Question
Define the claim before seeing the preferred answer.
Open discussions GLOBAL HEALTH · METHODS · EVIDENCE
Evidence
before certainty.
GHEAS builds methods and tools that make data analysis more objective, auditable and honest about what it cannot prove.
先让证据经得起检验，再让结论被世界看见。
GAEF · EVIDENCE SYSTEM PRE-VALIDATION V1
CLAIME0—E4Licensed by evidence
QUESTIONDATADESIGNFALSIFICATIONTRACEComplete
CONTRADICTIONSRetained
CAUSAL LANGUAGEGated
01 — THE METHOD COMES FIRST
A result is only as credible as the path that produced it.
Statistical software can calculate an answer perfectly while the research design, measurement or interpretation remains wrong. GHEAS therefore evaluates the whole evidence chain—not just the final coefficient.
Our first PM2.5 project is preserved, but it no longer dictates the method. The new framework is specified first, implemented as software, then tested against simulations, public benchmarks and real cases.
02 — ONE CONTINUOUS EVIDENCE TRAILFrom question to conclusion,
From question to conclusion,
nothing disappears.
Every decision is recorded. Every reasonable alternative remains visible. Every claim has a ceiling.
Provenance
Record where every observation and transformation came from.
Design
Separate description, association and causal identification.
Stress test
Run reasonable alternatives, placebos and failure gates.
Conclusion
License only the language the evidence can support.
03 — CURRENT WORKFramework, evidence,
Framework, evidence,
and validation in one system.
01
DISCUSSION PAPER · OPEN
Drinking-water arsenic and health
A bilingual independent reanalysis showing how reasonable modelling choices can change the apparent health effects of an intervention.
Read and comment →02FRAMEWORK · V1
GHEAS Auditable Evidence Framework
A pre-validation methodological blueprint for objective, reproducible and claim-calibrated analysis.
Explore the framework →03DISCUSSION PAPER · OPEN
PM2.5 and life expectancy
A bilingual dual-track reanalysis showing why a strong country-level association does not license a causal effect claim.
Read and comment →04EVIDENCE SYNTHESIS · ACTIVE
PM2.5 and mortality literature
A protocol-led review examining how design and measurement choices shape published health estimates.
Follow the research trail →04 — THE GHEAS STANDARD
“The purpose is not to produce a more impressive answer.Read the bilingual method paper ↗
It is to produce the most defensible answer the evidence allows.”