Does Gratitude Journaling Actually Work?

You have probably tried it, or at least thought about it. A notebook by the bed, a list of three things before sleep, the quiet discipline of noticing what is good. Maybe it helped. Maybe it felt forced after a week and you quietly stopped. Maybe you are still not sure whether it did anything at all.

The research has a lot to say about all three of those experiences — and it is more nuanced than the confident advice you will find almost everywhere else.

Gratitude interventions have a small average effect, but that category includes letters, lists, essays, and other exercises — not only journals. Against psychologically active alternatives, the average advantage is much smaller. Individual journal formats also produce different results: two targeted list tasks failed to beat an active control in one large trial, while an unconstrained gratitude list showed short-term benefits on gratitude and positive affect. The evidence is more conditional than a simple yes or no.

None of this makes gratitude journaling a bad idea. It makes it a more interesting one. This article goes through what the evidence actually shows, including the findings that rarely make it into popular summaries.

The practical implications are significant — if you want to skip straight to what the research recommends, jump to What this means in practice.

Key findings at a glance

• The overall effect of gratitude interventions on wellbeing is small — the largest 2025 meta-analysis reported Hedges' g = 0.19 across many intervention formats
• Across gratitude interventions, the average difference from psychologically active alternatives was essentially zero in one meta-analysis (d = −0.03)
• In one preregistered one-week trial of 958 people, two targeted lists did not beat an active control, but an unconstrained list improved gratitude and positive affect at post-test
• In a psychotherapy sample, the gratitude-letter condition used a lower proportion of negative-emotion words than expressive writing; the study did not show a decline across sessions
• Country and participant differences appeared in the broader intervention literature, but they should not be treated as universal rules for gratitude journals
• No study here establishes one minimum duration or optimal frequency for a gratitude journal

What the research actually shows: the numbers

The most comprehensive review of the evidence is Choi, Cha, McCullough, Coles, and Oishi (2025), published in Proceedings of the National Academy of Sciences. It pooled 145 papers, 163 samples, 727 effect sizes, and 24,804 participants from 28 countries. Crucially, it analysed gratitude interventions broadly rather than gratitude journaling alone.

The overall effect on wellbeing was Hedges' g = 0.19, a small average effect. The authors reported that the result was robust to their publication-bias analyses.

To put this in context: a g around 0.2 is conventionally described as small. It summarises the average contrast between participants assigned to varied gratitude interventions and their comparison groups; it is not an effect size for one specific journal routine.

Breaking this down by outcome type makes the picture more specific:

In this broader intervention literature, the average association was larger for positive outcomes than for negative affect. The table should not be read as a journal-format comparison.

The study everyone cites — and what it actually found

These numbers come from aggregating decades of research. The study that started the field is still worth understanding in detail, because it is cited everywhere but rarely described accurately.

Robert Emmons and Michael McCullough published "Counting Blessings Versus Burdens" in the Journal of Personality and Social Psychology in 2003. The paper is the primary source for most popular claims about gratitude journaling. It contains three separate studies, and the results are less uniform than popular accounts suggest.

Study 1 (weekly, 9 weeks, N = 192)

The gratitude group reported life as a whole more favourably and showed about 1.5 extra hours of exercise per week compared to the hassles group. However, there were no significant effects on positive or negative affect — the outcomes most people assume the study demonstrated.

Study 2 (daily, 13 days, N = 157)

The gratitude effect on a composite wellbeing measure was larger (d = 0.88 vs the hassles condition). Positive affect improved significantly, and participants offered more emotional support to others. But the exercise effects from Study 1 did not replicate, nor did any effects on physical symptoms or sleep.

Study 3 (clinical sample, N = 65)

Adults with neuromuscular diseases, daily practice for 21 weeks. This is the most compelling study — positive affect improved (d = 0.56), negative affect decreased (d = −0.51), sleep quality and duration improved, and partner/spouse reports corroborated the self-reports. That second-observer corroboration is unusual and strengthens the findings.

