# ALCOA+ in Practice: A Lab Notebook Audit Checklist for Data Integrity By the end of this article you will be able to audit any [lab notebook](/knowledge/molecular-biology/lab-notebook), paper or electronic, against the nine ALCOA+ attributes, score what you find, write findings that a colleague can act on, and fix the documentation habits that cause most audit findings. You need the notebook itself, a printed or digital copy of the checklist in this article, a pen, and optionally a CSV export from an electronic lab notebook (ELN) plus a spreadsheet program or Python if you want to audit timestamps at scale. ALCOA+ is a data integrity framework that grew out of regulatory expectations for records in regulated laboratory work. The original ALCOA term stands for Attributable, Legible, Contemporaneous, Original, and Accurate. The plus adds Complete, Consistent, Enduring, and Available. Regulators and quality professionals use these attributes to judge whether a record can be trusted, and the same logic applies to academic research notebooks, even when nobody is planning a formal inspection. ## What ALCOA+ Means in a Lab Notebook Each attribute describes a property of a record. In a notebook, each property translates into something you can physically check. **Attributable** means every entry identifies who did the work and when. In a notebook this means a full name or unique identifier with each entry, plus initials and a date on every page. **Legible** means a reader years from now can read the entry without guessing. It also means readable by someone who was not there when the entry was written. **Contemporaneous** means the record was created at the time of the activity, not reconstructed that evening or reconstructed from memory a week later. **Original** means the entry is the first capture of the data, or that the true original (an instrument printout, a raw file) is preserved and traceable. **Accurate** means the record reflects what actually happened, including errors, deviations, and failed runs. **Complete** means nothing is missing: all data, all metadata, all repeat measurements, and all deviations. **Consistent** means the record follows an expected chronological order and consistent formats across the notebook. **Enduring** means the record lasts. Permanent ink on paper that survives storage, or a durable digital record, not pencil or sticky notes. **Available** means the record can be retrieved for review, audit, or reuse during its entire retention period. A notebook audit is simply a systematic check of these nine properties, page by page or entry by entry, with written findings. ## Before You Start: Scope, Sampling, and Ground Rules An unbounded audit fails before it starts. Set three things first. 1. **Scope.** Decide which notebooks, which date ranges, and which record types (experiment entries, calibration logs, instrument printouts, meeting notes) fall inside the audit. 2. **Sample size.** For routine self-audits, a practical approach is to audit every entry in a fixed sample, for example the most recent 20 entries plus a random sample of 10 older entries. Smaller notebooks can be audited in full. Write the sampling plan down before you open the notebook, so nobody can accuse the audit of cherry-picking. 3. **Blindness to blame.** The audit evaluates records, not people. Findings should describe the record ("entry 14 has no date") rather than the person ("the postdoc failed again"). This framing keeps audits honest and keeps people from hiding problems. Agree on a scoring method before you start. A simple binary pass per attribute per entry works well, with a notes column. A more informative method scores each attribute as 2 (meets), 1 (minor gap, correctable), or 0 (gap that cannot be repaired from the record itself). ## The ALCOA+ Checklist, Step by Step Work through the following checks for each sampled entry. Record the entry identifier, the attribute, pass or fail, and a one-line finding. The worked example later in this article shows what a filled score sheet looks like. ### Step 1: Check attribution on every entry Look for the name (or unique user identifier) of the person who performed the work, the date the work was performed, and initials or a signature near the entry. In shared notebooks, check that two people never write in the same entry without both being named. **Red flags:** unsigned pages, a date with no name, a single signature covering several days of work, initials that do not match any name recorded anywhere in the front matter. ### Step 2: Check legibility Read the entry as if you had never seen the project. Can you read reagent names, concentrations, and sample identifiers without context? Crossed-out text should still be readable, because a single-line strike-through preserves the original record. **Red flags:** entries written over or taped over other entries, correction fluid, faded thermal printer paper, handwriting that hides numbers, abbreviations that appear nowhere in the notebook's abbreviation list. ### Step 3: Check that entries are contemporaneous Compare the date of each entry with the sequence of entries around it. Look for signs of backfilling: a block of entries all dated the same day describing work spread over a week, entries written in pencil and later inked over, or gaps followed by a cluster of "catch-up" pages. **Red flags:** past-dated entries written on visibly later pages, "recall" style summaries that lump several