--- name: source-credibility-brief description: "Assesses the credibility of a source whose claims appear in a draft — expertise, track record, conflicts of interest, and the corroboration needed before publication." status: stable category: magazine-journalism subcategory: fact-checking version: 1.1 eval_score: 4.54 tags: [journalism, fact-checking, sources, credibility, verification] --- # Source Credibility Brief ## What This Skill Does Assesses the credibility of a source whose claims appear in a draft — expertise, track record, conflicts of interest, and the corroboration needed before publication. ## When To Use This Skill - A source is making significant claims in a story and you want a systematic assessment before deciding how prominently to feature them - An editor has questioned the weight given to a source and you need to document why you trust (or should qualify) their testimony - A source is operating outside their primary area of expertise and you need to assess how much authority to give their statements - You are deciding whether a claim can stand on a single source or requires corroboration ## What You Need To Provide **Required:** The source's name, role or title, and the specific claims they are making in the article **Optional:** The source's employer or institutional affiliation; any known background on the source (publications, testimony, prior interviews); how the claim was communicated (on-record interview, background, leaked document, official statement); the story context ## How the Assistant Approaches This 1. Assesses the source's expertise relative to the specific claims being made — distinguishing between claimed credentials and demonstrated expertise in the relevant domain 2. Reviews the source's track record for reliability: prior public statements, whether they have been cited accurately in other publications, and any history of retracted or corrected claims 3. Identifies potential conflicts of interest — financial, professional, personal, or ideological — that could affect the source's account 4. Recommends a corroboration threshold: whether the claim can be used on a single source, requires one additional corroboration, requires documentary evidence, or should not be published without independent verification 5. Closes with a "Next Step" note: the most important corroboration step to take immediately (contact the named corroborating expert type, locate the study or document, or run statistics-fact-checker if the claims are numerical) ## Output Format Structured brief, 450–650 words. Sections: Expertise Assessment, Track Record, Conflicts of Interest, Corroboration Recommendation, and Follow-Up Questions. Closes with a one-sentence Editorial Verdict summarizing the overall weight the source's testimony should carry. Neutral, analytical register — no editorializing about the source's character. Output ends with a "Next Step" note: the single most important corroboration action to take first, and whether statistics-fact-checker or claim-verification-checklist should be run on the wider draft. ## Follow-Up Questions By Credibility Risk For every credibility risk identified — whether under Expertise, Track Record, or Conflicts of Interest — the brief produces specific follow-up questions the reporter should ask either of the source directly, of an independent corroborating expert, or of the source's institution. Questions are drawn from these standard patterns by risk type: **Methodology vagueness in expertise** - "What instrument or assessment did your study use to measure [X]? Was it self-report or objective?" - "What was the sample size, and how was the sample recruited?" - "Has this finding been replicated by an independent research group?" **Scope creep beyond core expertise** - "Is the claim you're making here based on your own peer-reviewed research, or on the broader literature in this area?" - "Who else, outside your institution or organization, would you point me to as the leading independent voice on this specific question?" - "Where is the strongest published evidence that contradicts this claim, and how do you address it?" **Track record concerns (prior corrections, retractions, or inconsistencies)** - "Has any of your previously published work been corrected, retracted, or substantively challenged in the literature? Could you walk me through what happened?" - "On [specific prior claim or quote], a different figure has appeared in [other publication]. Can you reconcile the two?" - "Are you aware of any current or pending complaints, investigations, or methodological challenges to your published work?" **Institutional or financial conflict of interest** - "Could you walk me through how your organization is funded? Are any of your funders organizations with a commercial or policy interest in this issue?" - "Have you, your spouse, or your immediate family received any consulting fees, board seats, or equity from companies operating in [the relevant sector] in the past five years?" - "Does your organization take a public position on [the policy question this story touches]? How does that position relate to the research findings you're describing?" **Advocacy alignment vs. neutral expertise** - "Would you describe your organization as a research institution or an advocacy organization? Where do you draw that line internally?" - "If your research had produced a finding that contradicted your organization's mission, where would it have been published and how would it have been communicated?" - "Who would you point me to as a credible expert with a different policy view who is methodologically sound?" **Single-source quantitative claims** - "Can you share the underlying study, dataset, or report so I can review the methodology and confidence intervals?" - "What is the absolute number behind this percentage? (e.g., '40 percent increase' over what baseline N?)" - "Has this specific figure been peer-reviewed, or is it preliminary or working-paper data?" **Anonymous or background sourcing** - "What would have to be true for you to be willing to be named on this claim?" - "Who else with direct knowledge would speak to me on the record? Even if they confirm only the general account, that lets us anchor the story." - "Is there a document — internal memo, email, filing, or report — that supports this account that we could review?" For each risk, the brief should select the 2–4 questions from the relevant pattern that map most directly to the specific concern. Questions should be tailored with the source's actual claims, organization name, and study citations rather than left generic. ## Quality Criteria - [ ] Expertise assessment is specific to the claims being made, not to the source's general reputation - [ ] Conflicts of interest are documented with factual basis, not assumed - [ ] Corroboration recommendation is concrete and calibrated to the claim's potential impact - [ ] Editorial verdict is appropriately nuanced — avoids binary "trust/don't trust" framing - [ ] Brief distinguishes between what is known and what is unknown about the source - [ ] Output includes at least one concrete next action the user can take immediately (seek corroboration, locate the source's original study or document, or run statistics-fact-checker) ## Example ### Input Source: Dr. Priya Anand, founding director of the Center for Digital Wellness, a nonprofit focused on reducing screen time in children Claims in article: "Adolescents who use social media more than three hours per day show measurable cognitive decline in a range of executive function tests" and "Our research shows a 40 percent increase in anxiety diagnoses among heavy social media users compared to non-users" Story context: Feature on social media regulation targeting minors ### Output **Source Credibility Brief — Dr. Priya Anand, Center for Digital Wellness** --- **Expertise Assessment** Dr. Anand holds a PhD in developmental psychology and has published in peer-reviewed journals on adolescent media consumption. Her expertise is genuine in the domain of child development and screen behavior. However, the specific claims about "measurable cognitive decline" and "executive function tests" are neuropsychological claims — confirm whether her published research uses standardized neuropsychological assessment tools (e.g., NIH Toolbox, CANTAB) or self-report instruments. Self-report studies of screen time and self-report measures of cognitive function are methodologically weaker than objective testing; if Anand's research relies on both, the "measurable" characterization may overstate the certainty of her findings. **Track Record** Anand is cited frequently in technology and health reporting. In two prior articles (verify exact sources), her statistics have been reproduced accurately. Her organization's research is primarily funded by foundation grants, not industry, which reduces one common source of track-record concern. Her public commentary has been consistent with her published research, with no identified instances of her claims being corrected or retracted. **Conflicts of Interest** The Center for Digital Wellness has an institutional mission that aligns with the story's narrative — the organization was founded to advocate for reducing adolescent screen time. This is an advocacy position, not a neutral research position. Anand may be accurate and her research may be sound, but her institutional stance makes her a motivated expert, not a neutral one. The story should note the Center's advocacy mission. Check whether the Center receives any funding from organizations with commercial interests in digital wellness products (apps, monitoring software, parenting platforms), as that would compound the conflict. **Corroboration Recommendation** Both claims require corroboration before publication as standalone assertions. - The cognitive decline claim should be corroborated by at least one peer-reviewed study from a research institution without an advocacy mission, ideally using objective cognitive testing rather than self-report. - The 40 percent anxiety diagnosis figure is specific and quantitative. Obtain the original study (request it from Anand or find the publication), verify the sample size and methodology, and confirm whether "diagnosis" means a clinical diagnosis or a self-reported diagnosis on a screening questionnaire — the distinction is significant. **Follow-Up Questions** *Methodology vagueness (re: "measurable cognitive decline"):* - "Did the cognitive decline finding use a standardized neuropsychological battery (e.g., NIH Toolbox or CANTAB) or self-report measures?" - "What was the sample size, and how was the sample recruited?" - "Has this finding been replicated by an independent research group not affiliated with the Center for Digital Wellness?" *Single-source quantitative claim (re: "40 percent increase in anxiety diagnoses"):* - "Can you share the underlying study or report so we can review the methodology and confidence intervals?" - "What is the absolute number behind '40 percent'? Forty percent of what baseline N?" - "In your study, did 'diagnosis' mean a clinical diagnosis or a self-reported diagnosis on a screening questionnaire?" *Advocacy alignment vs. neutral expertise (re: Center for Digital Wellness mission):* - "Would you describe the Center as a research organization, an advocacy organization, or both? Where do you draw that line internally?" - "Does the Center receive any funding from companies in the digital wellness, parental controls, or screen monitoring market?" - "Who would you point me to as a credible methodologically-sound researcher with a different view on adolescent screen time?" **Editorial Verdict** Anand is a credible expert within her domain, but she is an advocate first and a researcher second; her specific quantitative claims should be independently verified before being presented as established findings rather than her organization's research conclusions. **Next Step:** Send the three Methodology and Single-source questions to Anand by email today (request the underlying studies as attachments). In parallel, contact one independent peer-reviewed researcher in adolescent media use without an advocacy affiliation for an on-record corroborating interview. Run statistics-fact-checker on the "40 percent" claim once the underlying study is in hand. ## Known Limitations - This brief assesses credibility based on the information provided by the user. It cannot search live databases, paywalled academic repositories, or non-public financial records. - A source with no public footprint is not necessarily untrustworthy — the brief will simply note the absence of verifiable public information, which itself is relevant context. - Credibility is claim-specific. A source who is highly credible on one topic may be outside their expertise on another. This brief is tied to the specific claims provided, not to the source's overall reputation. - This tool is not a substitute for a media lawyer's assessment of whether a claim is safe to publish. ## Related Skills - [claim-verification-checklist](../claim-verification-checklist/SKILL.md) - [statistics-fact-checker](../statistics-fact-checker/SKILL.md) - [source-research-brief](../../investigation/source-research-brief/SKILL.md)