How Should Corrections to Research Records Be Documented?

How Can You Correct a Research Record Without Compromising Its Integrity?

You discover that a measurement was entered incorrectly, a participant identifier was transposed, or a laboratory observation was recorded under the wrong date. The mistake is genuine, and you know what the correct information should be. Can you simply replace the incorrect entry?

Research records are expected to be accurate, but accuracy does not mean pretending mistakes never occurred. The important question is how to correct an error while preserving a reliable account of the research process.

Correct the Error, but Preserve the Original Record and the Reason for the Change

What Makes a Research Record Correction Transparent and Defensible?

A Correction Is Different From Rewriting the Research History

A research record documents what was observed, measured, decided, or performed. It may contain mistakes because researchers are human, instruments malfunction, transcription errors occur, and interpretations change as evidence accumulates.

Correcting a mistake is part of maintaining reliable records. The problem arises when a correction conceals what was originally recorded or creates a misleading impression about when information became available.

For example, correcting a transposed measurement is different from replacing an inconvenient observation with a value that better fits an expected pattern. The first may be a legitimate correction supported by evidence. The second may constitute inappropriate data alteration if the replacement lacks a defensible basis.

The distinction depends on the evidence supporting the change, the procedure followed, and whether the correction is represented accurately.

What Information Should Accompany a Correction?

A defensible correction generally preserves enough information for another authorized person to understand what changed and why.

• Original entry: The information as it existed before correction, or a recoverable version of it.
• Corrected entry: The replacement information or amended interpretation.
• Date of correction: When the change was actually made, distinguished from the date of the original observation.
• Identity of the person making the correction: The researcher or authorized individual responsible.
• Reason for correction: A specific explanation of the error and the evidence supporting the change.
• Supporting evidence: A reference to an instrument log, original form, source file, calibration record, or other material when relevant.

Not every system requires these details in identical fields. What matters is whether the documentation makes the change understandable and verifiable.

How Should Corrections Be Made in a Paper Laboratory Notebook?

For a bound paper notebook, a widely recommended practice is to draw a single line through the incorrect entry so that it remains legible, write the corrected information nearby, and add the correction date and initials or signature. A brief explanation should accompany the correction when its reason would not otherwise be apparent.

Do not use correction fluid, erase the original writing, or remove pages when doing so would conceal the earlier record. Avoid overwriting numbers in a way that makes the original value impossible to identify.

If the correction requires a longer explanation, a dated supplementary entry can refer back to the original page and entry. The original record should remain intact.

These are general documentation practices rather than a universal legal standard. Institutional notebook procedures, regulated laboratory requirements, and validated recordkeeping systems may prescribe more specific methods.

How Should Corrections Be Made in Electronic Research Records?

Electronic records create additional possibilities and risks. A database may allow a value to be overwritten instantly, while an electronic laboratory notebook may automatically preserve previous versions.

When the system supports version history or an audit log, corrections should ordinarily use those functions rather than bypass them. Ideally, the system records the previous value, new value, user identity, and timestamp. A reason for the change should also be recorded when the system permits or requires it.

If the software does not preserve change history, researchers may need a separate correction log, versioned files, or a controlled amendment procedure. The chosen method should comply with the project's data-management plan and applicable institutional requirements.

How Do You Correct a Dataset Without Altering the Raw Evidence?

Research datasets frequently contain coding mistakes, transcription errors, duplicate records, and inconsistent identifiers. Correcting these problems may be necessary before analysis, but the correction process should distinguish source observations from analytical transformations.

A useful approach is to preserve the original dataset as a controlled source and create a cleaned or corrected analytical version. Changes can be documented through a reproducible cleaning script, transformation log, or dataset change register.

For each consequential correction, the documentation should identify the affected record, the original and revised values, the correction rule, and the justification. Where corrections are applied systematically, the script and its documented rules may provide stronger traceability than manually editing hundreds of cells.

For example, converting all valid responses coded as "Female" to the standardized category "F" is a coding transformation, not necessarily evidence that the original entries were wrong. By contrast, changing a participant's response because it appears implausible requires a separate evidentiary justification.

What if the Original Information Cannot Be Recovered?

Sometimes a correction is needed because an original entry was incomplete, ambiguous, or demonstrably incorrect, but the underlying source cannot be located.

In such cases, do not present an inferred replacement as though it were an independently verified observation. Record the uncertainty, explain the basis of any reconstruction, and consider whether the value should remain missing or be excluded under an established analytical rule.

A correction supported by a contemporaneous instrument export is different from a value reconstructed from memory several months later. The record should preserve that distinction.

Does the Timing of a Correction Matter?

Yes. A correction made immediately after data entry and one made after publication can have different implications.

The date of the original observation must not be changed to make a later correction appear contemporaneous. Backdating a correction may misrepresent the chronology of the research, even if the replacement value is accurate.

If an error is discovered after analysis, determine whether it affects derived variables, statistical results, figures, conclusions, or shared datasets. A corrected source record does not automatically update every downstream research output.

When published findings are materially affected, journal correction procedures and institutional research-integrity processes may become relevant. The appropriate response depends on the nature and significance of the error.

How Do Corrections Relate to Research Integrity and Auditability?

Good documentation makes it possible to distinguish an honest correction from an unexplained alteration. The ability to reconstruct changes is central to an audit trail that supports research integrity.

Correction documentation also complements broader requirements for defending what was actually done. A final dataset may be accurate, but without records explaining its development, important aspects of its provenance may remain uncertain.

Transparent corrections are therefore not merely administrative details. They help establish the relationship between original evidence, subsequent decisions, and reported findings.

Correcting a Transcription Error in a Research Dataset

The correction is defensible because the original value is preserved, the replacement is supported by an identifiable source, and the change can be traced to the person who made it.

If the questionnaire had been lost, the researcher could not justify the same correction merely because 45 appeared more plausible than 54. The uncertainty would need to be handled explicitly.

Common Mistakes When Correcting Research Records

Choose a Correction Procedure That Preserves the Evidence

Before making a correction, identify the record's role in the research. An original observation, an imported dataset, a coding decision, and an analytical output may require different amendment procedures.

Establishing these procedures before errors occur is preferable to inventing a method during a dispute. Researchers should also understand the distinction between legitimate modification of existing research records and changes that compromise the original evidence.

Before Finalizing a Research Record Correction, Check These Details

Frequently Asked Questions About Correcting Research Records

A Proper Correction Improves Accuracy Without Erasing History

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.