A blank page, a cheap biro, fifteen minutes. That was the entire treatment.
Students who did it four days running turned up at the campus health centre over the following six months at roughly half the rate of students who spent the same fifteen minutes describing their shoes.
That result belongs to James Pennebaker, whose 1986 paper with Sandra Beall built a research field more or less from scratch. Writing in Perspectives on Psychological Science three decades later, Pennebaker recalled running the experiment late in 1983 with the help of a new graduate student, and described it, with admirable bluntness, as horribly underpowered.
Odd foundation for several hundred follow-up studies to stand on.
What the students were actually asked to do
The brief was short and a little brutal. Write for fifteen minutes about the most traumatic or upsetting experience of your life. No one will read it, no one will mark it, spelling does not matter.
Controls got something deliberately inert. Describe your room. Describe your shoes.
Pennebaker also split the writers into separate conditions, and this is the detail that tends to vanish in the retelling. Some were told to stick to the facts of what happened, some to write only about how they felt, some to do both. As Karen Baikie and Kay Wilhelm set out in their review for Advances in Psychiatric Treatment, benefits showed up in the students who wrote about their deepest thoughts and feelings. In a later summary with Cindy Chung, Pennebaker noted that participants who recounted the bare facts of a trauma showed no improvement whatsoever.
Reciting events did nothing. The feelings had to be on the page.
Pennebaker’s retrospective also records a replication, run two years after publication with the psychologist Janice Kiecolt-Glaser and the immunologist Ronald Glaser, that turned up both fewer clinic visits and immune changes consistent with better health. That is roughly the point at which the finding stopped looking like a quirk of undergraduate behaviour.
When it left the psychology building
A campus clinic visit is a soft measure. Undergraduates go to the doctor for all sorts of reasons, not all of them medical.
So the serious test arrived in 1999, when Joshua Smyth and colleagues at Stony Brook ran the exercise past 112 patients with asthma or rheumatoid arthritis and measured them with clinical tests instead of questionnaires: spirometry for lung function, a rheumatologist’s examination for the joints. Their trial in JAMA reported clinically relevant improvement at four months among the patients assigned to describe the most stressful event of their lives, on top of the standard medical care every participant was already getting. Smyth’s team was careful about the limits of what they had shown, flagging that it remained unknown whether those gains persisted past four months or transferred to other diseases.
Where the evidence turns murky
Hundreds of trials later, the meta-analyses are still arguing.
The biggest of them, run by Joanne Frattaroli for Psychological Bulletin, pooled 146 randomised studies and landed on an average effect size of r = .075, a figure that is real and statistically significant but tiny. Meanwhile, as documented in a Frontiers in Psychology trial write-up, several other reviews found no supporting evidence for physical or mental health benefits at all, with the disagreements traceable to which studies each review let in and which statistical approach it used. Arguments like this tend to survive for decades precisely because both sides are using a defensible method.
A 2023 review in the British Journal of Clinical Psychology took a narrower slice: 31 trials with long follow-ups, about 4,000 people between them. Small effect on depression, anxiety and stress, and a curious pattern in the timing. Whatever the writing does appears to show up weeks or months afterwards rather than on the day.
The people it seems to make worse
“Contraindicated” is a heavy word to find in a paper about a homework exercise.
It turns up in a trial led by Andrea Niles with Annette Stanton and co-authors at UCLA, published in Anxiety, Stress & Coping. A hundred and sixteen healthy young adults went through four writing sessions, and on average nothing happened to their anxiety, depression or physical symptoms. Sorting participants by temperament changed everything: people who habitually express emotion were less anxious three months later, and people who do not normally talk about their feelings were more anxious. Identical instructions, opposite outcomes.
Which makes the standard advice to “just write it down” look less like a prescription and more like a bet on the reader’s personality.
Why it might work at all
Explanations are plentiful and none of them has won. Baikie and Wilhelm canvass the main candidates, including emotional habituation, the reduced physiological cost of no longer suppressing something, and the possibility that turning a mess of experience into a sentence with a beginning and an end changes how the memory is filed.
All plausible. All difficult to isolate in a study where the intervention is a stranger handing you paper.
What has survived, forty years on, is a procedure with no cost, no side effects worth insuring against, and an effect size that would embarrass a pharmaceutical company. There is no product attached to it and no patent to defend, which may be part of why the question of whether it works has been allowed to stay open this long.