The pattern across three studies: different outcomes appeared in different studies, and the most dramatic effects came from the clinical sample with serious illness. No study measured how long effects lasted after stopping.

The six-month claim — and what it actually means

Martin Seligman's team tested five positive psychology exercises in a large internet-based study (N = 577) with follow-ups at 1 week, 1 month, 3 months, and 6 months. The "Three Good Things" condition — writing three things that went well each day along with a causal explanation — showed improvements in happiness and reductions in depressive symptoms that were still present at 6 months.

This is the origin of the claim that gratitude journaling benefits can persist for months. What the study actually found is more specific: benefits at 6 months were contingent on continued practice. Participants who kept doing the exercise after the study ended showed sustained benefits. Those who stopped largely returned to baseline. It is a one-week introduction to a practice, not a one-week cure.

One important replication note: Mongrain and Anselmo-Matthews (2012, N = 344) replicated the happiness effects but did not replicate the depression effects. They also found that a control condition — writing about early childhood memories — produced similar improvements. This raises the question of whether gratitude specifically, or any structured positive writing, drives the benefit.

The finding most popular articles skip entirely

In 2016, Davis and colleagues published a meta-analysis of gratitude interventions in the Journal of Counseling Psychology that is rarely cited in popular writing. It is especially useful for honest reporting because it distinguishes between different types of control conditions — and the results look very different depending on which comparison you use.

Across gratitude interventions, comparisons with no treatment look more favourable than comparisons with psychologically active alternatives. That pattern does not show that every gratitude-journal format works or that every active comparison is equivalent.

This does not make gratitude journaling useless. It suggests that gratitude activities are among several structured positive practices and may not have a unique advantage. The comparison data do not identify a single active ingredient.

Two further meta-analyses reached the same conclusion. Dickens (2017) found d = 0.03 against positive controls. Kirca and colleagues (2023) found g = 0.06 against bona fide comparisons. Neither was significantly different from zero.

The authors of the Davis et al. analysis concluded that "enthusiasm for gratitude interventions should be tempered." This is not a dismissal of the practice. It is an honest calibration.

Why the Format Can Matter

This is the finding with the most direct practical implications — and the one most at odds with standard advice.

Regan, Walsh, and Lyubomirsky (2023) published a preregistered study of 958 Australian adults in Affective Science. Participants completed one of six conditions each day for one week: five gratitude activities — social letters, nonsocial essays, social lists, nonsocial lists, or unconstrained lists — plus an active control that recorded daily activities. Post-test was at the end of that week, with follow-up one week later; the overall study timeline was 15 days.

The targeted social and nonsocial lists did not differ from the active control on any outcome. But the unconstrained gratitude list — which allowed either people or things — produced greater gratitude and positive affect than the control at post-test. "Lists do not work" is therefore too broad.

Social letters and nonsocial essays produced several benefits relative to the control at post-test, although some nonsocial-essay comparisons weakened after correction for multiple testing. Long-form conditions also outperformed the two targeted list conditions on several post-test outcomes. Many differences were not maintained one week later, and social letters also increased indebtedness — a potential cost as well as a benefit.

The practical implication is narrower: in this one-week trial, targeted lists were less consistently beneficial than letters or essays, while an unconstrained list produced some immediate benefits. A short narrative is a reasonable option if stock lists feel mechanical, but this study does not establish narrative writing as universally superior across schedules or populations.

Researchers have proposed elaborative processing — engaging in depth with a specific experience — as one explanation for differences among formats. The Regan trial compared activities, not mechanisms, so it cannot show that elaboration caused the observed effects.

A counterintuitive finding about how it works

One of the more surprising findings in recent gratitude research concerns the mechanism. If gratitude journaling works by cultivating appreciation, you would expect the key factor to be an increase in positive emotional language over time.

