A Research Journal for PhD Students: Two Separate Records

2026-09-13 · Yesterself

Most doctoral programs assume you are keeping a notebook. Few of them say which one. There is the record of the work — what you ran, on what data, with which settings, and what came out. And there is the record of the person doing the work for four, six, sometimes eight years. Those are two different documents with two different readers, and the common mistake is to keep one book and hope it covers both.

A research journal for PhD students works better as two habits than one. This piece is about where to draw the line, and what each side is good for when you sit down to write up.

Why one notebook does both jobs badly

There are two failure modes, and they are mirror images.

If the personal material lives inside the methodological record, you censor it. Your supervisor might open that file. It may be attached to a data management plan, inherited by a collaborator, or shared with a lab you join later. So the entry that should say I have not understood this method for three weeks and I am afraid to admit it comes out as "revisit assumptions." The sentence you most needed to write is the one you edit away.

If the methodological material lives inside the personal journal, you lose it. Nobody goes looking through a diary for a random seed or a reagent lot number. The details that matter in year five are buried under paragraphs about a difficult meeting in year two.

The tones are incompatible too. The methodological record wants to be terse, dated, and boring on purpose. The personal record wants room to be uncertain. One file cannot do both.

The methodological record: what you tried, why, and what failed

Keep this one per work session, not per day and not per week. A usable entry answers six things:

• The question this session was supposed to answer.
• Exactly what you did — versions, parameter values, commit hash, sample or participant IDs, instrument settings, the date in an unambiguous format.
• What you expected before you looked. Write this first. It costs ten seconds, and it is what lets you tell later whether a result actually surprised you.
• What happened, described separately from what you think it means.
• What you concluded, marked plainly as a conclusion rather than an observation.
• What you will do next.

Then the part almost nobody keeps: the dead ends. Write down what you abandoned, the date, and why — in enough detail that a year later you can judge whether the reason was good. Some get dropped for real reasons and some because you were tired that week, and without the record you cannot tell them apart. Committees ask about them. So do reviewers. So do you, at two in the morning in year four, when an idea arrives that feels new and is not.

The personal record: the part of the project that is you

No methods training covers this one, and it holds a lot of what decides whether the thesis gets finished.

Three subjects come up again and again.

Isolation. A doctorate narrows your world to a handful of people who understand what you do, and sometimes to none. Write down the specific shape of it — which days were empty, who you did not hear from, what you stopped mentioning to friends because explaining it was too tiring. On paper it has dates and causes, some of which you can change. Unwritten, it just feels like a personality.

Friction with a supervisor. Write facts and dates, not verdicts. What was said, what was agreed, what was submitted and when, how long a draft sat unread. A record of facts is useful if a disagreement ever becomes formal, while a record of adjectives is not — and months later it often shows a pattern duller and more fixable than the story you were telling yourself. This is not legal advice. If you are moving toward a formal complaint or a change of supervisor, talk to your graduate school, your ombudsperson, or a student representative who knows your institution's procedures.

Motivation collapse. It usually arrives somewhere in the middle, and it does not feel like a phase while you are inside it. What helps is having last year's version of it on paper — the same flatness, the same certainty that it was permanent, and the dated record of what happened next.

None of this is treatment, and this article is not medical advice. If low mood is persistent, or you are struggling to function, speak with a doctor or your university's counseling service.

One thing the personal record should not be: a performance review of yourself. If every entry ends in a verdict on your worth, the record stops being useful and becomes another way to lose the evening. Describe the week. Leave the sentencing out.

A brief note on tools. The personal record only pays off if you reread it, which is the part most people skip. Yesterself is a journaling app built around that: each morning it mails you one letter drawn from your own past writing — yesterday's entry, an on-this-day one, or an older piece that touches what you have been writing lately — and your reply that evening becomes tomorrow's mail. You can also postmark a sentence to a chosen future date, which is worth doing the week before an upgrade panel or a defense. It is end-to-end encrypted and works without an account, an email address, or a password; your identity is twelve words you hold. Export is free forever, in open formats, on the free tier too. It is not released yet.

At write-up, the two records do different jobs

The methodological record becomes text almost directly. Your methods chapter is mostly transcription. Your limitations section is the honest version of the entries where something did not work. Reviewer questions — why did you not try X? — get a date, a description, and a result instead of a shrug. Viva preparation stops being an act of memory.

The personal record does something else. It gives you back the reason. In month four you knew why the question mattered; by month forty you can recite the literature and no longer feel the point. The introduction is a story about why this was worth doing, and the person who could tell that story is in your old entries. The acknowledgments come from there too — who helped, in the specific week they helped.

It also answers a question the methodological record cannot: whether a bad month is a bad month or a signal. Some people should leave a program and some should not, and nobody can tell which from inside a single week. Twelve months of dated entries make a better basis for that decision than a mood on a Tuesday.

Making both findable

Capture is the easy half. Retrieval is the half that fails.

• Date every entry, in a format that does not change.
• Tag methodological entries by research question or chapter, not by date alone. In year four you will search by topic.
• Keep both records in plain, exportable formats. Software you cannot export from is a liability across a project this long.
• Write a short monthly index: three lines on what moved, three on what stalled. It is the cheapest thing here and the most read later.

Where the two records touch

Some decisions are both. Switching instruments, dropping a study arm, changing supervisors — each has a methodological justification and a personal cost, and the two get remembered as one blurry thing. A separate decision log, holding the reasoning at the moment of choosing, keeps them straight. The same split shows up in software as the architecture decision record versus a private log, covered in decision records versus a personal decision log.

Start with the methodological record, because it is mechanical and pays off within weeks. Add the personal one when you can. A gap of a week or a month in either is not a failure; the record is there to be resumed, not maintained perfectly. What you want at the end is two documents: one that proves what you did, and one that remembers who did it.

Common questions

Should my lab notebook and my personal research journal be the same file?

No. The methodological record is written to be read by others and should stay terse and factual, while the personal record only works if you can be honest in it. Keeping them together means you either censor the personal material or lose the methodological detail.

My institution requires a specific lab notebook format. Where does the personal journal go?

Follow the required format for the methodological record and keep the personal one entirely separate, in something you control and can export. Check your program's data management and record-retention policy so you know what the official notebook may be shared with or subject to.

Is it wise to write about conflict with my supervisor?

Write facts and dates rather than judgments, and keep that record private. A factual log is more useful to you later, whether the situation resolves quietly or becomes formal. This is not legal advice; for a formal process, talk to your graduate school, ombudsperson, or student representative.

How often should I write in a research journal?

Write the methodological entry at the end of each work session, while the details are still exact. The personal record can be irregular, and a missed week or month is not a failure to correct.

Does a research journal actually help with writing the thesis?

The methodological record supplies most of the methods chapter, the limitations section, and your answers to questions about approaches you dropped. The personal record supplies the reason the question mattered, which is what the introduction is made of.