
Sara was six months into her PhD in Faisalabad, studying heat stress in dairy cattle, when her supervisor asked a simple question: what exactly did you feed the control group on 14 March? Sara opened her notebook and found loose numbers, a table with no units, and a sticky note that said "repeat??" in the margin. The feed trial had cost her lab six weeks, and she could not prove what she had done. A lab notebook is not a diary. It is the one document that proves your work happened, and this guide shows how to keep one that would survive any audit.
Most students write as if only they will ever read the notebook. Four other people may read it. Your supervisor reads it to check the work is real before signing a thesis. An examiner reads it when a result looks too neat, or when two chapters disagree. A journal editor may ask for the raw data behind a figure, and the notebook is where that data lives.
The notebook also matters outside the university. In the United States, the America Invents Act moved patent law to first-inventor-to-file in March 2013, but it is still evidence of when an idea was tested and what failed. The U.S. Food and Drug Administration's Good Laboratory Practice rules, 21 CFR 58.185, require study records to be kept and corrections made without obscuring the original entry.
The choice is not about which is more modern. It is about which one you will actually use every day, in the lab, with gloves on. A bound paper notebook works with wet hands, needs no battery, and cannot be lost to a software update. An electronic notebook is searchable, backs itself up, and lets your supervisor read your entries from another city. What matters is that you pick one and stay with it.
Ask your supervisor first, because some labs have a rule: paper only, or one named electronic system for everyone. If the choice is yours, test both for one week, then keep the one you reached for without thinking. One warning: never keep the record in a chat app or on loose sheets. Messages get deleted, and a loose sheet is gone the first time someone clears the bench. If it is not in the notebook, it did not happen.
Do this before your next experiment, not after it. Write your name, the project title, the lab name, and the start date on the inside cover, and number every page from one to the end. Then reserve the first two pages for an index: every time you start a new experiment, write one line — the date, a short title, and the page number. When Sara's supervisor asked about the 14 March feeding trial, she found it in ten seconds.
Decide your sample code system now, before the first sample exists. A code should carry the date, the project, and a running number, so a tube found in the freezer six months later tells you what it is. Sara's system was simple: FB for her Faisalabad cattle project, then the date, then a letter for the animal group and a number for the sample. A blood sample from the control group on 14 March 2026 became FB-2026-03-14-C01, and the matching feed sample became FB-2026-03-14-F01. Write the code key on the inside cover.
A good entry has five parts, in the same order every time: the date, the aim, the method, the results, and a short conclusion. Write the date in full at the top of every page — 14 March 2026, not 14/3, which means 3 December in some countries. Write the aim in one sentence: what are you testing today, and what would a good answer look like?
Here is a good entry, in the style Sara uses now. Date: 14 March 2026. Aim: test whether 200 g of bypass fat per cow per day raises milk yield in heat-stressed Sahiwal cows. Method: 12 cows in two groups of six; control on the standard ration, treatment plus 200 g bypass fat; feed weighed each morning at 7 a.m.; milk recorded at both milkings. Results: control 9.4 litres per cow per day; treatment 10.1 litres. Conclusion: a small rise, worth repeating with a larger group.
The bad version reads: 'Fed cows, milk ok, 9-10 L.' No date, no groups, no amounts, no conclusion. Three months later it told Sara nothing. The good entry took four more minutes to write, and saved her a six-week repeat.
The test for a method section is simple: could a new student in your lab repeat the experiment using only your notebook? Write the quantities with units, the equipment with its model and settings, and every reagent with its supplier and lot number. 'Added buffer' is useless. 'Added 5 mL of phosphate buffer, pH 7.4, Sigma P4417, lot SLCD2041' can be repeated, and a faulty lot number tells the lab which experiments are affected.
Write the settings you actually used, not the settings in the protocol, and write any change with its reason in the same entry. Usman, a lecturer in Karachi working on mastitis control, once spent two weeks chasing a result that vanished. His notebook showed why: the first run had used a different batch of culture medium, and he had not written the lot number down.
Never write a number without its unit, and never do the calculation later from memory. '9.4' means nothing; '9.4 litres per cow per day' means something. Write the raw reading first, then the calculation, then the result, all in the same entry. Sara's entry shows the working: six cows gave 60.6 litres in total, divided by six cows, which is 10.1 litres per cow per day. Bilal, a PhD scholar in Faisalabad, once fixed a mistake in his thesis data in an afternoon, because the raw readings were in his notebook and the error was one division done in his head.
A failed experiment is the most valuable entry in your notebook, because it stops you repeating the same mistake. Write what happened, what you think went wrong, and what you will change next time. Sara's entry for 21 March reads: 'Treatment group milk dropped to 8.2 litres. The bypass fat had clumped in the feed trough and the cows were leaving it. Next time, mix the fat into the concentrate the night before.' That entry explains a result that would otherwise look like failure.
Do not write 'did not work' and stop. Write the observation, the numbers, and your best guess at the cause, labelled as a guess. A guess written down can be tested; a shrug cannot. Examiners respect a thesis that reports failures honestly, and distrust one in which everything worked the first time, because in a real lab it never does.