experiments into one undated block, dates that contradict instrument timestamps on taped-in printouts. ### Step 4: Check that records are original Raw data should appear in the notebook first or the true original must be preserved. A handwritten transcription of an instrument readout is acceptable only if the printout or raw file is attached, referenced, and retained. Transcribing "clean" values into the notebook and discarding the raw readout destroys the original record. **Red flags:** printouts missing from pockets or taped pages that reference them, transcription notes where the raw file is gone, spreadsheet values pasted into the notebook with no file name or version reference. ### Step 5: Check accuracy, especially corrections Corrections must preserve the original entry. The standard method is a single line through the error, the correct value written nearby, and the initials and date of the person making the correction. The old value must remain readable. **Red flags:** erasures, correction fluid, overwritten digits, corrections with no initials, values changed without any explanation of why. ### Step 6: Check completeness Every experiment entry should let a reader repeat the work: what was done, with what materials, at what settings, with what result, and what went wrong if anything. Blank fields and skipped pages need a reason. In regulated work, a blank field is filled with "N/A" or a diagonal line so nobody can add content later. **Red flags:** empty pages between dated entries, missing negative results or failed runs, deviations from the protocol that are described in email but not in the notebook, sample identifiers that appear in one experiment but not the follow-up. ### Step 7: Check consistency Scan the whole sampled range for chronological order, a single date format, consistent units, and consistent naming of the same sample across entries. Missing page numbers are a classic consistency failure in bound notebooks. **Red flags:** pages out of sequence, two date formats mixed (05/06/2026 read two ways), the same buffer called three different names, unnumbered pages in a bound notebook. ### Step 8: Check endurance Look at the physical medium. Pencil, water-soluble ink, and glued-in sticky notes fail the endurance test. Tape should cover only the edges of attached material, never the writing. Electronic records need a stable format and a controlled storage location, not a personal desktop folder. **Red flags:** pencil entries, gel pen ink that smears, sticky notes as the only record of a calculation, printouts on thermal paper that has already faded. ### Step 9: Check availability Ask the practical question: could someone else find and read this record today, and in five years? For paper, check that the notebook is stored where the lab says it should be and that an index exists. For electronic records, check that the file opens, that the storage location matches the lab's stated policy, and that backups are actually being taken. **Red flags:** notebooks in personal drawers, raw files on a departed member's laptop, ELN entries that only the author can access, no index for a shelf of completed notebooks. ## A Worked Example: Scoring 10 Entries ![Calculation where entry 41 scores 15 out of 18 after nine attribute scores are summed and the total is used as a fraction.](/article-images/illustrations/alcoa-principles-lab-notebook-checklist-entry-score.webp) Suppose you sample entries 41 through 50 of a bound notebook. You check all nine attributes on each entry and record the results. A compact score sheet might look like this: ``` Entry Attr Score Finding 41 A 2 Name, date, initials present 41 L 2 Legible, corrections struck through once 41 C 0 Entry describes work dated 2 days earlier (backfilled) 41 O 2 Printout 2026-03-04 taped in, edges clear 41 Ac 2 Units given for all measurements 41 Cp 1 Failed run mentioned but no conditions recorded 41 Cs 2 Chronological with entry 40 41 E 2 Permanent ink 41 Av 2 Notebook indexed on shelf 2 ``` Score each entry out of 18 (nine attributes, 0 to 2 each). If entry 41 sums to 15, the entry meets the record at 15/18, with the contemporaneous failure as the item to fix. Tally the attribute-level failures across all sampled entries, not just entry totals, because the pattern matters. If contemporaneous fails on 6 of 10 entries, the fix is a lab-wide habit change, not a one-off correction. For a quick summary calculation in a spreadsheet, put each attribute score in its own row and use: ``` =SUM(C2:C10)/18 ``` to get the entry score as a fraction, and: ``` =COUNTIF(D2:D200,"Contemporaneous fail")/COUNTA(D2:D200) ``` to get the fraction of sampled checks that failed on a given attribute. A score of 0.6 for contemporaneous means 60 percent of checked entries were not written at the time of the activity, which is a strong signal for a process fix. ## Auditing Timestamps in an Electronic Lab Notebook Most ELNs record a creation timestamp and a last-modified timestamp for each entry. An audit can compare them at scale. Export your entries to CSV with columns such as `entry_id`, `author`, `created_at`, `modified_at`, then run a short script that flags large edit lags, which are the electronic analog of backfilling. ```python import pandas as pd df = pd.read_csv("eln_export.csv", parse_dates=["created_at", "modified_at"]) df["edit_lag_days"] = (df["modified_at"] - df["created_at"]).dt.days flags = df[df["edit_lag_days"] > 