R0049/2026-03-31/Q001-H3¶
Statement¶
Partial implementations of analytical rigor frameworks exist as published AI/LLM prompts, but none achieve comprehensive coverage of a full research methodology — they address individual techniques or single phases of the research process rather than end-to-end analytical rigor.
Status¶
Supported. This is the best-supported hypothesis. The evidence reveals a landscape of narrow, single-technique implementations with no published example of a unified framework.
Supporting Evidence¶
| Evidence | Summary |
|---|---|
| SRC04-E01 | Roberts implemented 3 individual SATs (ACH, Starbursting, Key Assumptions Check) as separate LLM tools — partial, not unified |
| SRC02-E01 | Framework Chain-of-Thought addresses screening only, not full research lifecycle |
| SRC01-E01 | PRISMA-trAIce is a reporting checklist for AI in systematic reviews, not a system prompt |
| SRC05-E01 | Agent Laboratory automates research phases but without analytical rigor mechanisms |
Contradicting Evidence¶
| Evidence | Summary |
|---|---|
| — | No evidence was found that directly contradicts this hypothesis |
Reasoning¶
Every implementation found falls into one of these categories:
- Single-technique tools: Roberts' SAT implementations address individual structured analytic techniques but do not compose them into a methodology
- Phase-specific prompts: Framework Chain-of-Thought and other screening prompts address one phase of systematic review
- Reporting checklists: PRISMA-trAIce defines what to report when AI is used, not how to conduct the analysis
- Research automation platforms: Agent Laboratory, STORM, PaperQA2 automate research tasks but lack formal analytical rigor mechanisms (no bias assessment, no calibrated probability, no self-audit)
The gap between existing implementations and a comprehensive framework is substantial. No published work combines ICD 203 probability calibration, GRADE evidence scoring, ACH hypothesis testing, ROBIS self-audit, and search transparency logging into a single system prompt.
Relationship to Other Hypotheses¶
This hypothesis subsumes H1 (partial support exists) and H2 (which is eliminated). H3 provides the most accurate characterization of the evidence landscape.