Does Journaling Help With Stress?

The research tells a nuanced story. There is genuine evidence that writing about difficult experiences produces health benefits — but the specific claim that journaling reliably reduces stress rests on a foundation that is weaker, more conditional, and more honest about its limits than most popular accounts suggest. The average effect across meta-analyses is roughly d = 0.15 to 0.16: real but modest, and the benefits depend heavily on the type of stressor, the writer's personality, and the structure of the writing itself.

This article covers the evidence. We have already written a practical guide on how to journal for stress covering specific techniques with evidence behind them. This companion piece focuses on the underlying research: the effect sizes, the mechanisms, the boundary conditions, and — importantly — the findings that most accounts leave out.

Where does the evidence come from?

The field of expressive writing research began with James Pennebaker at the University of Texas in 1986. His foundational study asked healthy college students to write for fifteen minutes a day across four days about the most stressful or traumatic experience of their lives. Compared to students who wrote about neutral topics, the writing group visited the student health centre at roughly half the rate over the following six months.

That finding launched four decades of research. But the picture that has emerged is substantially different from the one popularised in wellness culture.

The three most important meta-analyses tell a consistent story about the size of the effect:

Smyth (1998)

Analysed thirteen randomised studies and found an overall effect size of d = 0.47 — a medium-sized benefit across psychological wellbeing, physiological measures, reported health, and general functioning. This was the figure that launched a thousand wellness articles.

Frattaroli (2006)

Expanded the scope dramatically, analysing 146 randomised studies in Psychological Bulletin. With that larger, more representative sample and more appropriate statistical methods, the overall effect shrank to r = .075. The earlier, widely-cited figure appears to have been inflated by the small pool of early studies.

Pennebaker (2018)

In a retrospective review in Perspectives on Psychological Science, placed his best estimate at around d = 0.16 — and was explicit that the benefits were real but modest.

None of these meta-analyses isolated stress as a separate outcome with its own effect size. The benefits were pooled across psychological wellbeing, physical health, physiological markers, and various behavioural outcomes. When studies have directly measured perceived stress — using scales like the Perceived Stress Scale or the DASS-21 stress subscale — the results are frequently null.

For a broader look at the evidence behind writing as a health practice, see our article on whether journaling actually works.

What is the most important finding most articles leave out?

In 2020, a randomised trial published in Frontiers in Psychology studied the effects of expressive writing during the COVID-19 pandemic. One hundred and twenty participants were assigned to either expressive writing or a control condition. The writing group showed no benefits — and on some measures showed significantly elevated stress compared to controls.

This is not a minor footnote. It is a direct test of whether writing about stress helps during an ongoing, real-world stressor. The answer was: not only did it not help, it may have made things worse.

The researchers' interpretation explains a great deal about the conditional nature of the evidence. Pennebaker's expressive writing paradigm was designed for processing past events — experiences that have a beginning, a middle, and an end. Writing about an experience that has ended allows the construction of a coherent narrative, the kind of meaning-making that the research consistently identifies as the active ingredient.

Writing about an ongoing, uncontrollable crisis is different. There is no narrative closure available because the story is still happening. Under these conditions, writing about the stressor may function as repeated exposure to a source of distress rather than processing of it — rumination with a pen.

The distinction between past and ongoing stressors is not peripheral to the evidence on writing and stress. It is central to it.

Is the physical evidence stronger than the psychological evidence?

One of the more striking patterns in the research is that writing appears to affect the body more reliably than it affects the mind.

Frisina, Borod, and Lepore's (2004) meta-analysis of clinical populations found that expressive writing significantly improved physical health outcomes (d = 0.21) but did not significantly improve psychological health outcomes (d = 0.07). The difference between the two was statistically significant.

The most clinically significant study in the entire field makes this concrete. Smyth and colleagues (1999), published in the Journal of the American Medical Association, recruited 107 patients with asthma or rheumatoid arthritis — conditions significantly affected by chronic stress — and had them write about the most stressful experience of their lives for twenty minutes across three days.

At four-month follow-up, 47% of those in the writing condition showed clinically relevant improvement, compared to 24% of controls. Asthma patients showed measurable improvement in lung function. Rheumatoid arthritis patients showed reduced disease severity.

But there is a crucial detail: psychological stress was not the measured outcome. The study assessed lung function and disease severity through clinical evaluation — objective physical markers. The improvement in physical symptoms is real. Whether participants felt less stressed is a different question that this study did not answer.