30] print(flags[["entry_id", "author", "created_at", "modified_at", "edit_lag_days"]]) print(f"flagged {len(flags)} of {len(df)} entries") ``` Expected output looks like this: ``` entry_id author created_at modified_at edit_lag_days 7 108 j.rivera 2026-05-02 14:10:00 2026-06-20 09:45:00 49 12 113 m.okafor 2026-04-18 11:02:00 2026-06-01 16:30:00 44 flagged 2 of 87 entries ``` Interpretation, not automation, does the rest. A large edit lag is not proof of backfilling. Adding a long-awaited analysis result months later is legitimate, and the audit trail (if enabled) will show exactly what changed. The point of the flag list is to tell you which entries deserve a manual look at their audit trail, not to serve as findings by themselves. In a spreadsheet, the same check is one column: ``` =IF(C2-B2>30,"REVIEW","OK") ``` Common export problems and their fixes appear in the common mistakes section below. ## What a Lab Notebook Template Should Contain A template does most of the compliance work before an entry is ever written. Whether you use a paper notebook, a printed pro forma, or an ELN template, include these fields: 1. **Header block:** experiment title, notebook owner, project code, page number. 2. **Date and time** of the work, not just the date of writing. 3. **Personnel:** who performed each step, and witnesses or collaborators where relevant. 4. **Objective:** one or two sentences stating what the experiment tests. 5. **Materials and conditions:** reagent identities with lot or batch numbers where used, equipment with asset or serial identifiers, and all setpoints (temperatures, speeds, concentrations, incubation times). 6. **Method reference:** the protocol or SOP name and version, or the full method written out. 7. **Raw results:** the unprocessed observations, with attached printouts or file references. 8. **Deviation section:** anything that differed from the plan, recorded when it happened. 9. **Conclusion and sign-off:** interpretation, next steps, signature and date. 10. **Verification line:** a second signature for witness-based work, where your lab requires it. A printed one-page version of this list, taped inside the front cover of every paper notebook, converts the ALCOA+ attributes into daily habits. In an ELN, the same list becomes a form with required fields, which prevents the most common completeness failures automatically. ## Paper Notebooks, Hybrid Systems, and ELNs Compared No medium satisfies ALCOA+ by itself. The table below summarizes how the three common options handle the attributes, what the main risks are, and where each fits. | Option | How ALCOA+ attributes are handled | Main risks | Best fit | |--|--|--|--| | Paper notebook | Attribution by signature, originality by physical record, endurance by ink and binding | No automatic timestamps, legibility depends on the writer, storage and retrieval are manual, search is impossible | Field work, quick bench notes, labs without ELN infrastructure | | Hybrid (paper plus scanned or printed attachments) | Paper keeps originality, scanning adds availability and endurance if the scan is indexed and backed up | Uncontrolled scans, orphaned printouts, unclear which copy is the original | Labs transitioning to electronic systems | | Electronic lab notebook (ELN) | Automatic timestamps, user accounts give attribution, audit trails record every change, search gives availability | Audit trail disabled or never reviewed, shared logins destroying attribution, data locked in a discontinued product, weak import of legacy records | High-volume data, collaborations, regulated work | The key point in any electronic lab notebook comparison is that the software enforces nothing unless it is configured to. Attribution requires individual accounts, not shared credentials. Enduring requires an export format you can still open after the vendor disappears. Available requires backups someone actually tests. An ELN with defaults left untouched can score worse on ALCOA+ than a disciplined paper notebook. ## Correcting Errors and Handling Loose Records ![Path for correcting an error: on paper strike one line, add initials and date; in an ELN change the entry so the audit trail keeps the old value.](/article-images/illustrations/alcoa-principles-lab-notebook-checklist-correction-path.webp) Two operations cause a large share of real findings, so they deserve explicit procedures. **Correcting an error (paper).** Draw one line through the error, write the correct value beside it, add your initials and the date, and, where the correction is substantive, a short reason ("transposition error, rechecked against printout 2026-03-04"). Never erase, overwrite, or use correction fluid. **Correcting an error (electronic).** Make the change in the ELN and rely on the audit trail to preserve the old value. Do not delete and re-create the entry, because that breaks the link between the record and its history. If the system allows deletion, that is a configuration problem to raise with your administrator. **Loose records.** Instrument printouts, photos, and raw data tables are originals and must be retained. In a paper system, tape printouts in along their edges (never over writing), label each one with date, source instrument, and entry number, and reference it from the body text. For oversized items, mark the envelope or folder with the same identifiers and reference it in the entry. ## How to Run the Audit and Document It The audit itself is a record too, and it should meet the same standards. 