly worse, though the difference was small.

Newer analyses point the same way. A second large English schools trial, INSPIRE, found that mindfulness made no difference overall. When enough of it was delivered, it seemed to help secondary pupils but made younger primary pupils' low mood slightly worse. And when researchers used computer models on MYRIAD data to find which children benefit from school mindfulness, even the best-matched pupils gained almost nothing.

Short practices and breathing

A review of 65 brief mindfulness trials found a small effect on negative mood, which nearly disappeared after correcting for unpublished null results. A 2023 trial found that five minutes a day of slow breathing with long exhales improved mood slightly more than five minutes of mindfulness meditation.

Hope

A 2011 review of 27 studies found that hope-building programmes raised hope and life satisfaction a little but did not reduce distress, with signs that negative studies went unpublished. Long-term studies of teenagers link falling hope with later depression, but these are correlations.

A 2026 review of 99 long-term studies found that people with more optimism and life satisfaction tend to live longer, even when researchers allow for depression, so feeling good seems to be more than just not feeling bad.

If low mood is severe, lasts for weeks, or comes with thoughts of self-harm, talk to a doctor or a trusted adult.

What this means for you

These habits are cheap, safe for most people and modestly helpful. Expect a small lift, not a transformation. Pick the practice you will actually keep doing.

Gratitude

Choi et al. (2025; preregistered; 145 papers, 163 samples, 727 effect sizes, N = 24,804, 28 countries) estimated g = 0.19 (95% CI 0.15 to 0.22) on wellbeing, robust to publication-bias checks. Effects were larger for positive-affect outcomes, combined gratitude activities and randomised designs. Between-country heterogeneity was significant but unexplained; effects were non-significant in France, India, Japan, the Netherlands and the UK. Davis et al. (2016) showed why comparators matter: d = 0.31 versus measurement-only (k = 5), d = 0.14 versus activity-matched controls (k = 18) and d = -0.03 versus psychologically active comparisons (k = 9). Cregg and Cheavens (2021; k = 27; N = 3,675) found g = -0.29 post-test and -0.23 at follow-up on depression and anxiety combined, larger against waitlists. In youth, a meta-analysis of 16 RCTs of "wise interventions" found the largest post-intervention effect on depressive symptoms for gratitude (g = 0.29), from few studies (Parlak et al., 2025).

Replication. Seligman et al. (2005) reported lasting effects of Three Good Things. Mongrain and Anselmo-Matthews (2012), with an improved control and a positive placebo, replicated the happiness gain but found it matched by the placebo, and found no advantage over control on depression. This points to a common-factor mechanism.

Funding. The John Templeton Foundation funded the Greater Good Science Center's "Expanding the Science and Practice of Gratitude" project (from 2011, nearly $4 million distributed in research grants, with Robert Emmons as lead collaborator). A large share of the field has grown with foundation support that favours positive findings about virtues; this does not invalidate results, but preregistered syntheses such as Choi et al. are the better guide.

Mindfulness in adults

Galante et al. (2021; 136 RCTs; N = 11,605) found versus no intervention: anxiety SMD -0.56, depression -0.53, distress -0.45, wellbeing +0.33, with wide prediction intervals crossing zero. Versus specific active controls, no superiority. GRADE certainty was moderate to very low, and only distress survived exclusion of high-risk trials. Galante et al.'s (2023) individual participant data meta-analysis (13 trials; N = 2,371) found distress SMD -0.32 with high confidence and no moderation by baseline distress, gender, age or education.

Apps and industry. A definitive NHS staff trial (N = 2,182) found Headspace reduced stress more than an active control, with small effects (Taylor et al., 2022). Mindfulness apps for students showed small effects on depression and anxiety (LaMontagne et al., 2024). Conflict of interest: some app trials are run by company staff; for example, seven of nine authors of Kirk et al. (2023) were Headspace employees. Trial quality in mental-health apps has improved little over time; fewer than 35% of trials report adverse events (Linardon et al., 2025).