The cortisol evidence is similarly instructive. Creswell and colleagues (2005) found that writing about personal values before a laboratory stress test completely eliminated the cortisol response — participants who wrote about their values showed no significant cortisol increase, while controls showed a significant spike. DiMenichi and colleagues (2018) found that writing about past failures before a stressor attenuated the cortisol response. But Nazarian and Smyth (2013) tested five different versions of expressive writing instructions and found that none of them significantly influenced cortisol compared to controls.

The immune findings are among the most consistent. Pennebaker, Kiecolt-Glaser, and Glaser (1988) found that trauma writing enhanced immune function measured by T-lymphocyte response. Petrie and colleagues (1995) found that students who wrote expressively showed higher hepatitis B antibody levels at four- and six-month follow-ups — a functional immune benefit in response to an actual vaccine. These findings point to writing affecting the physiological underpinnings of stress without necessarily changing how stressed people report feeling.

Why is stress different from anxiety and depression?

The evidence for writing and anxiety is somewhat more established. The evidence for writing and depression is more mixed. The evidence for stress specifically is the thinnest of the three — and this is not accidental. It reflects something important about what stress is.

Lazarus and Folkman's foundational stress model defines psychological stress as the experience of demands exceeding available coping resources. Unlike anxiety, which is characteristically future-focused and tied to imagined threats, stress is typically tied to real external pressures — a workload, a relationship conflict, a financial situation, a health challenge.

Unlike depression, which involves a pervasive low mood and loss of engagement, stress tends to be more situational and more directly connected to identifiable causes.

This has a practical implication. Because stress involves identifiable external stressors, it might seem like an ideal target for writing-based cognitive reappraisal — articulate the problem, examine your thinking about it, develop a more accurate or adaptive perspective. In theory, this should work. In practice, it depends heavily on whether anything can actually be done about the stressor, and whether the stressor has ended.

Lazarus and Folkman distinguished between problem-focused coping — addressing the stressor directly — and emotion-focused coping — managing the emotional response to it. Expressive writing is primarily an emotion-focused approach. Lestideau and Lavallee (2007) found that for current stressors, planful writing — developing specific strategies to address the problem — was significantly more effective than standard emotional expressive writing, which produced no significant benefits.

This finding echoes the stress guide on this blog: for ongoing stress, action orientation tends to outperform emotional exploration.

What drives benefits when they occur?

Four decades of research have not produced consensus on why expressive writing works. Pennebaker himself shifted between theories over the years, ultimately acknowledging in 2018 that the original inhibition theory — suppressing thoughts about stressful experiences creates physiological burden, and writing releases it — was never empirically supported.

The most compelling current account involves narrative formation. Smyth, True, and Souto (2001) found that participants who wrote in narrative form — constructing a coherent story — showed health benefits, while those who wrote in fragmented form (listing thoughts and feelings) did not differ from controls.

Pennebaker's linguistic analyses across hundreds of writing samples found a consistent pattern: people who improved over writing sessions showed increasing use of causal words (because, reason, since) and insight words (understand, realise, notice) across sessions. They moved from describing what happened to making sense of it. That shift — from raw experience to integrated meaning — appears to be where the benefit lives.

Working memory offers a related explanation. Klein and Boals (2001) demonstrated across two semester-long experiments that expressive writing increased working memory capacity and reduced intrusive thinking. The proposal is that stress-related rumination occupies cognitive resources, and writing externalises these concerns — a form of cognitive offloading that moves worries from the active mental workspace to external storage.

Ramirez and Beilock (2011), publishing in Science, found that ten minutes of writing about worries immediately before a high-stakes exam eliminated the anxiety-performance gap in high-anxiety students. This is one of the most directly applicable findings for stress management — writing before a stressful event, not after it, can serve as a form of performance inoculation. However, attempts to replicate this specific finding have produced mixed results, and it should not be treated as settled.

Who benefits and who does not?

The average effect size conceals substantial variation. Some people benefit substantially. Others show no benefit. A meaningful minority show harm. Understanding who falls into which group is more useful than knowing the average.

Emotional expressiveness.

This is the most consistently documented moderator. Niles and colleagues (2014), in a randomised trial with 116 healthy adults, found that the effect of expressive writing on anxiety depended strongly on trait expressiveness. People who naturally externalise emotions showed significant anxiety reductions at three-month follow-up. People who typically suppress emotions showed significant anxiety increases. The practical implication: if you habitually keep your feelings to yourself, standard expressive writing may not be the right starting point.

