Does Journaling Help With Depression? What the Research Says

Journaling may help some people with depression, but the evidence is inconsistent and does not support using it as treatment on its own. A depression-focused meta-analysis found no significant long-term average effect from standard brief expressive writing, while several smaller studies of other structured exercises reported benefits. Open-ended writing can also become repetitive rumination for some people, so the method and the person's response both matter.

The contrast shows why each protocol, population, comparison condition, and outcome must be examined separately. It does not divide writing into one form that always works and another that always harms.

This article covers what the research says and what remains uncertain for someone considering writing alongside depression care. For a practical, safety-focused companion, see our guide on how to journal for depression. Our article on journaling and anxiety reviews a related but distinct evidence base.

If depression is significantly affecting your daily life, please speak to a doctor or therapist. Journaling can complement professional care; it cannot replace assessment or treatment. If you may act on thoughts of suicide or self-harm, or someone is in immediate danger, contact local emergency services now. Find A Helpline lists verified crisis services in more than 175 countries.

If you find yourself relying on alcohol or other substances to manage how you feel, or writing repeatedly leaves you more distressed, contact a healthcare professional rather than trying to push through alone.

What does the evidence base look like?

Several meta-analyses have examined instructed writing about thoughts and feelings. Their depression findings range from null to small and differ in included populations, interventions, comparisons, and follow-up periods.

An early study by Pennebaker and Beall (1986) assigned 46 healthy undergraduates to write on four consecutive days about traumatic events or trivial topics. Later protocols vary substantially; the common 15-to-20-minute, three-or-four-session format describes a research tradition, not a depression dose.

Across all psychological outcomes, the effects are small. Frattaroli's 2006 meta-analysis of 146 randomised studies in Psychological Bulletin found an overall effect of r = .075 — small but statistically significant. For depression specifically, the effect was r = .073.

That small pooled association spans heterogeneous outcomes. A depression-focused review gives a narrower answer.

The Reinhold finding

Reinhold, Bürkner and Holling (2018) published a meta-analysis in Clinical Psychology: Science and Practice that focused exclusively on depression as an outcome, analysing 39 randomised controlled trials. Their finding: standard brief expressive writing produced no significant long-term effect on depressive symptoms in physically healthy adults.

A small decrease appeared immediately after writing, but did not hold at follow-up.

In this depression-focused review, the standard brief expressive-writing protocols did not show a significant long-term average effect.

There are important caveats. Only around 6% of the included studies involved people formally diagnosed with depressive disorders — most tested whether writing helped prevent or reduce subclinical symptoms in healthy people.

The review found associations between outcomes and factors such as participant age, number of sessions, topic specificity, and spacing. Moderator findings across a group of studies do not establish an ideal prescription for an individual. The null overall result means the standard brief protocol should not be presented as a reliable treatment for depression.

Other meta-analyses

Sohal and colleagues (2022) in Family Medicine and Community Health examined 20 randomised trials and reported larger average changes for anxiety and PTSD symptoms than for depression. A subgroup of interventions lasting more than 30 days showed greater improvement in depression scores than shorter interventions, but the comparison does not prove that duration caused the difference or establish a minimum course of writing.

Lin Guo's 2023 meta-analysis combined depression, anxiety, and stress across 31 randomised trials and reported a small delayed average effect (Hedges' g = −0.12). Trials with one to three days between sessions had stronger average effects than trials with longer spacing, but this study-level moderator is not a personal schedule or a depression-specific effect.

Frisina, Borod and Lepore (2004) examined nine small, heterogeneous studies of clinical populations and found d = .21 for physical outcomes and d = .07 for psychological outcomes, with a significant planned contrast. The limited evidence cannot establish a general body-versus-mind rule or a depression effect.

Across these meta-analyses, effects for depression are small and inconsistent, ranging from null to roughly d = 0.15 depending on the review and outcome. That is not evidence that journaling treats a depressive disorder. For a broader view of the evidence across all outcomes, see our article on whether journaling actually works.

Which study matters most for depression specifically?

Krpan and colleagues (2013) reported a small randomised trial of expressive writing in people diagnosed with major depressive disorder through structured clinical interview.

Forty participants with MDD were randomised to either expressive writing (20 minutes per day for three consecutive days) or control writing about daily organisation. Depression was measured using the BDI-II and PHQ-9.

The expressive writing group showed significant reductions in depression scores by day five, with benefits persisting at the four-week follow-up.

This is encouraging but limited: it was a 40-person brief report, and this article does not identify an independent replication of the same protocol in a comparable diagnosed sample.

