The research-to-output workflow for journalists addresses a problem that's different from most professions: the output — a published news article, investigation, or feature — often must be produced under significant time pressure while meeting the evidentiary standard that distinguishes journalism from speculation.
The workflow creates the infrastructure for both speed and accuracy: a knowledge base check that retrieves prior beat intelligence before any story research begins, a structured research process that targets the specific verification needs of each story, a source and document tracking discipline that maintains the evidentiary record, and a post-publication routing step that converts each story's research into permanent beat intelligence for the next story.
The Four Research-to-Output Failure Modes in Journalism
Failure 1 — Starting from scratch on every story. A new development occurs on the beat; the reporter begins researching the institutions and context without checking the knowledge base for prior coverage, established facts, or related story threads. Time is spent reconstructing context that's already been developed.
Failure 2 — Publication without adequate verification. A source provides a significant claim; the claim makes it into the published article on source attribution alone without independent verification or document corroboration. The claim is later found to be incorrect; the retraction damages credibility.
Failure 3 — Research not routed after publication. The reporting done for a story — institutional intelligence developed, source relationships established, documents obtained — is filed with the story and doesn't feed into the beat knowledge base. The next story on the same topic starts from scratch.
Failure 4 — Source attribution without reliability assessment. A source is quoted or cited without the reporter explicitly assessing what that source's perspective and interests are and what those create as potential blind spots in their statements. The story reflects the source's interests without acknowledging the potential bias.
Stage 1: The Research Anchor
Before any story research begins, the research anchor defines the story's core question and checks what the knowledge base already has.
STORY RESEARCH ANCHOR — [Date]
STORY TYPE: [Breaking news / Enterprise / Investigation / Feature]
CORE QUESTION: [What is this story answering? One sentence.]
PUBLICATION DEADLINE: [When must this be filed?]
RESEARCH TIME AVAILABLE: [Hours]
KNOWLEDGE BASE CHECK:
Institutions involved: [Found in KB? What's there?]
Related story threads: [Any prior threads — active, cold, published, killed — relevant?]
Relevant sources: [Source identifiers with expertise in this area?]
Prior coverage: [Published stories that this connects to?]
WHAT'S ALREADY ESTABLISHED: [What the KB and prior coverage already support]
WHAT NEEDS VERIFICATION FOR THIS STORY: [Specific claims that require independent support]
WHAT REMAINS UNKNOWN: [Open questions this story needs to address]
RESEARCH SCOPE:
[What specific research needs to happen for this story to meet evidentiary standards?]
[Targets: documents to obtain, sources to contact, public records to check]
The research anchor converts the story from an open-ended investigation into a targeted verification project: here's what we know, here's what we need to establish, here's how to get there.
Stage 2: Document and Records Research
Documents are the evidentiary foundation for strong journalism: they establish facts independently of any source's account, they can be reviewed by editors and legal counsel, and they don't change their account in the way that memory-based sources sometimes do.
Document research session:
DOCUMENT RESEARCH — [Date]
Story: [Thread/story identifier]
Time budget: [Hours]
Target: [What specific documents would support the story's core claims?]
SOURCES CHECKED:
[ ] Court records: PACER, local court systems — [Any filings relevant to this story?]
[ ] Federal regulatory filings: SEC EDGAR, FDA, EPA, FCC, etc. — [Relevant documents?]
[ ] State regulatory filings: State-specific databases — [Relevant documents?]
[ ] FOIA requests: [Any pending requests with relevant agencies? File new requests?]
[ ] Public financial records: IRS Form 990 for nonprofits; corporate filings — [Relevant?]
[ ] Government accountability offices: GAO reports, IG reports — [Any relevant reports?]
[ ] Academic and research databases: [Any relevant research?]
DOCUMENTS OBTAINED:
Document 1: [Description] | Date: | Source: | Significance:
Document 2: [...]