The nature of the stressor.

Research consistently finds that expressive writing works best for past, completed stressors — experiences that can be narrated because they have ended. For current, ongoing stressors, the evidence is weaker and the risk of amplification higher. Cameron and Nicholls (1998) found that combining emotional expression with concrete coping plans outperformed emotional expression alone for current stressors.

Baseline stress level.

Frattaroli's meta-analysis found stronger effects in samples with higher initial distress. This makes theoretical sense: people who have more to process have more to gain from processing. For people experiencing mild, everyday frustrations, the benefits are likely to be smaller.

Rumination.

This interacts with context in complex ways. Sloan and colleagues (2008) found that among college students, those with higher brooding tendencies actually showed fewer depression symptoms after expressive writing over a six-month follow-up — suggesting the writing redirected maladaptive brooding into more productive reflection. But Sbarra and colleagues (2013) found the opposite for people going through divorce: high ruminators in that acutely distressing situation fared significantly worse with expressive writing, showing elevated emotional distress at up to nine months post-study.

The difference may come down to whether the stressor is resolved. Completed events can be ruminatively processed into meaning. Ongoing losses cannot.

What is the honest picture?

The research does not support the simple claim that writing in a journal reduces stress. It supports something more conditional and more interesting.

Writing about stressful experiences — specifically experiences that have ended, written in a way that constructs narrative meaning rather than recycling distress, by people naturally inclined toward emotional expression, for fifteen to twenty minutes across several sessions — produces small but real improvements in health outcomes. The effects on physical health markers are somewhat more robust than the effects on psychological stress measures.

The mechanism appears to involve meaning-making and cognitive offloading more than catharsis or inhibition release.

What writing in a journal does not reliably do is reduce perceived stress in the moment, during an ongoing stressor, for people who habitually suppress emotions, or in a single session of free emotional venting. The wellness industry's unconditional version — write your feelings, feel less stressed — does not have strong evidence behind it.

This is not a reason to dismiss the practice. A small effect size from a free, brief, private activity is genuinely valuable. The physical health evidence — immune function, chronic disease markers, health centre use — is more consistent and clinically meaningful than popular summaries suggest.

And the sleep finding from Scullin and colleagues (2018), who measured brain activity directly in a sleep laboratory, is robust: writing a specific to-do list for five minutes before bed reduced sleep onset by approximately nine minutes — a practically significant effect for anyone whose stress manifests as lying awake at night. For more on the sleep-specific evidence, see our article on journaling before bed and sleep.

The more honest framing is this: writing in a journal is not a stress-reduction tool in the way that exercise or meditation are — with moderate to large effects and broad applicability. It is a cognitive processing tool that, under the right conditions, facilitates the kind of meaning-making that helps people integrate difficult experiences and free cognitive resources consumed by unresolved concerns. That is a more specific claim, but it is also a more accurate and more useful one.

What does this mean in practice?

If you are using a journal to work through a stressful period, a few implications follow from the research.

Timing matters.

Writing about stressors that have passed is more likely to produce benefits than writing about stressors that are still happening. For ongoing stress, action-oriented writing — specifying what you can do, what is within your control, what your next step is — tends to outperform pure emotional processing.

The goal is meaning, not venting.

Writing the same distress in the same words session after session is closer to rumination than to processing. Writing moves toward benefit when it shifts — over sessions — from describing what happened to understanding why, what it means, and what comes next.

Fifteen to twenty minutes, several times.

This is the best-validated format. Single sessions produce modest or no effects. Very short sessions (two to three minutes) may have some benefit, but the research support is weaker.

Privacy enables honesty.

Pennebaker built confidentiality into his protocol from the beginning. The research on secrecy — Slepian and colleagues' programme of work showing that keeping significant concerns private creates measurable physiological and cognitive burden — suggests that genuine disclosure requires genuine privacy. If there is any chance your writing might be seen, the honest writing that the mechanism requires may not happen. For more on what this means in a digital context, see our article on who can actually read your digital journal.

For a detailed breakdown of specific techniques and when to use each one, see our practical guide on how to journal for stress.

Start today: after your next obligation winds down, open your journal and spend fifteen minutes writing about a stressful situation that has already resolved. Do not aim for eloquence — aim for understanding. Write what happened, why it bothered you, and what you see now that you did not see then.