A transparent methodology note tells readers how a news story was reported, checked, and limited. It builds trust by making the reporting process understandable without turning the article into a diary of every email, search, or editorial discussion.
What a methodology note should explain
A methodology note is a short, reader-facing account of the methods behind a story. It is especially useful when reporting involves data, documents, surveys, public records, social media posts, investigative techniques, or claims that are difficult to verify independently.
A useful note should answer five practical questions:
- What was the story trying to establish?
- What sources, records, data, or interviews were examined?
- How were important facts checked?
- What choices or definitions affected the results?
- What could not be confirmed or included?
The note should help a careful reader understand the evidence without requiring specialist knowledge. It is not a defense brief, a list of credentials, or a substitute for sourcing claims in the story itself.
Start by identifying the reporting method in plain language. For example, write “We reviewed 184 inspection reports from January 2022 through June 2024” rather than “A comprehensive document-based investigation was conducted.” The first version gives readers something they can evaluate.
Decide when a note is needed
Not every short news update requires a separate methodology section. A brief report based on one on-the-record interview and a public announcement may only need normal source attribution. A methodology note becomes more valuable when the reporting process could affect how readers interpret the findings.
Consider adding one when your story includes:
- A dataset assembled from multiple agencies or records systems.
- Analysis of a large collection of documents, messages, images, or videos.
- A survey or poll conducted by your newsroom or an outside organization.
- Estimates, projections, rankings, maps, or calculations.
- Verification of user-generated content or eyewitness material.
- A comparison across years, regions, groups, or institutions.
- Anonymous sources whose roles or corroboration need explanation.
- A reconstruction of events from incomplete or conflicting evidence.
- Automated tools, transcription software, scraping, translation, or artificial intelligence.
- Important exclusions, missing records, or unresolved uncertainty.
A methodology note can appear near the end of the article, in a clearly labeled box, or on a linked methods page. For a short story, three to six paragraphs may be enough. For a complex investigation, use a longer page with sections and supporting files.
Define the reporting question before describing the process
A methodology note is easier to write when the reporting question is specific. “We investigated corruption” is too broad to guide readers. “We examined whether city contracts awarded between 2021 and 2023 went to companies connected to current officials” is more precise.
Write down the question and the boundaries before gathering evidence. Clarify:
- The time period covered.
- The geographic area.
- The people, organizations, or events included.
- The meaning of key terms such as “complaint,” “fatality,” “connection,” or “contract.”
- Whether the work measures reported incidents, confirmed incidents, or all incidents.
These definitions matter because small wording changes can produce very different results. If “housing complaints” means complaints logged by one department, do not present the total as the number of all housing problems in a city. State what the number represents and what it does not represent.
A strong opening sentence often follows this pattern:
To examine [specific question], we reviewed [source or material] covering [time and place], then compared the findings with [independent check].
This gives readers the purpose, evidence, and broad verification method immediately.
Describe the sources precisely
Readers need enough information to understand where the evidence came from. Name the source when possible, identify the relevant date or version, and explain how you obtained it.
Instead of saying “official data,” specify “monthly reports published by the state labor department, downloaded on March 4, 2026.” If the source was supplied by a person or organization, say so when disclosure is safe and appropriate. If a source requested confidentiality, describe its role and explain why the source was not named.
For each major source, record:
- The organization or person responsible for it.
- The title, file name, URL, or archive reference.
- The date it was created, published, downloaded, or accessed.
- The relevant time period covered.
- Whether it is original, copied, summarized, or derived from another source.
- Any known gaps, revisions, or restrictions.
Distinguish between primary and secondary material. A court filing, inspection record, meeting recording, or raw dataset is generally different from a news release or a report summarizing that material. Secondary sources can be useful, but readers should know when your account relies on an interpretation rather than the underlying record.
If you used interviews, describe the number and type of interviews without exposing confidential details. “We interviewed 12 current and former employees, including four people who directly reviewed the records” is more informative than “Several sources confirmed the findings.” Do not imply that every interview supported the same conclusion if the evidence was mixed.
Explain how you selected material
Selection rules can shape a story as much as the sources themselves. Explain what was included, excluded, or unavailable.
For document reporting, state whether you reviewed every record in the defined period or a sample. If you used a sample, explain how it was selected. A random sample, a convenience sample, and a set of records obtained through a public-records request are not interchangeable.
For a dataset, describe filters and exclusions in ordinary language. You might write:
- We included cases marked “closed” or “resolved.”
- We excluded duplicate entries with the same case number.
- We removed records missing a date because they could not be placed in the study period.
- We counted each contract once, even when it had multiple amendments.
Avoid presenting a filtered result as if it were an untouched total. If you selected examples to illustrate a broader pattern, label them as examples rather than evidence that the pattern applies universally.
A compact table can make the scope easy to scan:
| Element | What to report |
|---|---|
| Population | The full group or record set considered |
| Time period | Start and end dates, including time zone if relevant |
| Included | The records, people, or events counted |
| Excluded | Duplicates, incomplete records, out-of-scope material, or other removals |
| Final set | The number of items actually analyzed |
Do not use a precise final number without explaining how it was reached. Precision can create false confidence when the underlying records are incomplete.
Show how you verified the reporting
Verification is the part readers usually care about most. Describe the checks that support your central claims, not just the existence of sources.
Useful verification steps may include:
- Comparing a figure with the original document rather than relying on a press release.
- Checking names, dates, addresses, and titles against an independent record.
- Asking the person or organization named in a claim for a response.
- Confirming that multiple records refer to the same person, event, or location.
- Reviewing a calculation manually or having another reporter reproduce it.
- Checking whether a document was later corrected, withdrawn, or superseded.
- Examining contradictory evidence and explaining how it affected the story.
Use specific language about the level of confirmation. “We confirmed the date in three independent records” is stronger and clearer than “The date was thoroughly verified.” If a claim could not be independently confirmed, say that directly.
For images and videos, explain what was checked: original upload information, visible landmarks, weather, shadows, timestamps, reverse-image results, geolocation clues, or testimony from a person who was present. These checks may establish that media existed at a certain time or place, but they may not prove what happened before or after the recording.
For anonymous sources, explain the source’s relevant access and the corroboration obtained. Avoid revealing identifying details through unnecessary descriptions. A methodology note should increase transparency without making confidentiality meaningless.
Explain calculations, coding, and tools
When a story includes numbers, give readers enough information to understand the calculation. State the formula in plain language, identify the denominator, and distinguish counts from rates or percentages.
For example, “The rate was calculated by dividing reported injuries by the estimated workforce and multiplying by 100,000” is useful. “We calculated the rate using standard methods” is not.
Mention important technical steps, including:
- How dates and categories were standardized.
- How duplicate records were identified.
- How missing values were treated.
- Whether currency was adjusted for inflation.
- Whether a boundary or population estimate changed during the period.
- What software, spreadsheet formulas, scripts, or databases were used.
You do not need to publish every line of code in the article. If reproducibility matters, link to a cleaned dataset, code repository, calculation sheet, or downloadable methodology document when doing so does not expose personal information or create security risks.
If automated tools or artificial intelligence assisted with transcription, translation, classification, searching, or summarization, disclose the role they played. State what a human checked afterward. Do not claim that a tool “verified” information unless a person independently reviewed the relevant evidence.
A practical disclosure might say: “A transcription tool produced a first draft of the interviews. Reporters checked the audio against the transcript before quoting it.” That tells readers both what the tool did and where human responsibility remained.
State limitations and uncertainty clearly
A transparent note does not make a story look weak. It shows that the newsroom understands the boundaries of its evidence.
Describe limitations that could reasonably change how readers interpret the findings. Common examples include:
- Records were unavailable for some months or locations.
- The source used changing definitions over time.
- The dataset records reported incidents, not all incidents.
- Some people declined interviews or did not respond.
- Public records may contain errors or later amendments.
- A sample was too small to support broad conclusions.
- Correlation does not establish causation.
- A video or image could be authenticated but not fully contextualized.
- Estimates are sensitive to assumptions or population changes.
Avoid vague disclaimers such as “There may be limitations.” Name the limitation and explain its practical effect. For example: “Because two counties did not provide comparable records, the statewide comparison excludes them and should not be read as a complete statewide total.”
Separate unknowns from disagreements. An agency may dispute your interpretation while the underlying document remains available. Conversely, a claim may be impossible to settle because the relevant record does not exist. Readers benefit from knowing which situation applies.
Write the note for ordinary readers
Use direct verbs and short paragraphs. Replace institutional language with concrete descriptions:
- “We requested” instead of “A request was submitted.”
- “We counted” instead of “Counts were generated.”
- “We could not verify” instead of “Verification was not possible.”
Put the most important methodological information first. Readers should not need to reach the final paragraph to discover that your dataset covered only one city or that a key estimate was based on a small sample.
Keep the note separate from advocacy. Do not use it to attack critics, praise the newsroom, or repeat the article’s main argument. The purpose is to explain how the reporting was done and how confidently the evidence supports the claims.
Before publication, ask a colleague who was not involved in the reporting to read the note. Have them identify what they still cannot tell from it. If they cannot determine the period, selection rule, or verification process, revise those points before polishing the prose.
Troubleshoot common methodology problems
If you cannot describe the source clearly, pause and trace the information back to its earliest reliable origin. A widely repeated figure may have been copied from an outdated or misinterpreted report.
If the numbers in your notes do not match the published total, check for duplicate records, revised files, inclusive versus exclusive date ranges, rounding, and filters applied in different spreadsheet tabs. Preserve the original files and record each transformation.
If sources disagree, do not silently choose the number that best fits the story. Present the disagreement, explain which source you used, and give the reason—for example, a later corrected filing or a definition that more closely matches the story’s question.
If a limitation is serious enough to undermine the central claim, change the claim rather than hiding the limitation. Narrow “the program reduced crime” to “crime reports fell during the period studied,” if that is all the evidence supports.
If confidentiality prevents you from describing the reporting process in detail, provide as much non-identifying information as is safe: the source’s relevant role, the nature of access, the number of independent confirmations, and the evidence that can be published.
Final editing checklist
Before publishing, confirm that the methodology note:
- States the reporting question and scope.
- Identifies major sources and relevant dates.
- Explains inclusion, exclusion, and selection rules.
- Describes verification of central claims.
- Defines important calculations and terms.
- Discloses meaningful use of software or automated tools.
- Identifies missing information and important limitations.
- Distinguishes confirmed facts from allegations, estimates, and unresolved claims.
- Uses plain language without overstating certainty.
- Does not reveal confidential or personally sensitive information.
- Matches the methods actually used in the reporting.
The final test is simple: could a careful reader understand what you examined, what you did with it, and where the evidence stops? If the answer is yes, the methodology note is doing its job.