Source Quality Rubric
Rating System (A-E)
A - Highest Quality (Primary Sources)
| Criteria |
Examples |
| Peer-reviewed systematic reviews |
Cochrane Reviews, PRISMA-compliant studies |
| Meta-analyses |
Statistical aggregations of multiple studies |
| Randomized controlled trials |
Clinical trials, A/B testing at scale |
| Official government publications |
FDA, NIH, WHO, regulatory bodies |
| Major institution research |
MIT, Stanford, Google Research, OpenAI |
Trust Level: Can be cited as definitive evidence
Verification: Check DOI, publication venue, author credentials
B - High Quality (Authoritative Sources)
| Criteria |
Examples |
| Peer-reviewed original research |
Journal articles with methodology |
| Official standards documents |
ISO, IEEE, W3C specifications |
| Established organization white papers |
Gartner, McKinsey, Big Tech research |
| Clinical/practice guidelines |
Medical associations, professional bodies |
| Official documentation |
Product docs, API references |
Trust Level: Strong evidence, verify with A-level if critical
Verification: Check author affiliation, funding disclosure
C - Moderate Quality (Expert Sources)
| Criteria |
Examples |
| Expert opinion pieces |
Named experts with credentials |
| Conference presentations |
Academic/industry conferences |
| Case studies |
Documented implementations |
| Reputable news analysis |
NYT, WSJ, BBC technology sections |
| Industry analyst reports |
Paid reports with methodology |
Trust Level: Useful for context, verify key claims
Verification: Check author expertise, potential bias
D - Lower Quality (Supporting Sources)
| Criteria |
Examples |
| Preprints |
arXiv, bioRxiv (not peer-reviewed) |
| Expert blog posts |
Personal blogs from domain experts |
| Conference abstracts |
Preliminary findings |
| Press releases |
Company announcements |
| Trade publications |
Industry news sites |
Trust Level: Use as leads, verify before citing
Verification: Cross-reference with higher-quality sources
E - Use with Caution (Weak Sources)
| Criteria |
Examples |
| Anonymous sources |
Unverifiable claims |
| Opinion without evidence |
Hot takes, speculation |
| Outdated information |
>3 years for fast-moving fields |
| Clear bias/conflicts |
Vendor marketing, advocacy sites |
| Social media posts |
Twitter/X threads, Reddit comments |
| AI-generated content |
Unverified LLM outputs |
Trust Level: Do not cite as evidence
Verification: Only use if no alternatives exist, flag uncertainty
Quality Assessment Checklist
For each source, evaluate:
Authorship
Publication
Content
Accessibility
Domain-Specific Adjustments
Technology/AI
- Preprints may be acceptable (field moves fast)
- Company research blogs (Google AI, OpenAI) often quality
- GitHub repositories count as primary sources for code
- Benchmark papers essential for performance claims
Medical/Health
- Peer-review mandatory for health claims
- Clinical trial registration required
- FDA/regulatory status must be verified
- Patient population and sample size matter
Business/Market
- Market size claims need methodology
- Analyst reports vary in quality
- Company financials from SEC filings (A-level)
- Press coverage alone insufficient
Legal/Policy
- Primary legislation and regulations (A-level)
- Court documents and rulings (A-level)
- Legal analysis from law firms (B-level)
- News coverage of legal matters (C-level)
Minimum Quality Thresholds
| Content Type |
Minimum Rating |
| Definitive claims |
A or B |
| Statistics/data |
A, B, or C with verification |
| Expert opinions |
C or higher |
| Background context |
D acceptable |
| Future predictions |
C with attribution |
Red Flags - Automatic Quality Reduction
- No author attribution → E
- No date → Reduce by 2 levels
- Broken citations → Reduce by 1 level
- Known retraction → E
- Predatory journal → E
- SEO-optimized content farm → E
- Circular citations (sources cite each other) → Investigate
