Does journaling help with anxiety? What the studies show and what they do not

In 1986 James Pennebaker asked a group of students to write about the most painful event of their lives on four consecutive days. Fifteen minutes each time, alone, no editing. On those days their blood pressure went up and their mood went down. Over the next six months they visited the student health centre less often than the students who had written about neutral topics.

That experiment started what is now called expressive writing. It has been repeated hundreds of times in forty years, and the question "does journaling help with anxiety" has an answer that does not fit in one word. It helps. Not everyone, not by much, and not in the way self-help covers promise. Below is what I could check against the original papers rather than the retellings.

What was actually measured

One thing first: in most of these studies nobody kept a journal in the everyday sense. The classic protocol, as Baikie and Wilhelm describe it in their 2005 review, is three to five sessions, often on consecutive days, 15 to 20 minutes each, on "the most emotional experience", with an instruction to ignore grammar. That is closer to a one-off exercise than to a daily habit.

They did not measure "anxiety" as such either. Smyth's 1998 meta-analysis split the results into four groups: self-reported physical health, psychological well-being, physiological measures, and general functioning. All four moved in the right direction. Health behaviours, like sleep, exercise and smoking, did not change.

146 randomised studies, average effect r = .075: positive, statistically significant, and small.

The largest meta-analysis is Joanne Frattaroli's from 2006. One hundred and forty-six randomised studies, an effect that exists, is significant, and is small. The more useful part of her paper is the moderators. The effect was stronger for people who already had a trauma or a health problem, when there were more sessions, and when people wrote at home rather than in a lab.

A recent meta-analysis by Sohal and colleagues in 2022 looked at journaling specifically for anxiety disorders, depression and post-traumatic stress: twenty randomised trials, a difference of roughly five per cent in favour of journaling, an effect the authors call small to moderate. In the same breath they say the data are heterogeneous, the methods often weak, and the recommendation only reaches grade B. Their argument for journaling is modest and honest: it costs almost nothing and risks almost nothing.

Three mechanisms with evidence behind them

Why writing does anything to anxiety at all is not fully understood, and the researchers say so. Three things do come up paper after paper.

Naming. Anxiety does not survive specifics well. While it is "just bad", it fills the whole room. Once the page says "I am afraid that on Thursday's call they will ask about the budget and I am not ready", it takes up one sentence. In Pennebaker and Beall's study, the group that wrote about both the facts and the feelings got the delayed benefit; the group that wrote about feelings alone did not.

Processing, not venting. Ullrich and Lutgendorf in 2002 separated two ways of writing: focusing on emotions only, or on emotions together with thoughts about the event. One hundred and twenty-two students, one month of journaling. Those who wrote about feelings and what they thought about them were more likely, by the end of the month, to see something useful in the event. The emotions-only group reported more severe illness symptoms over the same month than everyone else. Pouring it out, on its own, can make things worse.

Distance. A written line separates you from a thought by exactly one line. A week later you can reread it and notice the prediction did not come true. This is not listed as a separate mechanism in the meta-analyses, but it is how the delayed effects are usually explained: the anxious thought stops being the weather and becomes text you can argue with.

What does not work

Two findings from the same papers get quoted less, because they are inconvenient.

First: "just get it out" does not help and can hurt. That is the Ullrich and Lutgendorf result, and it fits an older one from Smyth: right after writing about something hard, people almost always feel worse, and that temporary dip is unrelated to any later benefit. If the evening entry turns into the same fears retold in a loop, with no line that starts "what I think about this is", the journal works like chewing gum.

Second: daily obligation is not part of any protocol. Three to five sessions. Baikie and Wilhelm list the groups where writing did not help or did harm: adult survivors of childhood abuse, veterans with post-traumatic stress, people in acute grief. For them the advice is the same: do not force it, and stop when it gets heavy.

Eight journaling ideas for anxiety

Everything below leans on the same three mechanisms: name it, process it, step back. Each idea takes between one line and ten minutes.

1. The specific fear. Not "I feel anxious" but "I am afraid that…" with a place, a time and the people involved. One sentence.
2. What I think about it. The second line after the first. Not "how do I calm down" but "what do I think about the fact that I am scared". This is the exact move from Ullrich and Lutgendorf.
3. A prediction with a date. "Friday will be bad because…" Reread it a week later. Most of anxiety's predictions do not come true, and that has to be seen, not heard.
4. What has happened like this before. Three times it was this scary, and how each one ended.
5. The body. Where the anxiety is right now: throat, shoulders, stomach. No interpretation, just the location.
6. The worst case and the next step. Describe the worst outcome, then one action you could take if it happened.
7. What was plain today. Plain, with no grade attached. Anxiety erases the neutral hours; a journal gives them back.
8. A letter to someone you will not send. Pennebaker's classic: you need an addressee, not a delivery.

A note on frequency. The studies worked with three to five sessions. If you write every day and then skip a week, nothing broke. If anything, pauses are in the protocol and "days in a row" is in none of them.

What a journal does not do

It does not replace talking to a professional, and that is not a disclaimer for its own sake. The effect in the meta-analyses is small, and in some groups it is absent. A journal is good for understanding what is happening to you, and as the thing you bring to a person who knows how to work with it.

It does not help during a panic attack. Writing is impossible in those minutes, and unnecessary. A journal is for afterwards.

And it is not a place where you should feel worse for long. A short dip after writing about something hard is documented, and it passes. If it does not pass, or if thoughts of harming yourself appear, that is a reason to reach for help rather than a notebook: your local emergency number, or a helpline for your country at findahelpline.com.

How this works in Lanternly

We built the journal while reading the same papers, and a few decisions came straight out of them.

Mood is recorded as a point on two axes, how pleasant and how much energy, with no number anywhere. "Anxious" and "sad" sit in different places because they are different states, and the post about mood journaling without scores explains why a single scale blurs them.

In the morning Luna asks one question you can hold on to in one sentence, something like "what took up more room today than it deserved". The "See it differently" tool makes the move from feelings to thoughts that separated benefit from harm in the studies: first what you feel, then what you think about it, then how it looks from the outside. Luna writes the weekly overview in words, with no chart sliding downwards.

There is no counter of days in a row. Skip a week, come back, Luna is glad. And all of it stays on your device: what you write about your fears goes to no server, and why that matters is explained separately.

If you want to begin, start with one line: "I am afraid that…", with the day of the week and the place. That is enough for the anxiety to take up one sentence instead of the whole evening.