How to Write a Lab Notebook That Survives an Audit
Knowing how to write a lab notebook is mostly an evidence problem, not a writing one. A lab notebook is the systematic, tamper-evident record of a research project — what was done, by whom, on what date, under what conditions, and with what result — kept so that a third party can reconstruct the work years later, on paper or in electronic form.
In Korea’s national R&D programs, that record is not a nice-to-have. The Guideline on Lab Notebooks for National R&D Programs (‘국가연구개발사업 연구노트 지침’, MSIT Notice No. 2021-102, in force since January 2022) is the rulebook that institutions must build their own notebook regulations around — and it is what a verification team leans on when they ask to see your notes. This guide walks through what the guideline actually requires: what to record, how to preserve it, who confirms it, and what a compliant lab notebook format looks like in paper and electronic form.
What Counts as a Lab Notebook?
The guideline’s definition is broader than most people expect: a lab notebook is “a document systematically recording the information, data, and know-how obtained while performing an R&D project.” It does not have to be a bound paper book. The guideline explicitly requires institutions to allow notebooks “in various formats — paper, electronic notes, audio, or video.” What makes a record a lab notebook is not the medium but the discipline: dated, attributed, and resistant to alteration.
Two boundary cases are worth knowing. For certain project types — preliminary surveys, planning evaluations, humanities and social sciences research, workforce training, infrastructure build-out — and for institutions that genuinely cannot manage notebooks, such as solo researchers and early-stage startups, the guideline allows the annual or final report to count as the notebook (Art. 8(3)). And when several institutions share one project, each institution keeps its own notebook as a rule, whether researchers write individual notebooks or share one jointly is up to the institution’s own regulation.
What Should Every Lab Notebook Entry Contain?
The core requirement is threefold, and everything else follows from it: the institution must be able to verify the date of the record, who recorded it, and whether it has been altered (Art. 7(2)). In practice, a defensible entry looks like this:
- The date — written when the work happens, not reconstructed afterward. A notebook backfilled before an audit fails the one test it exists to pass.
- The recorder — who ran the work and wrote the entry, plus who else participated. The guideline names these roles explicitly: the recorder writes, and a separate confirmer reviews and signs.
- Conditions — materials, equipment, parameters, environments: enough that someone else could rerun the work.
- Results as they came out— including the failures and dead ends. The guideline’s standard for content is that the recorder writes “objective facts without falsification or alteration, so that a third party can use them to reproduce the research process and its results” (Art. 8(6)). A notebook with only the successes is a brochure, not evidence.
- A visible correction history — corrections are allowed; invisible corrections are not. On paper that means crossing out, never erasing. Electronically it means an edit history that survives.
The reason to be this strict is not bureaucracy. When a result is disputed — a misconduct allegation, a patent conflict, a question about whether a grant produced what it claimed — the notebook is the only witness that was there on the day. The alternative is knowledge that lives in one person’s head, which is exactly the tribal knowledge tax that sinks teams when that person is unavailable.
How Long Is a Lab Notebook Preserved — and by Whom?
Two answers surprise people. First, ownership: the notebook belongs to the institution that managed its creation, not to the researcher who wrote it (Art. 9). When you move on, the record stays and the institution must keep managing it — its regulation is supposed to cover what happens to notebooks when research is interrupted or a researcher leaves.
Second, duration: the default retention period is 30 years from the project end date (Art. 10(3)), though an institution may set different periods by project type in its own regulation. Thirty years is longer than most careers, and that is the point — the record is meant to outlive every individual who could be asked about it. Alongside retention, the institution must run an access system: a reading log, defined disclosure rules, and a guarantee that the recorder can always read their own notes (Art. 11).
Who Verifies a Lab Notebook?
Verification happens at three levels. Day to day, the confirmer— as the guideline defines it, “a person who confirms the contents of the lab notebook and signs it” — reviews entries and countersigns. Institution-wide, the head of the institution must establish and operate a notebook regulation covering creation, storage, and management (Art. 5), and is explicitly told to support researchers with training and tools rather than use notebooks to control them. And from outside, the responsible ministry can audit whether an institution’s notebook system actually works, not just whether it exists on paper (Art. 13).
Note what verifiers look for: not beautiful prose, but the three properties from Art. 7 — date, recorder, tamper-evidence — applied consistently. A messy notebook with an unbroken chain of dated, signed entries beats an immaculate one nobody signed.
Does an Electronic Lab Notebook Format Count the Same as Paper?
Yes — and this is the part of the lab notebook format question that matters most in practice. The guideline puts electronic notes on exactly the same footing as paper, and applies the same three tests. What changes is how you meet them: instead of bound pages and ink, an electronic lab notebook format needs system timestamps, per-user attribution, access control, an alteration-proof audit trail, and backups. A shared wiki or a folder of word-processor files meets almost none of these — no reliable attribution, no preserved edit history, no guarantee the file today is the file that was written then.
A Lab Notebook Checklist You Can Use Today
- Every entry carries the date it was written and the name of the recorder.
- Participants, materials, equipment, and parameters are recorded per experiment.
- Results are entered as they come out — failures included.
- Corrections are visible; nothing is deleted or overwritten invisibly.
- A confirmer reviews and signs entries on a regular cadence.
- The format (paper, electronic, audio, video) is one your institution’s regulation allows.
- An electronic system keeps timestamps, attribution, access control, and an audit trail.
- Notebooks are retained for the set period — 30 years by default — under the institution’s management, with a reading log.
- There is a plan for what happens to the record when a researcher leaves or the project stops.
FAQ
Can I keep my lab notebook in a wiki or a cloud document folder?
Usually not. The guideline’s tests — recording date, recorder identity, and detection of alteration — are exactly what generic wikis and shared folders lack. Unless the tool preserves attribution and a tamper-evident history, the record will not hold up when it is checked.
How long do lab notebooks have to be kept?
Thirty years from the end of the project by default, with institutions allowed to set different periods by project type in their own regulations. In practice, assume the record must outlive everyone on the team.
Who owns the lab notebook — the researcher or the institution?
The institution that managed the notebook’s creation owns it. The recorder keeps the right to read their own notes and to request their use for protecting themselves, filing patents, or defending results — but the copy on file is institutional property.
Do all projects have to keep one?
Most national R&D projects do, but the guideline allows exceptions — for project types where notebooks add little (surveys, planning work, training programs) and for institutions that cannot realistically manage them, the annual or final report can count as the record. That exception is narrower than teams hope; check your institution’s regulation before relying on it.
If you would rather not assemble those properties from shared folders and trust, that is the standard Rootr was built around — records that are attributable, time-stamped, tamper-evident, and checkable against live data, for research and engineering teams alike. Pricing has the plan breakdown.