Journal Entry Prompts Sorted By Entry Type

Most journal entry prompts fail for a reason that has nothing to do with the prompt. They fail because you don't know what you're trying to produce, so a question that would make a perfect two-minute log entry gets attempted as a soul-searching session at eleven at night, stalls out at four lines, and gets filed as evidence that journaling isn't for you.

Short answer up front. There are four kinds of entry worth writing, they have different structures and different lengths, and the prompt only works if it matches the type. A log entry is two minutes and answers what happened. A processing entry is fifteen and answers what's actually going on. A decision entry is ten to twenty and produces a written commitment. A record entry is however long it takes and exists entirely for the version of you reading it in five years. Below is a prompt set for each, plus the thing almost nobody covers, which is how you know an entry is finished.

Four Entries Worth Writing, And The One That Isn't

Before the prompts, the map.

The fifth kind, the one I'd skip, is the venting entry with no shape. Pure discharge onto the page, same content each time, no sequence and no question. It feels productive because it's emotionally loud, and it's the closest journaling gets to just rehearsing the thing. Some people find it helps and I'm not going to tell them they're wrong. I'd just say it's the entry type with the weakest case behind it.

Log Entries

The most undervalued of the four, and the one that survives bad weeks.

Log prompts have to be answerable when you have nothing. That's the entire design constraint. If a prompt requires a feeling or an insight, it isn't a log prompt.

• What happened today, in order, five bullets maximum?
• Who did I speak to and about what?
• What did I finish, and what did I move to tomorrow?
• What did today cost me, in money or hours or both?
• What's the state of the thing I'm currently worried about, in one line?
• Where did I actually spend the afternoon?
• What did I eat, sleep and do, roughly?

None of that will feel meaningful while you're writing it. That's expected. The value shows up at volume, when six months of two-minute entries turn into something you can search, and you find out that the fortnight you remember as a disaster contained four good days you'd completely forgotten and one specific conversation that was actually the problem.

If even five bullets is too much friction, the sentence-stem approach to entry starters drops the barrier further.

Processing Entries

This is the one people mean when they say journaling, and it's the one with an actual research literature behind it.

The research is on expressive writing, where participants typically wrote for 15 to 20 minutes across three to five sessions about something emotionally significant. The instruction in the classic version is close to "write your very deepest thoughts and feelings about the most traumatic experience of your entire life or an extremely important emotional issue," as summarised in Baikie and Wilhelm's clinical review. Which is a heavier instruction than most people expect and explains why the sessions ran that long.

Prompts that hold fifteen minutes:

• What's the situation, described with no adjectives, in the order it happened?
• What do I know for certain here, and what have I assumed without checking?
• What was I expecting instead, and where did that expectation come from?
• Which part of this is about the event and which part is about what it says about me?
• What would have to be true for the other person's version to make sense?
• What have I already tried, and what actually happened when I did?
• What am I protecting by not resolving this?
• If a friend brought me this exact situation, what's the first question I'd ask them?

Keep one topic per entry. The temptation is to do a sweep of everything currently unresolved, and a sweep produces a list rather than an entry.

One honest caveat about this type specifically. Writing usually feels worse immediately afterwards rather than better, which the review describes as a short-term rise in distress and low mood that doesn't appear to carry longer-term risk. If it's consistently still with you hours later, that's not a prompt problem and a journal isn't the right tool for it. Talk to someone qualified. Nothing on this page is a substitute for that.

Decision Entries

Different shape entirely, and the only entry type with an output you can check.

A decision entry isn't finished when you feel clearer. It's finished when there's a sentence on the page that starts "I'm going to" and has a date attached. Everything before that sentence is scaffolding.

• What am I choosing between, stated as two options in plain sentences?
• What's the actual deadline, and who set it?
• What do I already suspect the answer is?
• What's the cheapest reversible version of the option I'm avoiding?
• What information would change my answer, and can I get it this week?
• What would I advise someone else with these exact facts?
• What am I going to do, by when?

That fourth one does most of the work. A very large share of stuck decisions aren't actually between two options, they're between committing hard to something and doing nothing, and there's nearly always a small reversible version sitting in between that you haven't considered because it isn't dramatic enough to feel like a decision.

Date these entries. Write down what you expect to happen. Then, and this is the part that stings, reread them a year later, because a prediction you never recorded is a prediction you get to quietly rewrite in hindsight.

Record Entries

For the reader who isn't you yet.