That is not what the evidence shows.

Wong and colleagues (2018) in Psychotherapy Research randomised 293 adults seeking university-based psychotherapy to therapy only, therapy plus expressive writing, or therapy plus gratitude-letter writing. The gratitude group wrote once weekly for three weeks. Mental health was assessed about four and twelve weeks after the writing intervention ended, and the gratitude condition reported better mental health than both comparison groups at those follow-ups.

When the researchers analysed the writing, a lower proportion of negative-emotion words in the gratitude condition, relative to expressive writing, mediated the condition's later mental-health effect.

This was a between-condition language pattern, not evidence that participants steadily removed negative words over successive sessions. It also came from gratitude letters used alongside psychotherapy, so it should not be treated as a proven mechanism for ordinary gratitude journals.

The delayed follow-ups are noteworthy: benefits were observed after the three-week writing intervention had ended. They do not show that participants wrote continuously for 12 weeks, that effects necessarily kept growing, or that every gratitude practice requires at least four weeks.

Most participants in this study did not send their gratitude letters. Sending was optional rather than randomly assigned, so the study cannot show that writing without sending was sufficient or equivalent to sending.

Who May Benefit — and When It Might Not Help

Choi et al. reported larger average effects among participants lower in dispositional gratitude, which is compatible with a ceiling effect. This was a moderator across varied gratitude interventions, not proof that every gratitude-journal format works best for people who feel least grateful. Moderator analyses also describe averages across studies and should not be used to predict an individual's response with certainty.

Beyond that moderator, three other patterns emerge from the research:

People with at least moderate baseline distress benefit more.

A 2025 JMIR study found a significant medium effect only in participants with at least moderate baseline symptoms — no significant effect in the full low-symptom sample. Froh and colleagues found comparable moderation in adolescents.

Older adults tend to benefit more than undergraduates.

Several meta-analyses have found larger effect sizes in older adults. The likely explanation is motivation: undergraduates often participate for course credit with limited engagement, while adults practising by choice are more invested.

Cultural context matters substantially.

The PNAS 2025 meta-analysis found country-level variation in the broader gratitude-intervention literature, with nonsignificant estimates in several countries. The tested cultural variables did not explain that variation after correction for multiple comparisons. This is an unresolved population-level result, not a verdict that gratitude journaling cannot help any individual in those countries.

When gratitude journaling may not be the right tool

Active grief or loss.

Focusing on gratitude during acute grief can feel invalidating or create pressure to feel differently. No writing exercise is a substitute for grief or crisis support; change or stop the practice if it intensifies distress.

Clinical depression.

Cregg and Cheavens (2021) found a small effect of gratitude interventions on depression (g = −0.29) but explicitly stated this would not be something they would recommend as a treatment. For clinical depression, the evidence supports cognitive behavioural therapy and other structured interventions considerably more strongly.

When it feels forced.

If writing gratitude entries feels like performing positivity rather than noticing something you value, changing or stopping the exercise is reasonable. Motivation and engagement are plausible influences, but the cited evidence does not establish a consistent moderator specific to gratitude journals. Do not turn difficulty with the exercise into another reason to judge how grateful you think you should feel.

How often, and what about the health claims?

No single optimal frequency is established. Gratitude studies use markedly different activities, schedules, comparison groups, and follow-up periods, so percentages of "significant" studies at each frequency are not a reliable head-to-head comparison. Choose a sustainable schedule and adjust if the exercise becomes mechanical or distressing rather than treating daily or weekly practice as evidence-based universally.

On physical health: gratitude journaling is frequently said to improve immune function and lower blood pressure. This is mostly unsupported. The one well-evidenced physical benefit is sleep quality — the gratitude condition in Emmons and McCullough's Study 3 reported sleeping longer and feeling more refreshed, and a systematic review of 19 RCTs found sleep quality improved in 5 of 8 studies that measured it. The broader claims about immune enhancement and blood pressure are extrapolations from limited, poor-quality studies.