Mindfulness in young people

Dunning et al. (2019; 33 RCTs) found small effects (d 0.16 to 0.30); the 2022 update (66 RCTs; N = 20,138) found d 0.12 to 0.35 versus passive controls, only anxiety/stress (d = 0.11) and mindfulness versus active controls, and no significant effects at follow-up. The authors judged the evidence low quality. MYRIAD (Kuyken et al., 2022; 84 schools; N = 8,376; Wellcome-funded) found SMDs near zero at one year (depression risk 0.005; wellbeing 0.02). Secondary analyses found small harms for at-risk pupils and worse functioning with higher dose (Montero-Marin et al., 2022).

A complier average causal effect analysis of the INSPIRE cluster RCT (about 14,400 pupils) found null intention-to-treat effects, but dose-dependent reductions in depressive symptoms for secondary-school mindfulness and increases for primary-school mindfulness (d = 0.32 to 0.48), suggesting possible harm at younger ages; CACE estimates rest on strong assumptions (Thornton et al., 2026). In MYRIAD (N = 8,376), causal forest and elastic net models identified adolescents predicted to benefit, but the realised differential benefit was negligible (d = 0.07 to 0.08), limiting the case for personalised universal delivery (Webb et al., 2026).

Brief practice and breathing

Schumer et al. (2018; 65 RCTs; N = 5,489) found g = 0.21 for negative affectivity, falling to g = 0.04 after adjustment for publication bias. Balban et al. (2023) found daily five-minute cyclic sighing improved positive affect more than mindfulness meditation over one month; it is a single remote trial and needs replication.

Adverse effects

Farias et al. (2020; 83 studies; 6,703 participants) estimated adverse-event prevalence at 8.3% overall (3.7% in experimental studies; 33.2% in observational), most often anxiety and depression, including in people with no prior history. Using a structured interview after mindfulness-based cognitive therapy, Britton et al. (2021) found lasting negative effects in 6 to 14% of participants, linked to hyperarousal and dissociation; childhood trauma predicted adverse effects (Canby et al., 2025). These are clinical samples, but they argue against presenting meditation as risk-free.

Hope and optimism

Snyder's model defines hope as goals plus agency and pathways thinking (Ritschel & Cassiello-Robbins, 2023). Weis and Speridakos (2011; 27 studies; N = 2,154) found d = 0.22 for hope, 0.16 for life satisfaction and 0.04 (non-significant) for distress, with evidence of publication bias. Hope is inversely related to depression cross-sectionally, but longitudinal findings are more complex (Ritschel & Cassiello-Robbins, 2023). In 2,823 rural Chinese adolescents followed for three years, declining hope trajectories mediated the link between economic stress and later depression (Wang & Zhao, 2025); this is observational. Optimism predicts lower cardiovascular events (RR 0.65) and all-cause mortality (RR 0.86) across 15 cohorts (N = 229,391), with high heterogeneity and residual confounding possible (Rozanski et al., 2019).

Across 99 prospective studies controlling for ill-being, psychological wellbeing (including optimism) predicted lower all-cause mortality (HR = 0.86, 95% CI 0.84 to 0.88; prediction interval 0.72 to 1.04), with similar estimates in unadjusted studies, supporting wellbeing as distinct from absence of distress, though causality is untested (Genecov & Yeager, 2026).

Forced positivity

Repeating positive self-statements made people with low self-esteem feel worse (Wood et al., 2009). Habitual acceptance of negative emotions predicted better psychological health six months later (Ford et al., 2018). Both support teaching positivity alongside acceptance, not instead of it.

Claims we deliberately do not make

That gratitude treats depression; that Three Good Things works through a specific mechanism beyond positive self-focus; that school mindfulness programmes improve pupil mental health; that meditation is harmless for everyone; that hope interventions reduce distress; that optimism causes longer life; that positive thinking should replace accepting difficult feelings; that these practices change brain chemistry.