Is It Ever Acceptable to Delete an Incorrect Research Record?

If a Research Record Is Wrong, Why Should You Keep It?

A researcher discovers that a spreadsheet contains an incorrect measurement, an experimental run was performed using the wrong settings, or a participant record was duplicated. The record is clearly erroneous and should not be used in the final analysis.

Would deleting it be the most sensible way to prevent future confusion?

Not necessarily. A record can be unsuitable for analysis while remaining important evidence of what occurred. Whether it can be deleted depends on its evidentiary role, applicable retention requirements, and the reason for removing it.

Deletion May Be Permissible, but an Error Alone Is Not Sufficient Justification

When Is Deleting a Research Record Appropriate, and When Is It a Research Integrity Risk?

An Incorrect Record Can Still Be Authentic Evidence

Research records serve two related but different purposes. They preserve information used to produce findings, and they document the activities through which that information was generated.

Suppose an instrument records a temperature of 85 degrees Celsius because its sensor malfunctioned. The value may not represent the actual experimental temperature. Nevertheless, the record may accurately document the instrument's output at that moment.

Deleting the observation would remove evidence of the malfunction. Retaining it with an appropriate annotation could help another researcher understand why the value was excluded from analysis.

The distinction is consequential. A failed experiment, incorrect entry, or invalid measurement may still form part of the historical record of the study.

Deleting a Record Is Different From Excluding It From Analysis

Researchers often use the word "delete" to describe several different operations. These should not be treated as interchangeable.

For analytical errors, exclusion or correction will often achieve the researcher's objective without destroying the original evidence. The principles for documenting research record corrections remain relevant when a record must be amended rather than removed.

When Might Deletion Be Legitimate?

Not every file created during research must necessarily be retained forever. Research projects may generate temporary files, redundant exports, disposable working copies, and materials that have reached the end of an authorized retention period.

Deletion may be appropriate in circumstances such as:

• Authorized disposal after retention: The applicable retention period has ended, and no continuing obligation requires preservation.
• Removal of redundant working copies: Duplicate files can be removed when an authoritative record remains securely preserved and the copies are not independently required.
• Approved records-management procedures: A documented institutional schedule authorizes disposal of a particular category of records.
• Data-protection obligations: Applicable law, consent arrangements, or approved protocols may require certain personal information to be erased, subject to relevant exceptions and competing obligations.
• Security remediation: An improperly exposed copy may need to be removed from an unauthorized location while the institution preserves necessary evidence and follows incident-response procedures.

These circumstances do not create automatic permission to delete. The researcher must determine whether the record falls within the relevant category and whether any retention, investigation, contractual, or regulatory restriction applies.

When Should an Incorrect Record Normally Be Preserved?

Preservation is particularly important when the record is original evidence, supports a reported finding, documents a deviation from the protocol, or explains a consequential analytical decision.

For example, a laboratory instrument's failed run may need to remain in the record because it documents the experiment's chronology. Similarly, a survey response excluded under a quality-control rule may still be relevant to assessing how exclusions affected the study.

Researchers should be especially cautious about deleting records associated with disputed results, suspected irregularities, or pending reviews.

What Do Formal Research and Laboratory Standards Say About Deleting Original Entries?

Some regulated research environments impose explicit requirements for preserving original entries and documenting changes.

For example, the United States Food and Drug Administration's Good Laboratory Practice regulation at 21 CFR 58.130(e) requires changes to recorded data in covered nonclinical laboratory studies to preserve the original entry, indicate the reason for the change, and identify when and by whom it was made.

This is a binding requirement within its regulatory scope, not a rule that automatically governs every academic research project worldwide. Nevertheless, it illustrates an important data-integrity principle: a correction should not conceal the earlier record.

Electronic record requirements can also be significant. Where applicable, 21 CFR Part 11 includes controls for electronic records and audit trails. The FDA's data-integrity guidance further explains the importance of retaining information needed to reconstruct regulated activities.

Outside regulated environments, institutional research policies, funder conditions, disciplinary standards, and records-management schedules may govern retention and disposal. Researchers should consult the requirements that actually apply to their project.