DOCUMENTS REQUESTED:
[ ] [FOIA request filed: agency, request date, tracking number, expected response]
WHAT DOCUMENTS ESTABLISH:
[What the documents obtained confirm, with document references]
[What remains to be established from other sources]
The URL and web source archiving discipline:
Web sources — company websites, organization pages, government databases — are captured with a date-stamped screenshot or archive.org snapshot at the time of research. Pages change; published articles are embarrassed by sources that reference web content that's since been modified or removed.
Stage 3: Source Development and Interview
Sources provide the human dimension of journalism — the context, the motivation, the accounts of what happened — that documents often can't provide alone.
Pre-interview research:
SOURCE PREPARATION — [Date]
Source identifier: [Pseudonym/identifier]
Interview type: [On record / Background / Off record — confirmed before interview]
Relevant context: [What this source's KB entry says about their reliability and perspective]
INTERVIEW OBJECTIVES:
Primary objective: [The most important thing this interview needs to establish]
Secondary objectives: [Other questions this interview should address]
KEY QUESTIONS:
1. [Most important question]
2. [Second question]
3. [...]
VERIFICATION TARGETS:
[What specific claims does this source have firsthand knowledge to support?]
[What do I want this source to confirm, challenge, or contextualize?]
ATTRIBUTION STANDARD CONFIRMED:
[ ] On record — name and title — confirmed before interview
[ ] On background — no name, role/description — confirmed before interview
[ ] Off record — for context only, not for publication — confirmed before interview
During and after interview:
INTERVIEW NOTES — [Date + Source identifier]
KEY STATEMENTS (verbatim for significant quotes, paraphrase flagged as such):
[Statement 1: exact quote OR "Paraphrase: [content]"]
[Statement 2: ...]
CONTEXT PROVIDED:
[Background context the source provided that's not for direct attribution]
RELIABILITY ASSESSMENT:
[Did this source's account align with or contradict the documents?]
[Were there gaps or inconsistencies in their account?]
[Any perspective or interest that should influence how to weigh their statements?]
FOLLOW-UP NEEDED:
[ ] [What requires verification from another source or document?]
[ ] [Any claim that needs a response from the subject of a critical claim?]
Stage 4: Verification and Fact-Checking
The verification stage is what distinguishes journalism from rumor and aggregation.
The verification protocol:
STORY VERIFICATION CHECKLIST — [Date]
For each factual claim in the story that will be published:
Claim: [The specific claim as it will appear in the article]
Source: [Where this claim comes from]
Verification status:
[ ] Corroborated: [Second independent source or document that confirms]
[ ] Single source: [Only one source, no independent document corroboration]
[ ] Document-supported: [Document that independently establishes this]
[ ] Alleged: [Being reported as an allegation, not a verified fact]
SUBJECT RESPONSE PROTOCOL:
[ ] Subject of significant critical claims contacted for response
[ ] Response received: [What they said]
[ ] Deadline provided for response: [When]
[ ] No response received (document in file)
LEGAL REVIEW REQUIRED?
[ ] Does this story contain claims that could be defamatory if wrong?
[ ] Does it involve private individuals' private lives?
[ ] Does it involve national security concerns?
→ If yes to any: legal review before publication
The single-source discipline:
Claims with a single source — no document corroboration, no independent second source — are either withheld from publication or published with explicit acknowledgment that they're single-sourced (typically: "according to a person familiar with the matter"). This isn't a prohibition on single-source reporting; it's a transparency discipline about what level of evidentiary support a claim has.
Stage 5: Story Production
With research complete and verification documented, the story is drafted and edited. The research-to-output workflow's contribution to this stage is that the evidence is organized before drafting begins — the verification checklist is complete, the key quotes are identified, and the story's structure flows from the established facts.
The pre-draft organization:
- The verification checklist is complete for all significant claims
- The most important documents are identified and quotation-ready
- The key quotes (verbatim) are identified from the interview notes
- The story's narrative structure is clear from the research
Stage 6: Post-Publication KB Routing
After publication, route any new beat intelligence from the story's research to the knowledge base before the story folder is filed.
