A student may save a final chart and still be unable to explain which data file produced it. A research notebook helps connect a result to the source, method and version used. The goal is not to make every project use a particular software product; it is to leave an intelligible record that can be checked under the project's rules.

The US National Institutes of Health discusses documentation and reproducibility in its electronic laboratory notebook guidance. Those institutional requirements apply within their stated context and are not a universal rule for student projects. The fictional analysis below illustrates an independent, small-scale provenance map without using personal or laboratory data.

Follow a result backward

Suppose an invented project counts study-room observations on three days. The original data file contains values 2, 4 and 6. A documented script calculates the mean as (2 + 4 + 6) / 3 = 4. The student saves a chart displaying that value.

Record Fictional identifier What it establishes
Raw data observations-v1.csv The three original values
Analysis mean-v1.py The calculation applied to that file
Output chart-v1.png The displayed result of 4
Notebook entry entry 01 The links, date and explanation

The filenames are illustrative, not a required naming standard. A chart alone cannot establish which of several similarly named files was used. The notebook entry should make the connection explicit and include enough context to distinguish the dataset, method and output.

Record a correction without erasing the history

Imagine the student discovers, from an authorized source, that the third observation was entered incorrectly and should be 9. Save a corrected file as observations-v2.csv, document the reason and evidence for the correction, and rerun the same stated calculation. The revised mean is (2 + 4 + 9) / 3 = 5. The difference from the previous result is one unit.

Keep the original and corrected versions according to project policy. Record which chart is superseded and which one is used in the final submission. Do not describe the revised figure as if it came from the original data, and do not alter the old output silently. If the proposed correction cannot be verified, mark it as unresolved rather than presenting it as established.

Capture the steps that affect interpretation

A useful entry records the question, date, source, units, relevant file identifiers, transformation and output. It should also explain exclusions, assumptions and unexpected events. For code-based analysis, note the program or environment version where that information affects reproducibility. For a physical experiment, the relevant details might include a protocol version, equipment settings and sample identifiers under the supervisor's approved system.

Reproducibility does not mean that every later user must obtain identical results under changed inputs or an altered method. It means the record should make the inputs and steps clear enough for an authorized reviewer to understand and, where appropriate, repeat the analysis. A version number without an accessible record or description may not be sufficient.

Respect access and project rules

Do not put sensitive records, participant identifiers or restricted data into a public notebook merely to make an example easier to share. Follow the institution's access, consent, retention and security requirements. A notebook can point an authorized reader to a protected location without exposing the underlying data. The responsible supervisor or data steward can clarify which storage system is approved.

A personal draft spreadsheet and a formal research record may serve different purposes. Before deleting, renaming or sharing a file, check the project's versioning and retention rules. An analysis may also require approval before collecting new information; documentation does not replace that approval.

Check the trail before submission

Starting from each reported figure, ask which source and version produced it, what transformation was applied, whether a correction is documented and where the authorized record can be found. Then work forward from the source to confirm that the cited output follows the stated process.

This map gives the student a way to spot mismatched files before a report is submitted. It also gives a reviewer a clear route through the work, while keeping restricted information within the systems designed to protect it.