The study supports further investigation of a directed protocol; it does not prove that ordinary journaling treats depression or establish the size and durability of an effect.

Why Should Depression and Anxiety Evidence Stay Separate?

Some reviews report different subgroup estimates for depression, anxiety, and PTSD, but protocols and populations vary. A result for a combined distress outcome or one diagnosis should not be imported into another. Our anxiety review likewise describes small, mixed effects rather than a simple hierarchy.

Several hypotheses may help explain this difference, although none has been established as a complete mechanism.

Rumination versus worry.

Rumination — repetitive, often past-focused dwelling — is common in depression, while worry is more often future-focused and associated with anxiety. These patterns overlap and do not describe every person's experience.

The Pennebaker protocol asks people to explore their deepest thoughts and feelings. For a person already caught in brooding, that open instruction may become written rumination rather than organised reflection. This is a plausible risk, not a rule separating everyone with anxiety from everyone with depression.

The need for action.

Depression often involves reduced motivation and behavioural withdrawal. Behavioural activation — gradually reconnecting behaviour with meaningful or rewarding activity — is one of several recommended psychological treatments for depression. Writing may help someone plan or track activities, but writing alone does not provide the full intervention.

Self-focused language.

In an observational study by Rude, Gortner and Pennebaker (2004), currently depressed students used more negatively valenced and first-person singular words than never-depressed students. That association does not show that an expressive-writing instruction reinforces depression or that changing pronouns would help.

How does rumination turn journaling against you?

Monitor whether a writing task adds perspective or action, remains neutral, or increases repetitive brooding and distress. The page itself does not guarantee processing.

A review by Nolen-Hoeksema, Wisco and Lyubomirsky describes rumination as repetitive thinking about distress, its causes, and consequences, and links it with depression and several proposed cognitive and social pathways. It did not test journals or show that every negative entry is harmful.

Treynor, Gonzalez and Nolen-Hoeksema (2003) separated brooding and reflective-pondering factors in questionnaire data. Their longitudinal associations differed, but these were not experimental journal conditions and do not create a simple rule for classifying an entry.

Gortner, Rude and Pennebaker (2006) tested three consecutive days of expressive or control writing in 97 college students selected for prior elevated but currently low depressive symptoms. Among participants above the sample median on emotional suppression, the expressive-writing group reported lower depression at six months.

In that study, reduced brooding statistically mediated the later difference in depression, while reflective pondering did not change. This offers one possible explanation for the result in that sample; it does not establish how writing works for everyone with depression.

In a different context, Sbarra and colleagues (2013) assigned recently separated adults to traditional expressive writing, narrative expressive writing, or control writing. Expressive-writing conditions were associated with worse later emotional outcomes among participants higher in baseline brooding. This moderator finding concerned marital separation, not depression treatment or everyone who ruminates.

The practical signal: if journaling consistently leaves you feeling worse, cycling through the same thoughts without movement, it may be functioning as rumination. That is useful information, not a failure. It suggests a different approach rather than more of the same.

What does the language of your writing reveal?

Pennebaker developed a text analysis tool called Linguistic Inquiry and Word Count (LIWC) that quantifies the psychological properties of written language. The patterns it has identified in depressed writers are striking.

Rude, Gortner and Pennebaker (2004) found that currently depressed students used significantly more first-person singular pronouns than never-depressed students. A meta-analysis by Edwards and Holtzman (2017) of 21 studies confirmed this: the correlation between depression and first-person singular pronoun use was r = .13 — small but robust across populations and conditions.

The small correlation does not establish whether first-person language is a symptom, cause, consequence, or context-dependent marker. Telling someone to remove “I” from an entry would not follow from this evidence.

Some studies have also associated changes in insight and causal words with better outcomes. Researchers interpret this as a possible sign of cognitive organisation, but the finding is not consistent enough to locate a universal therapeutic mechanism in particular words.

What types of writing help most with depression?

Standard Pennebaker expressive writing has weak to inconsistent effects on depression. Several alternative approaches show more promise.

Self-compassion writing

Shapira and Mongrain (2010) in the Journal of Positive Psychology randomised 188 participants vulnerable to depression (with moderate depressive symptoms at baseline) to write self-compassionate letters (comforting themselves as they would a close friend), optimism letters, or control writing about early memories, for seven consecutive evenings.

Both active interventions produced significant decreases in depression sustained to three months, and increased wellbeing observable at six months.

