How To Journal, Using The Protocols From The Research

Almost every guide to how to journal is written from the author's own habit. Buy a notebook you like, pick a time, don't judge yourself, keep going. That's fine advice and it's also unfalsifiable, which is why every one of those pages says roughly the same thing in a different font.

Here's the direct version. The researchers who study writing about your life didn't hand people a blank notebook and good intentions. They handed them a protocol with a word count, a clock, and a specific paragraph of instructions, and those protocols are published, and you can just copy them. The most-used one is three or four sessions of fifteen to twenty minutes each, spaced a day or two apart, with an instruction that tells you to stop worrying about spelling. And the single most interesting finding I came across while writing this is that in one large trial the instructions only mattered for the people who wrote a lot, with the cutoff sitting at 1,722 words across three sessions.

So this page is the mechanics. Not what to write about, which is a separate problem with its own prompt library, but how the sitting-down part is actually structured when somebody is measuring the result.

The Four Protocols, Side By Side

I went and pulled the actual written procedures rather than describing them from memory. Four of them, all still online, all with their session structure specified.

Sources for that table, in order, are the 2023 Frontiers in Psychology trial, the same paper, the VA's clinician-facing page on Written Exposure Therapy, and the JMIR Mental Health randomised trial.

Two things jump out at me. First, none of them are daily. The heaviest one is five sessions total. Second, the shortest session in the set is fifteen minutes, which is roughly three times what most people picture when they picture journaling.

The Bit About Volume That Nobody Mentions

Susan Rude and three colleagues ran 833 undergraduates through three fifteen-minute writing sessions, spaced roughly a day or two apart, and compared standard instructions against an acceptance-flavoured version against a control (Frontiers in Psychology, 14, 1192595).

The headline result was nothing. No significant difference between conditions, p equals 0.21. If you stopped reading there you'd conclude that the instructions don't matter, and honestly that was my first read too.

Then they looked at how much people wrote. Across the three sessions the average essay totals were 1,520 words in one condition, 1,576 in another, 1,230 in the third. And there was a significant interaction between condition and essay length, F(2, 826) equals 4.12, p equals 0.017, with an eta squared of 0.010. The contrast they report is that the acceptance-enhanced condition beat control only for people who wrote 1,722 words or more, which was the top 29 percent of writers.

Eta squared of 0.010 is a small effect and I'm not going to pretend otherwise. It's one trial, the moderator is essay length rather than something the researchers assigned, and people who write more are probably different from people who write less in a dozen ways nobody measured. Engagement could be doing all the work.

But 1,722 words across three fifteen-minute sittings is about 574 words a session, which is a real, checkable target, and it's the only number I've seen anywhere that tells you what "enough" looks like on a page. Roughly a page and a half of normal handwriting. Twice, then a third time.

Your First Four Sessions, Concretely

Copying a research protocol is a slightly odd thing to do at a kitchen table, so here's what it actually looks like when you strip out the lab.

Session one. Fifteen minutes, timer on, somewhere you won't be interrupted. Pick the thing that's currently sitting on you. Write about the event and how you feel about it, in whatever order it comes out, and do not fix the spelling. If you stop before the timer, write the last sentence again and keep going from there. That trick isn't from any paper, it's just what works when a page stalls.

Session two, a day or two later. Same length, same subject, and this is the one people skip. The protocols repeat the topic on purpose. You're not meant to move on.

Session three. Same again, and this time somewhere in the fifteen minutes try to say what the thing means now, as opposed to what happened.

Session four is optional and it's mine, not theirs. Read the three back in one sitting and underline the sentences you don't remember writing. Those are usually the ones worth having.

That's it. Three sessions is where most of the evidence sits, spacing of a day or two is what the trials used, and if you want the version with a much longer runway there's a whole separate approach for building self-awareness over months rather than days.

The Total Time Bill, Added Up

Nobody adds these up, so I did. Multiply the sessions by the minutes and you get the whole cost of each protocol, start to finish.

That bottom row isn't from any study. I put it there because it's the format most people mean when they say they want to start journaling, and it costs more than thirty hours a year, which is over three times the writing time of the heaviest protocol above and roughly forty times the classic paradigm.