You will write the wrong number. Everyone does. The rule for fixing it is one line: draw a single straight line through the wrong text so it can still be read, write the correction beside it, and add your initials and the date. The U.S. National Institutes of Health says the same: a record must show what was changed, when, and by whom.
Never erase, never use correction fluid, and never tear out a page. An erased number looks like a number changed after the result was known. If a whole page is wrong, draw one line across it, write 'cancelled, see page 47', and start again. The same rule applies to electronic notebooks: do not delete an entry — edit it, and let the system keep the change history.
A notebook entry that says 'see gel photo' without the photo has a hole in it. Glue or tape printouts into the book on the day, write the date and a one-line caption across the edge of the printout and onto the page, and sign across the join, so a printout cannot be swapped later. For photographs, write what is in the frame, the date, the scale, and the conditions. In an electronic notebook, attach the image file to the entry itself, not as a link to a folder, because folders get reorganised and links break.
Fleming's 1928 notebook shows why this matters. When a Penicillium mould contaminated his Staphylococcus culture, he recorded the observation, kept the plate, and photographed it. The notebook page and photograph are still held at the British Library. Record what you saw, keep the evidence with the record, and date both.
A notebook that exists in one copy is a notebook that can be lost. Paper books get left on buses, soaked in the lab, and borrowed and never returned. Pick one day a week and do three things: photograph or scan every new page, copy the images to a second place such as a university drive. For an electronic notebook, export your entries once a week as PDF and store the export where the lab can reach it.
Sara's lab lost a month of entries in 2025 when a free electronic notebook account was closed after the trial ended, and the only copy was inside the account. The lab now exports to PDF every Friday. Store the paper book in the lab, not at home, and when a book is full, write the end date on the cover and note the new book's number in the old index.
An examiner will not read your whole notebook. They will open it at random, or ask for the pages behind one figure, and judge the whole record by what they find. They look for three things: dates on every page, methods detailed enough to repeat, and results recorded on the day. A notebook that passes those three checks will survive almost any question.
Journals ask for the same things differently. When a reviewer questions a figure, the editor may ask for the underlying data. The UK Research Integrity Office advises that research records should be kept so the work can be verified: the notebook, the raw data, and the analysis must all be findable and linked. If your figure came from a spreadsheet, the spreadsheet should name the notebook pages the numbers came from.
Week one: set up the notebook, number the pages, write the index, and agree the sample code system with your supervisor. Week two: use the five-part entry for every experiment, do your first Friday review, and ask your supervisor to read two pages and tell you what is missing. Week three: add the backup habit — scan the week's pages, store the copy in a second place, and write a one-paragraph summary of the week. Week four: do a self-audit. Pick one figure and trace it back to the notebook: can you find the raw numbers, the method, the reagent lot, and the date, in under ten minutes? If you can, the system works. After a month the habit is set, and the notebook becomes the place you think.
| Feature | Bound paper notebook | Electronic notebook |
|---|---|---|
| Works with wet gloves | Yes, with a pen | No, needs a clean screen or keyboard |
| Search | Index only, by hand | Full-text search across all entries |
| Backup | Scan or photograph weekly | Automatic, but export a PDF weekly |
| Proving an entry was not changed | Ink, one-line corrections, numbered pages | Change history showing who edited what and when |
These authoritative external resources explain the standards and tools mentioned in this guide:
How long should I keep my lab notebook?
Keep it for at least ten years. The UK Medical Research Council advises retaining research data, including notebooks, for a minimum of ten years, and some funders and universities require longer. Check your institution's policy before you throw anything away, because a notebook you discard cannot be rebuilt when an examiner or a journal asks for it.
Can I use a loose-leaf binder instead of a bound notebook?
A bound notebook is safer, because pages cannot be added, removed, or reordered without leaving a trace. A loose-leaf binder invites doubt: a missing page looks like a missing experiment. If your lab uses printed sheets, glue or staple each one into a bound book, number the pages, and date every entry, so the record stays complete and in order.
What should I do if I make a mistake in an entry?
Draw one straight line through the wrong text so it stays readable, write the correction beside it, then add your initials and the date. Never erase, never use correction fluid, and never tear a page out. A visible correction proves the record is honest; a hidden one looks like the data was changed after the result was known.
Is an electronic lab notebook acceptable for a thesis or publication?
Yes, if it records the same things a paper book does: dated entries, the full method, raw results, and a change history that shows who edited what and when. Most universities accept established systems such as LabArchives, Benchling, or RSpace. Keep a regular export or backup outside the system, so your record survives if the subscription ends.
How much detail should I write in the notebook each day?
Write enough for a colleague to repeat the experiment without asking you anything. That means the aim, the exact method, every reagent with its lot number, the instrument settings, the raw results, and what you concluded. A short entry is fine; a vague entry is never fine, because the detail you skip is the one you will need later.
What happens to my notebook when I leave the lab?
The notebook normally belongs to the institution or the research group, not to you, because it is the primary record of work the lab funded and supervised. Hand it to your supervisor or lab manager, keep a copy for yourself with permission, and note the handover in your final entry so nobody argues later about where the record went.
Submit your manuscript ← Back to GAP Insights
Useful publishers, databases, funders, tools and institutions for researchers: