What Counts as an Adequate Research Record?

When Is a Research Record Actually Good Enough?

You kept the spreadsheet. Your laboratory notebook still exists. The analysis code is somewhere in the project folder. Does that mean you have an adequate research record?

Not necessarily. A record can exist yet still be inadequate if nobody can determine who created it, when it was created, what it describes, which version is authoritative, how it relates to other evidence, or what changed later.

The practical standard is therefore more demanding than simply retaining files. The record must remain capable of supporting an accurate account of the research.

An Adequate Record Must Be Usable as Evidence

Adequacy Is About Evidentiary Quality, Not the Number of Files

The simplest test of a research record is not “Did I save something?” but “What can this record reliably establish?”

That distinction matters because records serve as evidence of the research process. U.S. Public Health Service research-misconduct regulations, for example, define the research record broadly and specifically address situations in which records do not adequately document questioned research. The regulation does not create a universal documentation template for all research, but it illustrates why adequacy can become consequential when the history of a study needs to be established.

An Adequate Record Identifies What Happened

The record should describe the relevant research activity with enough specificity to distinguish what actually occurred from what was merely planned.

Writing “experiment completed” or “data cleaned,” for example, says very little. Depending on the study, useful documentation might identify the procedure used, samples involved, instrument or software settings, deviations from the planned method, transformation rules, excluded observations, or other information necessary to understand the activity.

This does not require recording every mundane action. The appropriate level of detail depends on what another knowledgeable person would need to understand the consequential parts of the work.

It Should Be Attributable

A meaningful record should make it possible to determine who performed or documented an activity when authorship or responsibility matters. Anonymous changes, shared credentials, unexplained entries, and files whose origin cannot be established weaken the evidentiary value of the record.

In regulated contexts, the U.S. Food and Drug Administration uses the ALCOA principles for good documentation: records should be attributable, legible, contemporaneous, original, and accurate. These principles arise from specific regulatory settings and should not be misrepresented as a universal legal standard for every researcher. They nevertheless provide a useful way to think about characteristics that make records trustworthy.

It Should Be Created Close Enough to the Event to Be Credible

Timing matters. A note created during an experiment, immediately after an interview, or when an analytical decision is made generally provides stronger evidence of what happened than an account reconstructed months later.

That does not mean retrospective notes are inherently illegitimate. Sometimes researchers must document an earlier event after discovering that the original documentation was incomplete. In that situation, the retrospective nature of the entry should be clear rather than making the new account appear contemporaneous.

The distinction protects the integrity of the record. “Documented today from meeting notes and archived files” is different from creating an entry today and dating it as though it had existed six months earlier.

It Should Preserve Enough Context to Remain Understandable

A number without context is rarely useful evidence. Neither is a dataset containing variables nobody can define, a script that depends on unidentified input files, or an instrument output with no sample identifier.

NIH data-management guidance emphasizes that metadata and associated documentation help users understand how data were collected and interpreted. Depending on the research, documentation may include methodology, data labels, variable definitions, provenance, transformations, dates, units, coding rules, software information, or other contextual information needed to understand the evidence.

This is one reason the distinction between research data and research records matters. The numerical or textual data may be only one part of what is needed to interpret the research reliably.

It Should Preserve Consequential Detail Without Becoming a Diary of Everything

Adequacy should not be confused with maximal documentation. Recording every mouse click, casual conversation, or inconsequential formatting change would often create more noise than evidence.

The better question is whether omission of a detail would make a consequential part of the research difficult to verify or reconstruct. The stronger the effect of a decision or action on the evidence, analysis, or conclusions, the stronger the case for documenting it.

This is where judgment becomes necessary. Researchers routinely make small choices, but some apparently small choices later become important. The separate question of whether to document decisions that seem trivial at the time depends partly on whether those decisions could alter interpretation or reconstruction later.

It Should Distinguish Versions and Changes

An adequate record should not leave researchers guessing which version of a protocol, dataset, codebook, analysis script, or instrument was actually used.

Version identifiers, dates, controlled file locations, change histories, electronic audit trails, or similar mechanisms can help establish the sequence. The appropriate method depends on the research environment.

Changes themselves are not necessarily problematic. Research records sometimes need correction. What matters is whether the correction remains documented rather than making the previous record disappear without explanation.

It Should Preserve Negative, Unexpected, or Unsuccessful Results When They Are Part of the Record

A research record should not become a curated history containing only observations that eventually supported the preferred interpretation. ORI educational materials on laboratory record keeping emphasize recording both positive and negative results and enough methodological detail to allow others to understand what was done.

This does not mean every failed exploratory attempt must appear in a publication. Publication and record keeping serve different purposes. The concern arises when selective preservation distorts the evidentiary history of the work.

It Should Allow Important Steps to Be Traced

For many studies, no individual document is sufficient. Adequacy comes from the connections among records.

A source observation may connect to a participant or sample identifier, then to a cleaned dataset, then to a transformation script, then to an analytical output, and finally to a reported figure. If those connections disappear, all the individual files may still exist while the research process becomes difficult to reconstruct.

This traceability is closely related to the purpose of a research audit trail: preserving an intelligible sequence of consequential events, decisions, and transformations.

It Must Remain Available for the Required Period

A perfectly documented record that disappears prematurely is not adequate record keeping. Records need to remain intact, accessible, and appropriately protected for whatever retention period applies.

That period is not universal. Funders, institutions, sponsors, regulations, contracts, ethics requirements, and other authorities may impose different retention obligations. Some circumstances can also extend normal retention periods.

Two Records Can Describe the Same Analysis but Have Very Different Evidentiary Value

The second record is not adequate merely because it contains more files. It is adequate because the evidence is connected. A knowledgeable person can determine what happened without requiring the researcher to reconstruct the explanation from memory.

If the project had a different methodology, adequacy would look different. A qualitative study might instead require clear relationships among recordings, transcripts, coding decisions, analytical memos, and reported quotations. The underlying principle remains the same: the documentation should fit the research process it is supposed to evidence.

Adequate Documentation Is Not the Same as More Documentation

Use Reconstruction as the Practical Test of Adequacy

Choose one consequential result, observation, or methodological decision from your study and work backward. Can you identify the evidence supporting it? Can you determine who generated or changed that evidence? Can you identify the procedure and relevant version? Can you explain important transformations without relying on memory?

If the chain repeatedly ends with “I think this is what we did,” the record probably needs attention.

Do this while the research is active. Trying to create an adequate record after a problem appears is substantially harder, and retrospective reconstruction should never be disguised as contemporaneous documentation.

Can Your Research Record Pass These Checks?

Common Questions About Adequate Research Records

An Adequate Record Should Tell the Evidentiary Story of the Research

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.