What About Participant Privacy and Requests for Data Deletion?

Research integrity does not mean that all identifiable participant information must be retained indefinitely.

Data-protection requirements may impose obligations concerning retention limits, erasure, confidentiality, and secure disposal. Consent documents or ethics-approved protocols may also describe circumstances under which participant data can be withdrawn or deleted.

However, a participant's request for deletion does not always create an unconditional right to erase every associated research record. Applicable legal exceptions, anonymization, the stage of analysis, and legitimate research obligations may affect the outcome.

Where competing obligations arise, researchers should consult the institutional data-protection officer, research ethics committee, or other authorized office. It may be possible to remove identifying information while preserving permissible nonidentifiable research evidence, but the feasibility and legality of doing so depend on the circumstances.

Deleting personal data and deleting the scientific history of a study are not necessarily the same operation.

How Should an Authorized Deletion Be Documented?

When deletion is permissible, a record of the disposal decision may still be necessary. The documentation should be proportionate to the sensitivity, evidentiary value, and applicable requirements of the material.

A useful disposal record may identify:

• The category and identifier of the records being deleted.
• The reason deletion is appropriate.
• The retention schedule, policy, or authorization supporting the decision.
• The person approving and performing the deletion, where applicable.
• The date and method of disposal.
• Whether an authoritative copy or related record remains.
• Any confirmation required by the institution's records-management or information-security procedures.

The disposal log itself should not unnecessarily reproduce sensitive information. It should contain enough detail to establish accountability without creating a new privacy risk.

Documenting deletion is part of maintaining a traceable research audit trail. It does not, however, make an otherwise prohibited deletion acceptable.

Why Can Deleting Incorrect Records Distort Research Findings?

Deletion can change the dataset in ways that affect the interpretation of results.

Removing observations because they are demonstrably invalid under a defensible rule is different from removing observations merely because they contradict a hypothesis. The latter can introduce selection bias and may misrepresent the evidence.

Researchers should distinguish technical invalidity from statistical unusualness. An extreme value is not automatically an error, and a result inconsistent with expectations is not sufficient justification for deletion.

When observations are excluded, the exclusion criteria, number of affected records, and consequences for the analysis should be documented where relevant. Sensitivity analyses may help establish whether findings depend heavily on particular exclusion decisions.

The need to preserve such decisions is consistent with documenting research decisions that may not appear in the final paper.

What if an Incorrect Record Has Already Been Deleted?

If an important record was deleted, first determine whether it can be recovered from an authorized backup, version history, repository, or original source. Avoid making further changes that could overwrite recoverable evidence.

Document what is known about the deletion, including when it occurred, how it was discovered, what material may have been lost, and whether the research findings are affected.

If the deletion was unauthorized, involved sensitive information, or affects the reliability of the study, consult the appropriate institutional office. Do not silently reconstruct missing evidence and present it as an original record.

The appropriate response depends on the seriousness of the loss, the availability of independent evidence, and the applicable institutional requirements.

Should a Researcher Delete an Invalid Experimental Run?

The value of 240 is not excluded simply because it is unusually high. The exclusion is supported by independent evidence of instrument malfunction.

Deleting the original file would not improve the scientific justification. It would instead make it harder to verify why the run was rejected.

By preserving the record while excluding its measurement, the researcher maintains a distinction between what happened during the experiment and what evidence is appropriate for analysis.

Misconceptions About Deleting Incorrect Research Records

Decide Whether to Correct, Exclude, Preserve, or Delete

Before removing an incorrect record, identify what kind of information it contains and what function it serves. The key decision is not merely whether the record is correct, but whether its continued preservation is required or scientifically meaningful.

These decisions should be consistent with the project's requirements for adequate research records. When uncertainty remains, preserving the record temporarily while obtaining appropriate advice is generally safer than taking an irreversible action without authorization.

Before Deleting an Incorrect Research Record, Verify These Conditions

Frequently Asked Questions About Deleting Research Records

Incorrect Research Records Are Not Automatically Disposable

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.