1. Open the audit with a short header: date, auditor, notebooks sampled, sampling plan, checklist version. 2. Work through the checklist for each sampled entry, recording entry number, attribute, score, and finding. 3. For each finding, classify it: correctable in the record (add a missing date with a dated, initialed annotation), correctable by process (change the template), or uncorrectable (the original data is gone). 4. Summarize attribute-level failure rates and write two or three specific actions, each with an owner and a due date. 5. Sign and date the audit record and store it with the notebooks it covers. A finding like "entries 41, 44, 47 lack contemporaneous recording" with an action of "add a 10-minute end-of-session entry requirement and re-audit in 30 days" is usable. A finding like "documentation needs improvement" is not. ## Common Mistakes and Limitations These are the failures that show up again and again in real audits. - **Auditing only the newest pages.** Older entries reveal chronic problems that recent entries hide. Always sample a range that spans months. - **Treating the checklist as pass or fail with no written findings.** A tally with no descriptions cannot be acted on and cannot be defended later. - **Filling gaps after learning about the audit.** Backfilling to satisfy an audit converts a minor documentation gap into a data integrity violation. The correct response to a discovered gap is a dated, initialed note explaining the situation, never a fabricated entry. - **Allowing shared logins in an ELN.** A shared account makes every record attributable to nobody. Individual accounts are a precondition for the Attributable attribute, whatever the vendor's marketing says. - **Never reviewing the audit trail.** An ELN audit trail that nobody reads records nothing of value. Schedule audit trail review the way you schedule equipment calibration. - **Treating signed pages as verified work.** A signature attests that the person performed and recorded the work. It does not certify the science. Witness signatures, where required, mean the witness observed the work. - **Discarding raw data after transcription.** The transcribed value is a copy. The original readout, file, or printout is the record. - **Ignoring small failures.** A missing date seems trivial until it becomes the reason a result cannot be distinguished from a neighboring entry two years later. Small failures are cheap to fix and expensive to inherit. - **Confusing data integrity with data quality.** ALCOA+ asks whether the record faithfully captures what happened. It cannot make an experiment well designed or an instrument well calibrated. - **Assuming the checklist replaces institutional policy.** Your institution, funder, or quality system may impose stricter requirements. This checklist covers the documentation attributes themselves. Individual cases, especially anything touching regulated or clinical work, need review by the responsible quality or compliance function. ## Frequently Asked Questions ### What does ALCOA+ stand for? ALCOA stands for Attributable, Legible, Contemporaneous, Original, and Accurate. The "+" adds Complete, Consistent, Enduring, and Available. Together the nine attributes define a trustworthy record. ### How often should a lab notebook be audited? For active projects, a self-audit every one to three months catches habit drift early. Completed notebooks should get a final audit before they are indexed and archived, when corrections are still possible from memory. ### What is the difference between ALCOA and ALCOA+? ALCOA covers the five original attributes that describe an individual record. ALCOA+ adds four attributes (complete, consistent, enduring, available) that describe the record's context, lifespan, and retrievability. The plus attributes mostly concern systems and storage rather than single entries. ### Can you correct a mistake in a lab notebook? Yes, and you should. Draw a single line through the error so it stays readable, write the correct value nearby, and add your initials and the date. Never erase, overwrite, or use correction fluid, because those destroy the original record. ### Do electronic lab notebooks satisfy ALCOA+ automatically? No. An ELN makes compliance easier through automatic timestamps, audit trails, and user accounts, but only if those features are switched on and used. Shared logins, disabled audit trails, and untested backups all break ALCOA+ in an otherwise capable system. ### What should a lab notebook template include? At minimum: a header with title, owner, project, and page number, plus date and time, personnel, objective, materials with identifiers, method reference, raw results, a deviation section, and a conclusion with signature and date. Required fields in an ELN template enforce the same structure. ### Who should perform the audit? A person who did not write the entries under review. Peer auditors from the same lab work well for routine checks because they understand the science. For regulated work, quality staff or an independent reviewer performs or oversees the audit. ### How do you handle loose printouts and raw data? Treat them as originals. Tape printouts into the notebook along their edges, never covering writing, and label each with the date, source, and entry number. For oversized items, use a labeled folder referenced from the entry, and never discard the original after transcribing values by hand.