Why popular accounts sound more confident than the evidence warrants

Several structural problems in the research base explain the gap between what you read and what the studies actually show:

• The control-group problem. Most early studies compared gratitude journaling to writing about daily hassles — which actively worsens mood. Observed differences partly reflect the negative effects of the comparison, not only the positive effects of gratitude.
• Demand characteristics. Participants in a "gratitude exercise" know what kind of response is expected and may report feeling better partly because they expect to. No study has adequately controlled for this.
• Publication bias. After correcting for publication bias, Davis et al. (2016) found the effect against neutral activity controls dropped from d = 0.14 to d = 0.02 — essentially zero.
• Short follow-ups. No well-designed study has tracked participants beyond about six months. Long-term effects over years are genuinely unknown.

None of these concerns eliminate the evidence for a real effect. They calibrate it.

What this means in practice

The evidence points toward a more specific set of recommendations than "write a gratitude list every day."

Write narratively, not as a list.

A short account of one specific thing — what happened, who was involved, why it mattered, and what you noticed — is one evidence-informed alternative to a targeted list. In Regan et al.'s one-week trial, long-form conditions outperformed two targeted list conditions on several outcomes, but an unconstrained list also produced some immediate benefits.

To make the difference concrete, here is what the two formats look like side by side:

List entry: "Grateful for my friend Sara. Grateful for good weather. Grateful for my health. Grateful for coffee. Grateful for a productive day."

Narrative entry: "Sara called this afternoon out of nowhere — no reason, just checking in. We ended up talking for forty minutes. I kept thinking afterward how rare it is that someone does that. I felt lighter for the rest of the day in a way I hadn't expected. I don't think I tell her enough how much that kind of thing matters."

The narrative example invites more detail than the list. That may help some people engage, but the research does not show that every list is cognitively empty or that elaboration is the established mechanism.

Be specific.

"My colleague spent twenty minutes helping me untangle a problem I'd been stuck on for two days" preserves more context than "grateful for colleagues." Specificity is a useful writing choice; claims that it necessarily prevents habituation or is more genuine go beyond the cited trials.

If it feels forced, change the format before you quit.

If the practice feels forced, trying a different format is reasonable. Do not assume the problem must be list format, however; fit, context, and preference vary, and the evidence does not identify one response for everyone.

Do not over-prescribe frequency.

The literature does not establish that two or three times per week is superior to daily practice. Use a schedule that allows attention without turning the task into an obligation, and change or stop the exercise if it is unhelpful.

A gratitude letter can remain private, but delivery was not isolated.

In Wong et al. (2018), most participants did not send their letters. Because sending was optional rather than randomised, the study cannot establish that writing alone was sufficient, that sending adds nothing, or that the two choices are equivalent.

Do not infer one minimum duration.

Regan et al. tested daily activities for one week and measured a one-week follow-up. Wong et al. tested three weekly gratitude letters alongside psychotherapy and found effects at follow-ups four and twelve weeks after writing ended. Those designs show that timing matters, but they do not establish how long an ordinary gratitude journal must be used.

Gratitude interventions have a small average effect across many formats. Evidence for gratitude journaling specifically is narrower and varies with the writing task, comparison group, population, and outcome.

Comparisons with no treatment are generally more favourable than comparisons with other psychologically active exercises. That makes gratitude writing one accessible option among several, not a practice that reliably wins in every study or format.

If standard advice has not worked for you, format is one possible factor. You could try one specific narrative, an unconstrained list, a private gratitude letter, or a different structured positive activity altogether. No exact frequency or minimum duration is established. If you are looking for practical techniques for different kinds of journaling, our guide on how to journal for anxiety covers expressive writing approaches, and our article on whether journaling actually works reviews the broader evidence base across all outcomes. For tips on making any writing practice stick, see our piece on how to start a journaling habit.