HIE / methods
Methods & standards
Open questions, explicit sources, independent review. Different kinds of content need different kinds of justification.
No Claim Without Evidence.
This does not prohibit new questions. It requires that hypotheses remain hypotheses until warranted, and that findings remain bounded by the evidence that produced them.
Six kinds of content
Research questions define what we want to understand. Conceptual discussion clarifies ideas and arguments. Hypotheses make testable proposals. Empirical findings report observations under stated conditions. HIE interpretations compare records and alternatives. Future scenarios explore possibilities, not predictions.
Four separate judgments
Source scope tells you what was read. Scientific evidence strength describes methodological support. AI assessment confidence describes a model's judgment, not truth. Human review status requires an actual reviewer decision. None can substitute for the others.
Read the source scope
Metadata-only, abstract-reviewed, full-text-available, full-text-AI-reviewed, and human-reviewed are distinct. Full-text availability is not full-text review. Human approval does not change the scope the AI read.
A workflow, not a live-status claim
Question → Find evidence → Compare interpretations → Experiment / review → Revise. Existing discovery and evaluation code is disabled by default. The public catalog and evidence links are inspectable; this page does not assert continuous processing.
Publication & revision
Public relationships must connect public, unarchived records. Rejected records are excluded from live reads. Claims and evidence have revision history; reviewers are authorized separately from sign-in. Cached pages and dated snapshots are not live monitoring.