There's a finding in the large meta-analysis of writing studies that argues for writing close to events rather than long after. The average participant was writing about something that had happened around fifteen months earlier, and the further back it was, the smaller the measured effect, with a correlation of r equals -.283 overall (Frattaroli, Psychological Bulletin, 2006). I'd read that as a mild argument for promptness and a strong argument that a fifteen-month-old memory is not a source document.

• What does this room, this street, this desk look like right now?
• What are the prices of ordinary things this month?
• Who's in my life this week and how often do I see them?
• What do I believe is about to happen?
• What's the thing everyone around me currently assumes?
• What's a normal Tuesday like at the moment?
• What am I taking completely for granted?

The last one is nearly impossible to answer honestly and worth attempting anyway.

The Same Tuesday, Written Four Ways

An example helps more than another list, so here's one ordinary day through each type. I've kept it deliberately mundane because the mundane day is the one you'll actually be facing.

Log. "Client call moved to Thursday. Spent morning on the invoice thing, still not resolved. Walked to the shop, ran into D, first time in months. Bad sleep, maybe five hours. Didn't start the report."

Processing. "The invoice thing has taken four days and it isn't a four-day problem. What actually happened is that I sent it, got no reply, and have since checked my email roughly thirty times without following up once. What am I protecting by not following up. I think it's that as long as I haven't chased it, they haven't refused..."

Decision. "Choosing between chasing the invoice now and waiting until after the Thursday call. Deadline is my rent, which is the 1st. Cheapest reversible version is a two-line message today that doesn't mention money, just checks the file arrived. I'm going to send that by 6pm."

Record. "Rent's due on the 1st and it's 720 a month at the moment. D looked well, has moved out past the ring road, says the commute is fine which I don't believe. The shop has started closing at seven. I keep assuming this arrangement is temporary."

Four entries, same day, none of them interchangeable. The log takes ninety seconds and the processing one takes fifteen minutes, and if you attempt the second at the length of the first you'll produce four lines of vague dissatisfaction and conclude that journaling doesn't work for you.

How You Know An Entry Is Finished

Nobody answers this and it's the question I'd have wanted answered first.

For log entries, when the events are down. There's no insight target and looking for one turns a two-minute habit into an eight-minute chore that you'll drop by week three.

For processing entries, when you've written a sentence you didn't already know before you sat down. That's the actual signal. If you finish twenty minutes and everything on the page is something you could have said out loud beforehand, you transcribed rather than processed, and the usual cause is that the prompt was too broad.

For decision entries, when there's a commitment with a date.

For record entries, when the concrete details are down and you've resisted explaining how you felt about them.

And a general one across all four. If you're rereading your own paragraph as you write it to check whether it sounds good, you've slipped into writing for an audience, and the privacy finding from the meta-analysis suggests that matters more than most people think. Studies where participants wrote alone in a private room showed larger effects than those where they wrote alongside others, and studies conducted at home outperformed lab settings on psychological health measures. Being genuinely unobserved seems to be part of the mechanism.

What Goes Wrong When One Entry Does Two Jobs

Mixing types is the most common failure I'd expect, and it's invisible while it's happening.

The usual version is a processing entry that turns into a decision entry halfway through, which means you commit to something while still in the middle of feeling strongly about it. The other version is a decision entry that collapses back into processing, where you write six hundred words about the situation and never get to a sentence with a verb in it. Both feel like a full entry. Neither produces the output the type is for.

Rule I'd follow, and I'll admit it's a preference rather than something I can point to a study for. One entry, one job. If a processing entry starts producing decisions, write the decision entry tomorrow, when the material is available and the heat isn't.

The Honest Ceiling On All Of This

Structure improves the odds. It does not make the effect large.

The pooled average across those 146 randomised studies was r equals .075, with 36 of them landing on a negative effect, and the depression and anxiety subcategories came in at .044 and .028 respectively. There's also a 2022 study of 152 undergraduates using a prompted journaling app for 28 days that found the wellbeing benefit only appeared above a threshold of dispositional self-reflection, and the threshold the paper actually reports sits a little under one standard deviation below that group's average rather than at it (Behaviour Change). Prompts didn't manufacture reflectiveness in the people lowest on it.

So the fair claim is that matching the entry type to the prompt makes the habit stickier and the output more useful, and that the mental-health effect underneath it all is modest on average. I'd still do it. I just wouldn't sell it as more than that.

The broader theory of why some prompt wordings work better than others is in the main journal prompts guide, and if what you're really after is the self-knowledge angle rather than the daily-practice angle, self-discovery journaling and journaling for self-awareness take it from there.