What Should You Do When Two Researchers Have Different Records of What Happened?

What if Two Researchers Remember or Document the Same Event Differently?

Two members of a research team disagree about what happened during data collection. One researcher's field notes indicate that a participant withdrew before completing an interview. Another researcher's records suggest that the interview was completed and the participant's responses were retained.

Both records appear to have been created during the study. Neither researcher initially recognizes an error. Yet the discrepancy matters because it affects whether the participant's data should be included in the analysis.

Which account should the team accept? Should one record be corrected, both preserved, or the disagreement referred to someone outside the project?

Conflicting research records require careful evaluation because disagreement between records is not, by itself, evidence that either researcher acted dishonestly.

Preserve Both Accounts and Resolve the Discrepancy Through Evidence

How to Investigate and Reconcile Conflicting Research Records

Why Can Two Legitimate Research Records Contradict Each Other?

Research records are created by people working under different conditions, sometimes observing different aspects of the same activity. Even when researchers act carefully, their documentation may not be identical.

One observer may record when an interview began, while another records when the recording device was activated. A laboratory researcher may document the instrument's displayed reading, while a colleague records a value from a later exported file.

Differences can also arise because records were created at different times. An initial observation may subsequently be corrected, clarified, or reinterpreted. Without a clear history of those changes, the records may appear contradictory.

Common sources of disagreement include:

• Different observation points: Researchers witnessed different portions or aspects of an event.
• Recording errors: Dates, values, identifiers, or descriptions were entered incorrectly.
• Different operational definitions: Researchers applied different meanings to the same category or procedure.
• Timing differences: One record reflects an earlier state, while another reflects a subsequent change.
• Incomplete documentation: One or both accounts omit contextual information necessary for interpretation.
• Retrospective recollection: An account was reconstructed from memory rather than documented contemporaneously.

These possibilities should be examined before attributing the disagreement to negligence or misconduct.

First, Establish Exactly What the Records Disagree About

A disagreement may appear broader than it actually is.

For example, one researcher writes that an interview lasted 45 minutes, while another records 52 minutes. The difference may reflect whether introductory instructions and post-interview discussion were included.

Before deciding which record is correct, determine whether both accounts describe the same event, use the same definitions, and refer to the same time period.

It may help to prepare a neutral discrepancy record identifying:

• The event or observation being examined.
• The relevant original records and their identifiers.
• The specific statements or values that differ.
• The dates on which the records were created or amended.
• The possible consequences of the disagreement.
• The evidence needed to investigate the discrepancy.

This approach keeps the investigation focused on verifiable questions rather than personal disagreements.

Why Should Both Original Records Be Preserved?

Preserving both accounts protects the evidentiary history of the research.

If one record is overwritten or deleted before the discrepancy is examined, the team may lose information needed to determine how the disagreement arose.

For paper records, the original entries should remain legible. For electronic records, version histories, relevant metadata, and system logs should be preserved where available.

Any subsequent correction should follow the established procedures for documenting amendments to research records.

Preservation does not require treating both accounts as equally accurate. It means retaining the evidence necessary to evaluate their reliability.

If an investigation, audit, or formal research-integrity process is already underway, additional preservation requirements may apply. Researchers should follow the directions of the authorized institutional officials.

Which Evidence Should Be Used to Resolve the Disagreement?

Whenever possible, examine evidence that is independent of the conflicting recollections.

Relevant materials may include instrument logs, original data exports, timestamped system records, approved protocols, laboratory notebooks, audio recordings, contemporaneous correspondence, or other authorized documentation.

The strength of each source depends on its provenance, authenticity, completeness, and relevance to the disputed event.

For example, a timestamped instrument log may provide stronger evidence of when a measurement occurred than a researcher's recollection several months later. However, the log may not explain whether the instrument was configured correctly or whether its clock was synchronized.

Similarly, an interview recording may clarify what was said, but it may not establish whether a participant subsequently withdrew consent for a particular use of the material.

Evidence should therefore be assessed against the specific question being investigated. No single category of record is automatically decisive in every situation.

How Should Contemporaneous Records Be Weighed Against Later Recollections?

Records created close to an event often provide useful evidence because they may be less affected by later forgetting or reinterpretation.

Nevertheless, contemporaneous documentation is not necessarily accurate. A researcher can misunderstand an event or enter an incorrect value immediately after it occurs.

Later accounts may identify genuine errors, particularly when supported by independent evidence. Their retrospective nature should simply remain clear.

A reasonable evaluation considers when each account was created, what the author could directly observe, whether the information was subsequently modified, and whether independent evidence supports it.

Researchers should not convert an uncertain recollection into an apparently original record by changing its date or presentation.

Maintaining a traceable research audit trail helps preserve the chronology necessary for this evaluation.

What if the Researchers Used Different Definitions?

Some apparent conflicts arise from inconsistent terminology rather than incompatible observations.

Consider two researchers recording participant completion status. One defines completion as reaching the final survey page. Another defines it as submitting every required response.

The same participant may be classified as complete under one definition and incomplete under the other.

Resolving the discrepancy requires examining the study's intended operational definition, approved protocol, instrument design, and applicable analytical rules.

If the definition was ambiguous, the team may need to establish a defensible interpretation and apply it consistently. Any consequential clarification should be documented, including whether the rule was established before or after relevant outcomes were examined.

The original classifications should remain traceable. Reconciliation should not conceal that the earlier records reflected different assumptions.

Should the Principal Investigator Decide Which Account Is Correct?

The principal investigator may have responsibility for resolving routine project documentation issues, subject to institutional policies and the project's governance arrangements.

