Is It Misconduct to Reconstruct Missing Research Records From Memory?

Your Original Research Record Is Missing. Can You Reconstruct It?

A laboratory notebook disappears. A spreadsheet becomes corrupted. Interview notes cannot be located. A researcher still remembers much of what happened and considers recreating the missing record while the events are still reasonably fresh.

Is that itself misconduct?

Not necessarily. The central problem is what the reconstructed record represents. A retrospective account explicitly identified as reconstructed from memory is quite different from recreating a document and presenting it as though it were the original contemporaneous record.

Reconstruction Is Not Automatically Misconduct, but It Must Not Masquerade as the Original Record

Reconstruction, Recovery, and Fabrication Are Not the Same Thing

Start by Asking What Kind of Record You Are Creating

A reconstructed record is, by definition, created after the original record was lost, damaged, or found incomplete. That does not automatically make it false.

Suppose a researcher loses handwritten notes describing the sequence of procedures performed during an experiment. The researcher later creates a document stating: “Reconstructed on 4 October from surviving instrument logs, email correspondence, and my recollection.” That document does not claim to be the original notebook entry.

Now consider a different response. The researcher recreates notebook pages, inserts specific measurements that cannot be independently verified, dates the pages as though they were written on the day of the experiment, and places them among the original records without identifying them as reconstructed. The integrity problem is substantially different.

A Reconstruction Should Preserve Its Own Provenance

Provenance means knowing where information came from and how it entered the research record. For reconstructed records, provenance is especially important because the original documentation is unavailable.

A defensible reconstruction should make it possible to distinguish information recovered from surviving evidence from information recalled by the researcher. Useful supporting materials may include instrument outputs, electronic timestamps, database entries, emails, photographs, analysis files, questionnaires, consent or study records where appropriate, laboratory management systems, version histories, and other contemporaneous records.

ORI's current guidance emphasizes the importance of research records and other evidence in evaluating research misconduct allegations. The absence of primary documentation can make it considerably harder to determine what actually happened.

Reconstructing an Event Is Different From Reconstructing Exact Data

Memory may support some propositions better than others.

A researcher might confidently remember performing an experiment on Tuesday because several other records corroborate that event. That does not necessarily mean the researcher can accurately reproduce every measurement generated during the experiment.

The second claim is much more demanding because an exact numerical observation is being asserted. If the value cannot be verified, inserting a number because it seems familiar or plausible can cross into the problem of filling in a missing value from memory.

Reconstruction Should Not Quietly Become Invention

Gaps create psychological pressure. Once most of a lost record has been reconstructed, researchers may be tempted to complete the remaining blanks with what “must have happened.”

That is precisely where reconstruction can become dangerous.

If the evidence establishes that Samples A, B, and C were processed but does not establish the measurement for Sample B, the reconstructed record should preserve that limitation. A plausible number is still not an observed number.

Under the U.S. Public Health Service definition, fabrication means making up data or results and recording or reporting them. Consequently, unsupported invented details cannot be protected merely by calling the overall exercise “reconstruction.”

Do Not Backdate a Reconstruction to Make It Look Contemporaneous

A reconstructed record should normally identify when the reconstruction was actually created and, where useful, the period or research event it reconstructs.

Those are two different dates.

For example, “Reconstructed 4 October 2026 concerning experiment conducted 18 September 2026” communicates something materially different from creating the document on 4 October but dating it 18 September without explanation.

ORI has identified backdating and alteration of dates among conduct that can raise falsification concerns in particular contexts. The broader principle is straightforward: the record should accurately communicate its history rather than create a false appearance of contemporaneous documentation.

Missing Records Are Not Automatically Evidence of Misconduct Either

Research records can disappear for mundane reasons: hardware failures, accidental deletion, damaged storage media, misplaced physical files, poor recordkeeping, or organizational failures.

Under the current PHS framework, however, the absence or destruction of research records can have evidentiary consequences in a research misconduct proceeding under specified circumstances. ORI explains that such issues depend on factors including whether the respondent intentionally, knowingly, or recklessly destroyed records, failed to maintain them despite an opportunity to do so, or failed to produce maintained records, and whether the conduct significantly departed from accepted practices.

So neither conclusion is justified automatically: “the records are missing, therefore misconduct occurred” is too strong, but “the records are missing, so nothing can be evaluated” may also be wrong.

A Formal Misconduct Finding Requires More Than Poor Documentation

Research misconduct under the PHS framework means fabrication, falsification, or plagiarism. It does not include honest error or differences of opinion. A formal finding additionally requires a significant departure from accepted practices, the required level of culpability, and proof under the applicable evidentiary standard.

Poor recordkeeping can nevertheless be consequential even when it does not establish fabrication or falsification. It may violate institutional, funder, laboratory, regulatory, or disciplinary requirements, and it can make legitimate research difficult to verify.

A Lost Laboratory Notebook and Two Ways to Reconstruct It

Common Mistakes When Research Records Go Missing

Reconstruct the Evidence Trail, Not the Record You Wish You Had

If important research records disappear, begin by preserving the remaining evidence. Do not overwrite surviving files, quietly replace missing documents, or create retrospective records that could later be mistaken for originals.

If reconstructed information changes an existing dataset, keep an audit trail. The same principles used to document legitimate research data corrections can help preserve the distinction between the original surviving record and later amendments.

More broadly, prevention is considerably easier than reconstruction. Preserving original data, metadata, notebooks, code, image files, and version histories provides the evidentiary foundation needed to demonstrate the authenticity of research data and images if questions arise later.

Before Reconstructing a Missing Research Record, Check This

Frequently Asked Questions About Reconstructing Research Records

A Reconstruction Should Reveal Its Limitations, Not Hide Them

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.