1---2name: 2566-quality-rubric-5c8b32c03description: Source Quality Rubric4---5# Source Quality Rubric67## Rating System (A-E)89### A - Highest Quality (Primary Sources)10| Criteria | Examples |11|----------|----------|12| Peer-reviewed systematic reviews | Cochrane Reviews, PRISMA-compliant studies |13| Meta-analyses | Statistical aggregations of multiple studies |14| Randomized controlled trials | Clinical trials, A/B testing at scale |15| Official government publications | FDA, NIH, WHO, regulatory bodies |16| Major institution research | MIT, Stanford, Google Research, OpenAI |1718**Trust Level**: Can be cited as definitive evidence 19**Verification**: Check DOI, publication venue, author credentials2021---2223### B - High Quality (Authoritative Sources)24| Criteria | Examples |25|----------|----------|26| Peer-reviewed original research | Journal articles with methodology |27| Official standards documents | ISO, IEEE, W3C specifications |28| Established organization white papers | Gartner, McKinsey, Big Tech research |29| Clinical/practice guidelines | Medical associations, professional bodies |30| Official documentation | Product docs, API references |3132**Trust Level**: Strong evidence, verify with A-level if critical 33**Verification**: Check author affiliation, funding disclosure3435---3637### C - Moderate Quality (Expert Sources)38| Criteria | Examples |39|----------|----------|40| Expert opinion pieces | Named experts with credentials |41| Conference presentations | Academic/industry conferences |42| Case studies | Documented implementations |43| Reputable news analysis | NYT, WSJ, BBC technology sections |44| Industry analyst reports | Paid reports with methodology |4546**Trust Level**: Useful for context, verify key claims 47**Verification**: Check author expertise, potential bias4849---5051### D - Lower Quality (Supporting Sources)52| Criteria | Examples |53|----------|----------|54| Preprints | arXiv, bioRxiv (not peer-reviewed) |55| Expert blog posts | Personal blogs from domain experts |56| Conference abstracts | Preliminary findings |57| Press releases | Company announcements |58| Trade publications | Industry news sites |5960**Trust Level**: Use as leads, verify before citing 61**Verification**: Cross-reference with higher-quality sources6263---6465### E - Use with Caution (Weak Sources)66| Criteria | Examples |67|----------|----------|68| Anonymous sources | Unverifiable claims |69| Opinion without evidence | Hot takes, speculation |70| Outdated information | >3 years for fast-moving fields |71| Clear bias/conflicts | Vendor marketing, advocacy sites |72| Social media posts | Twitter/X threads, Reddit comments |73| AI-generated content | Unverified LLM outputs |7475**Trust Level**: Do not cite as evidence 76**Verification**: Only use if no alternatives exist, flag uncertainty7778---7980## Quality Assessment Checklist8182For each source, evaluate:8384### Authorship85- [ ] Author clearly identified86- [ ] Author has relevant credentials87- [ ] Author affiliation disclosed88- [ ] No undisclosed conflicts of interest8990### Publication91- [ ] Publication date clearly stated92- [ ] Publication venue reputable93- [ ] Peer review or editorial process94- [ ] Still current/relevant9596### Content97- [ ] Methodology described (if research)98- [ ] Data sources cited99- [ ] Claims supported by evidence100- [ ] Limitations acknowledged101102### Accessibility103- [ ] URL accessible and stable104- [ ] DOI available (for academic)105- [ ] Archive link created (for ephemeral content)106107---108109## Domain-Specific Adjustments110111### Technology/AI112- Preprints may be acceptable (field moves fast)113- Company research blogs (Google AI, OpenAI) often quality114- GitHub repositories count as primary sources for code115- Benchmark papers essential for performance claims116117### Medical/Health118- Peer-review mandatory for health claims119- Clinical trial registration required120- FDA/regulatory status must be verified121- Patient population and sample size matter122123### Business/Market124- Market size claims need methodology125- Analyst reports vary in quality126- Company financials from SEC filings (A-level)127- Press coverage alone insufficient128129### Legal/Policy130- Primary legislation and regulations (A-level)131- Court documents and rulings (A-level)132- Legal analysis from law firms (B-level)133- News coverage of legal matters (C-level)134135---136137## Minimum Quality Thresholds138139| Content Type | Minimum Rating |140|--------------|----------------|141| Definitive claims | A or B |142| Statistics/data | A, B, or C with verification |143| Expert opinions | C or higher |144| Background context | D acceptable |145| Future predictions | C with attribution |146147---148149## Red Flags - Automatic Quality Reduction150151- No author attribution → E152- No date → Reduce by 2 levels153- Broken citations → Reduce by 1 level154- Known retraction → E155- Predatory journal → E156- SEO-optimized content farm → E157- Circular citations (sources cite each other) → Investigate
tools-only