POST-PUBLICATION ROUTING — [Date]
Story: [Headline or working title]
Publication date: [Date]
INSTITUTION ENTRIES TO UPDATE:
[ ] [Institution 1]: [New intelligence from this story's reporting]
[ ] [Institution 2]: [...]
STORY THREAD HISTORY:
[ ] Published thread updated: [Status → Published; link to published story]
[ ] Any related threads updated with new intelligence from this story?
ANALYTICAL FRAMEWORK:
[ ] Any new pattern or dynamic observed in this story worth adding to the beat framework?
SOURCE ASSESSMENTS:
[ ] Any source reliability updates based on this story's outcomes?
[Did sources' accounts prove accurate when independently verifiable? Any discrepancies?]
NEW STORY THREADS:
[ ] Did this story surface threads that weren't pursued? → New thread entries created
A Worked Example: Breaking News Investigation
Context: a regulatory filing released late Friday afternoon contains data suggesting a major healthcare company has been systematically underpaying claims. The reporter has two hours before the evening's digital deadline.
Research anchor (10 minutes):
KB check: the healthcare company has an institution entry with prior coverage history. Two related story threads are documented (one published 18 months ago, one cold). Source identifiers: three sources with healthcare regulatory expertise are active in the beat KB. Research anchor defines: what this filing specifically claims, what independent verification is achievable in two hours, and what goes in as an allegation vs. a verified fact.
Document research (30 minutes):
The filing itself is the primary document. Two prior regulatory filings from the same agency are retrieved from the regulatory database and provide context. One prior enforcement action against a different company for similar conduct provides regulatory context.
Source development (45 minutes):
Two sources from the beat KB with regulatory expertise are contacted by phone. Source A (on background) confirms the filing's significance from a regulatory standpoint and provides context on typical enforcement timelines. Source B (on background) provides industry context on the practice alleged. Company contacted for comment with a 30-minute response deadline; spokesperson declines to comment on pending regulatory matters.
Verification (15 minutes):
Verification checklist: the filing's existence and content — document, verified. The regulatory significance — Source A on background, corroborated by prior enforcement precedent in documents. The company's response — documented attempt and "decline to comment." The claim about the impact on claimants — single-sourced, published as "according to the filing" (document-supported, not independently corroborated through additional sources).
Story production (20 minutes):
Drafted from organized research with verification notes complete. Pre-draft organization was done during verification stage.
Post-publication routing:
Institution entry updated: new development documented with link to published story. Cold story thread updated: new development reactivated; thread status changed to Active.
Key Takeaways
-
The journalism research-to-output workflow has six stages: research anchor with KB check, document and records research, source development and interview, verification and fact-checking, story production, and post-publication KB routing.
-
The verification checklist is what maintains the evidentiary standard under time pressure. Rather than holding all verification in the reporter's head, the checklist makes explicit which claims are corroborated, which are single-sourced, and which are published as allegations.
-
The research anchor's KB check converts each story from a from-scratch investigation into targeted gap-filling. For experienced beat reporters with populated knowledge bases, the knowledge base often provides 30-60% of the contextual research that a new reporter would need to conduct from scratch.
-
Source attribution standards are confirmed before the interview begins. On record, background, or off record: confirmed before the interview starts, not negotiated after. This prevents the post-interview dispute where a source claims they said something off record that the reporter understood as attributable.
-
Post-publication KB routing converts each story's research into permanent beat intelligence. The institution profiles, source assessments, and story threads updated after each published story ensure that the next story on the same beat starts from a richer baseline.
Conclusion
The research-to-output workflow for journalists converts the story development process from an ad hoc, deadline-driven exercise into a structured sequence that starts from accumulated beat intelligence, targets research at the specific verification needs of each story, maintains the evidentiary record that supports publication, and routes the story's research intelligence back into the beat knowledge base for the next story. The verification checklist is the discipline that maintains accuracy under deadline pressure; the post-publication routing is the discipline that converts each story from an isolated output into a contribution to a compounding beat intelligence system.
See also: Clip Articles for Later Reading.