A specialized skill for the researcher and fact-checker agents covering investigative research. Provides primary/secondary source collection, source reliability assessment, data triangulation, and bias analysis methodologies. Use for 'research,' 'fact-checking,' 'source verification,' 'investigative techniques,' and similar topics.
Specialized knowledge used by the researcher and fact-checker agents when investigating and verifying materials for documentaries.
A documentary's credibility depends on the depth and accuracy of its research. A single incorrect statistic can undermine the entire program's trust. There is a fundamental difference between "casually searching and using" versus "verifying at investigative journalism standards."
| Tier | Description | Reliability | Example |
|---|---|---|---|
| Primary | Primary sources — original data, direct interviews | Highest | Census raw data, firsthand testimony |
| Refereed |
| Peer-reviewed academic sources |
| High |
| Academic papers, meta-analyses |
| Institutional | Reports published by institutions | Medium-High | WHO reports, central bank statistics |
| Media | Verified news coverage | Medium | Major news outlet articles, reportage |
| Anecdotal | Anecdotes, personal experiences | Low | Social media, blogs, community posts |
| Criterion | Question | Weight |
|---|---|---|
| Currency | When was it published? Updated? | 20% |
| Relevance | Is it directly related to the topic? | 15% |
| Authority | What are the author/institution's expertise and credentials? | 25% |
| Accuracy | Does it match other sources? Is there evidence? | 25% |
| Purpose | Information? Persuasion? Advertising? Political purpose? | 15% |
| Score | Grade | Usability |
|---|---|---|
| 80-100 | A (Highly Reliable) | Can be used as core fact |
| 60-79 | B (Reliable) | Use as supporting material, cross-check recommended |
| 40-59 | C (Caution) | Reference only, cross-check required |
| 20-39 | D (Suspect) | Do not use directly, cite only as a perspective |
| 0-19 | F (Disqualified) | Do not use |
Method Triangulation
(Different research methods)
^
/ \
/ \
/ \
Data Triangulation --- Investigator Triangulation
(Different data sources) (Different analysts)
| Type | Method | Example |
|---|---|---|
| Data Triangulation | Confirm the same fact from different sources | Government statistics + academic paper + field observation |
| Method Triangulation | Verify the same phenomenon with different methods | Interview + statistical analysis + literature review |
| Perspective Triangulation | Confirm from different stakeholders' viewpoints | Government + NGO + affected parties |
| Bias Type | Detection Question | Countermeasure |
|---|---|---|
| Confirmation Bias | Are we only collecting data that supports our desired conclusion? | Intentionally search for counter-evidence |
| Survivorship Bias | Are we only looking at success stories? What about failures? | Investigate failure cases with equal weight |
| Selection Bias | Are we only representing a specific group? | Include materials from diverse demographic groups |
| Recency Bias | Are we overvaluing recent events? | Include historical context |
| Framing Bias | Are we defining the problem in only one way? | Try redefining with a different frame |
## Source #[Number]
- **Title**: [Material title]
- **Author/Institution**: [Name]
- **Publication Date**: [YYYY-MM-DD]
- **Type**: [PRIMA tier]
- **URL/DOI**: [Access path]
- **CRAAP Score**: [Score/100]
- **Key Citation**: "[Direct quote — quotation marks required]"
- **Summary**: [1-2 sentence key content]
- **Related Section**: [Which part of the documentary treatment will use this]
- **Cross-Check**: [Does it match other sources? Compare with Source #XX]