A meta-analysis of the effectiveness of gratitude interventions on well-being across cultures
Gratitude practices produce a small but real increase in well-being. Pooling 145 papers and 24,804 participants across 28 countries, this preregistered meta-analysis found an effect of Hedges' g = 0.19 - a modest gain, robust against publication bias, and the most honest single number available on whether gratitude journaling works.
The most credible number we have on gratitude journaling is 0.19 - and we would rather quote it than a bigger one.
This 2025 PNAS paper is a preregistered meta-analysis: the authors committed to their method before they saw the results, which is the strongest protection against a study finding whatever it hoped to find. It pools 145 papers, 163 samples, 727 effect sizes and 24,804 participants from 28 countries.
The headline: gratitude interventions led to small overall increases in well-being, Hedges’ g = 0.19, 95% CI [0.15, 0.22]. The confidence interval does not cross zero, and the result held up against publication bias and influential cases.
What a small effect actually means. It means gratitude journaling is worth doing and is not going to rebuild your life. Anyone selling you a bigger number is either citing a single flattering study or not citing anything.
Three method choices moved the effect: it was larger when well-being was measured as positive affect, larger when several kinds of gratitude exercise were combined, and larger in randomised controlled trials.
The cross-cultural finding is the paper’s real news and its open question. Effects genuinely differed between countries, but no tested moderator explained the variation - so the authors call for more primary research rather than offering a story.
g = 0.19 is small. This is not a treatment, and the paper makes no clinical claim. Effects differed significantly between countries and the authors found no moderator explaining why, which they flag as needing more primary research. Effects were larger for positive affect than other outcomes, and larger in randomised trials - a reminder the number varies with how you measure it.
Articles citing this study
Summarized by Tiago Martins · Last updated: · How we write these