However, authority over the project is not equivalent to independent evidence that a particular account is accurate.

A responsible decision should be grounded in the records, relevant procedures, and available corroboration. Where the principal investigator participated in the disputed event or has a potential conflict of interest, additional review may be appropriate.

For consequential methodological disagreements, the team may seek advice from a qualified methodologist, data steward, research administrator, or other authorized reviewer.

If the disagreement involves credible concerns about fabrication, falsification, deliberate concealment, or interference with evidence, the matter may require formal research-integrity procedures rather than ordinary project reconciliation.

Researchers should avoid assuming that a senior colleague's account is more reliable solely because of their position.

How Should the Resolution Be Documented?

Once the available evidence has been evaluated, the team should preserve a record of the reconciliation.

A useful reconciliation record may contain:

• The identifiers of the conflicting original records.
• The precise nature of the discrepancy.
• The evidence examined and its provenance.
• The explanation for the disagreement, if established.
• The conclusion reached and the basis for accepting it.
• Any remaining uncertainty or limitation.
• The person responsible for the determination and the date.
• Any correction, analytical change, or reporting action resulting from the decision.

Where the disagreement is resolved, the reconciliation should explain why one interpretation or value was adopted without erasing the earlier accounts.

Where the disagreement remains unresolved, the record should state that explicitly.

This documentation supports the reconstruction of consequential research decisions by someone who did not participate in the original discussion.

What if Neither Account Can Be Verified?

Not every discrepancy can be resolved conclusively.

Sometimes the original source evidence is unavailable, the surviving records are incomplete, or the available documentation supports more than one plausible interpretation.

In these circumstances, researchers should resist the temptation to select the more convenient account merely to complete the dataset or manuscript.

The appropriate response depends on the significance of the disputed information. Options may include retaining an uncertainty flag, treating a particular value as missing, conducting sensitivity analyses, or excluding an affected observation under a justified rule.

These are methodological decisions, not automatic consequences of disagreement. Their appropriateness depends on the research design, available evidence, and applicable requirements.

If the unresolved discrepancy materially affects a reported conclusion, the limitation should be considered when interpreting or communicating the findings.

When Does a Record Discrepancy Become a Research Integrity Concern?

Conflicting records do not automatically constitute research misconduct.

However, additional concerns may arise when evidence suggests that records were knowingly altered to misrepresent observations, original entries were deliberately concealed, or researchers were pressured to adopt an unsupported account.

Under the U.S. Public Health Service framework, research misconduct includes fabrication, falsification, and plagiarism as defined by the applicable regulation. The determination of misconduct requires consideration of specific standards and evidence; an ordinary disagreement does not satisfy those standards by itself.

Researchers should distinguish a good-faith disagreement about evidence from a credible allegation of intentional or reckless misrepresentation.

When misconduct is reasonably suspected, the appropriate institutional reporting process should be followed. Informal reconciliation should not be used to suppress a legitimate concern or interfere with a formal investigation.

How Should Conflicting Records Affect Published or Submitted Research?

The consequences depend on whether the discrepancy affects the reliability of reported findings.

A disagreement about a minor administrative detail may have little effect on the manuscript. A conflict involving sample eligibility, treatment assignment, measurement values, participant consent, or analytical exclusions may be more consequential.

If the discrepancy is discovered during manuscript preparation or peer review, the researchers should determine whether the analysis or methodological description requires revision.

For published research, a material discrepancy may require consultation with the journal and relevant institutional officials. The appropriate response could range from clarification or correction to a more substantial examination of the published findings.

Not every discrepancy requires a public notice. The relevant question is whether the published record remains accurate and sufficiently supported.

How Can Research Teams Prevent Conflicting Records?

Some disagreements are unavoidable, especially when researchers make independent observations. Nevertheless, consistent documentation practices can reduce preventable discrepancies.

Teams should establish operational definitions, agreed data-entry procedures, common record identifiers, clear responsibilities, and appropriate version controls.

Where independent observations are part of the research design, differences should be preserved until the prescribed comparison or reconciliation procedure has been completed.

For example, independently coded qualitative materials should not necessarily be forced into agreement before the analytical approach calls for comparison. Differences may themselves be informative, depending on the methodology.

Periodic checks can identify inconsistencies before they affect later analyses. A clear process for resolving disagreements also reduces the risk that authority or interpersonal pressure will substitute for evidence.

When team membership changes, appropriate research record handover procedures can preserve the context needed to interpret apparently conflicting accounts.

Resolving Two Conflicting Accounts of a Participant's Withdrawal

In this hypothetical situation, the records appeared contradictory because they referred to different aspects of the participant's involvement. The interview was completed, but the subsequent withdrawal affected the permitted use of its data.

Neither original account needed to be erased. The reconciliation clarified the chronology and allowed the team to make an informed decision about the analytical dataset.

If the withdrawal record had been unavailable, the team would have needed to acknowledge the unresolved uncertainty rather than assume that either researcher's account was necessarily correct.

Common Mistakes When Research Records Contradict Each Other

Resolve Record Conflicts According to Evidence and Consequence

When records disagree, begin with the evidence rather than the identities or preferences of the researchers involved.

The objective is to establish what can reasonably be concluded, preserve the original accounts, and determine whether any unresolved uncertainty affects the study.

A documented reconciliation should be treated as an evidence-based conclusion, not an opportunity to produce an artificially consistent project archive.

Where essential source material is unavailable, the procedures for handling missing research records may also become relevant.

Before Resolving Conflicting Research Records, Check These Details

Frequently Asked Questions About Conflicting Research Records

Resolve Conflicting Records Without Erasing the Disagreement

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.