Forty times. For a practice with a thinner evidence base than the forty-five-minute version.

I'm not saying the daily habit is worthless, and plenty of people who do it would tell you it's the only thing keeping the week in order. I'm saying the accounting is worth seeing once, because "just five minutes" is the phrase that sells the most journals and it's doing a lot of hiding. Forty-five minutes, spread over four days, is the version with the most evidence attached, and it fits in one weekend.

Where This Stops Being The Right Tool

Say this plainly, because the mechanics above are the mechanics of an experiment, not of care.

A journal is not treatment, and I'd be doing you a disservice if the tidy table above implied otherwise. If what you're writing about is heavier than a page can hold, the useful move is a qualified professional rather than a fourth session, and if you're in immediate crisis then a crisis line is the right call and this article isn't. In the US that's 988, which takes calls, texts and chats every hour of every day, and it's explicitly for emotional distress rather than only for people who are actively suicidal. Elsewhere, your national line exists and is staffed tonight.

Notice also that one of the four protocols in my table is a clinical treatment delivered by a therapist across five sessions. Written Exposure Therapy is not a thing you administer to yourself out of curiosity, and the VA page describes the clinician giving feedback on the previous session's writing before each new one, which is a component you cannot supply on your own.

The normal, expected reaction to a session that goes deep is feeling worse for a while afterwards. Short of it. If that "while" is turning into whole evenings, that's information, and the response is to stop and talk to someone rather than to write harder.

What To Do About The Notebook Question

Nothing. Buy any notebook. Or don't.

I say this as somebody who has watched people spend more energy choosing a system than using one, and I include myself, on entirely different projects. I've shipped seven paid toolkits and I can tell you the setup phase is always the most enjoyable and the least productive part of anything.

The relevant evidence on handwriting versus typing is in the pillar's moderator table, and the short version is that it made no measurable difference. The positive affect journaling trial ran entirely online. Written Exposure Therapy uses paper in a clinic room. Both work in their own contexts.

What I'd actually spend the decision on is where. Two of the protocols specify a private setting, and the one thing I'd protect is the certainty that nobody's reading it, because the moment you're writing for a possible audience you start editing, and the edited version is the version you already believe.

The Questions That Come Up Every Time

Do I have to write every day? No, and nothing in the four protocols above is daily. Positive affect journaling ran three times a week for twelve weeks, which is the most frequent design in the set, and the classic paradigm is three or four sessions in total. The daily framing comes from habit apps and the morning routine genre, not from the research.

What if I've got nothing to say? Then write the plainest possible account of the last two hours until something interesting turns up, which it usually does around minute four. If you want the version of this that starts you with half a sentence instead of a question, entry starters are built for exactly that stall.

How long should an entry be? On the one number I could find, about 570 words a session. That's more than most people write. I'd treat it as a target rather than a rule, and I'd rather give you a real number from a real trial than the usual "as long as feels right".

Should I write about good things or bad things? Both are studied. The classic paradigm is about difficult events, positive affect journaling is deliberately the opposite, and the 2018 trial found benefits from the positive version in people who were anxious to begin with. Anna Freud's own self-care page for young people quotes a young person saying "Don't feel that you have to write loads," which is decent counterweight to my 570-word figure and I'm including it on purpose.

Do I reread it? I don't know, and I looked. My instinct is that rereading last week is riskier than rereading last year, but that's a hunch and it isn't in any of these papers.

What I'd Skip Entirely

I should say I don't keep a daily journal, so treat this section as judgement rather than practice.

I'd skip the prompts-per-day calendar. I'd skip the streak. I'd skip any product that promises the whole thing takes five minutes, because the shortest protocol in the research is fifteen, though I've written a longer piece on whether the five-minute format holds up that's fairer to it than that sentence sounds.

And I'd skip the idea that you're doing it wrong. The largest trial in this article found no difference between two carefully worded instruction sets. If professional psychologists can rewrite the instructions and get nothing, the exact phrasing of your own first sentence is not the variable that decides whether this works for you.