The instructions focused on self-compassion or optimism rather than trauma disclosure. This was one online study in a vulnerable sample, with a high dropout rate and no standardised effect sizes reported in the article. It does not establish that discomfort during the exercise is meaningful or that either task treats depression.

Gratitude writing

Cregg and Cheavens (2021) meta-analysed 27 gratitude-intervention studies and reported a small post-test effect (g = −0.29) for depression and anxiety combined. The review included varied gratitude activities and does not isolate journaling or clinical depression.

Seligman and colleagues (2005) reported that an online “three good things” exercise outperformed a placebo exercise on depressive-symptom scores at several follow-ups in a self-selected sample. It was not a trial in people diagnosed with depression, and other tested exercises had different time courses.

Neither study establishes specificity as the active ingredient or shows that a detailed entry is universally better than a short list.

Best Possible Self writing

Laura King's 2001 study assigned 81 undergraduates to write for 20 minutes on four consecutive days about trauma, a best possible future self, both, or a non-emotional control. Best-possible-self writing was less upsetting and improved subjective wellbeing at three weeks; at five months, the three emotional or goal-writing groups had fewer illness visits than control. It did not test depression treatment.

In King's study, the exercise caused less distress during writing than trauma-focused writing. That makes it a possible alternative to explore, not a guarantee that it will be lower-risk or effective for every person with depression.

Writing as part of therapy

Kazantzis, Whittington and Dattilio (2010) meta-analysed homework effects in cognitive and behavioural therapy. Assignments included behavioural and other tasks, not journaling as one isolated component. The result supports neither free-form writing nor a specific depression journal.

Gerger and colleagues (2022) reviewed writing interventions for trauma survivors with PTSD symptoms, not depression. Although some network estimates for enhanced writing and psychotherapy overlapped, the authors reported substantial heterogeneity and inconsistency and said a definite comparison could not be made. Those findings should not be generalised into a depression treatment rule.

If you are working with a therapist, ask which written exercises fit your treatment plan. A thought record or activity log has a defined clinical purpose; free-form journaling is not interchangeable with those tools.

For more on the boundaries between journaling and professional care, see our article on whether journaling can replace therapy.

What does privacy have to do with it?

Pennebaker's standard instructions emphasise confidential writing and sometimes tell participants they may destroy or hide it afterwards. Privacy may make honest disclosure easier, but the evidence does not establish it as a universal precondition or a single mechanism of expressive writing.

For some writers, concern about a partner, colleague, or service provider reading an entry may change what they record. The studies reviewed here do not show that privacy matters more for depression than for other conditions or that encryption changes clinical outcomes.

When choosing a digital journal, check what is encrypted, who controls the keys, what metadata remains visible, and how account recovery works. Strong technical safeguards reduce provider access, but no design makes every endpoint, backup, or account immune to risk.

What Can and Cannot Be Inferred?

The studies above test different tasks and samples. They support cautious possibilities, not a forecast for one person.

Smyth's 1998 meta-analysis found higher immediate distress after expressive-writing sessions (d = .84), without evidence that this immediate response explained later outcomes. Brief discomfort can occur, but worsening or persistent distress is a reason to stop and seek support — not proof that the exercise is working.

Follow-up timing changes what is measured.

Lin Guo's 2023 review found larger average effects at one- to three-month follow-ups than immediately after writing. Sohal and colleagues reported a duration-related subgroup difference. Together these findings justify measuring outcomes beyond the first session, but they do not establish that everyone should write for a particular number of weeks.

There is no validated universal schedule.

Studies have used very different protocols, from a few concentrated sessions to longer programmes. Moderator and subgroup results do not justify prescribing several sessions per week for months. If you use writing alongside treatment, agree the purpose and frequency with your clinician; for self-help, choose a manageable trial and reassess rather than treating persistence as evidence of effectiveness.

Language patterns are research measures, not personal tests.

Exploratory linguistic studies have linked some changes in perspective, causal language, and insight words with better outcomes. These patterns may help researchers study cognitive organisation, but they cannot tell an individual whether a particular entry is therapeutic.

If writing repeatedly returns to the same dark thoughts and increases distress without adding perspective or action, that is a reason to pause and discuss another approach with a clinician. A single entry cannot diagnose rumination or tell you whether a treatment is working.

If you want to try writing, choose a bounded, low-pressure exercise and notice how you feel during and afterwards. Stop if distress rises or the same thoughts become more intense. A self-compassion letter or a specific activity plan may be easier to contain than open-ended disclosure, and a clinician can help decide whether writing belongs in your care.