Primary evidence before commentary
Original and controlling records are preferred over summaries and commentary where reasonably obtainable.
Evidence assurance
Evidence Standards for Responsible Reliance
If you are going to rely on a conclusion, the evidence behind it should be traceable, appropriately independent and strong enough for the consequence of the decision.
Information is not automatically evidence.
Source independence before volume
Repeated reporting does not create independent corroboration when several sources depend on the same origin.
Contradictions remain visible
Evidence that weakens the leading interpretation remains part of the assessment.
Alternative explanations are tested
Where more than one explanation fits the evidence, alternative explanations are examined.
Uncertainty is disclosed
Unavailable records, unresolved questions and material gaps are not hidden by confident drafting.
Reliance is explicit
The final assessment states what the evidence can and cannot responsibly support for the intended decision.
Model memory is not evidence.
Artificial intelligence, generated recollection, apparent familiarity, repeated statements or confident language cannot establish a material fact. Material propositions in professional work require an identifiable evidentiary basis.
Evidence in practice
See what the evidence standards do in practice.
The MH370 dossier shows how factual status, reliance levels, source independence, model dependence, contradictions and unresolved gaps are kept separate rather than blended into one confident narrative.
No conclusion stronger than the available evidence.
Discuss a MatterTHE EVIDENCE MUST FIT THE CONCLUSION.
Material factual propositions must be traceable to an identifiable evidential basis. Model memory, repetition and plausibility are not substitutes for evidence.
Primary sources matter.
Where relevant primary or controlling records exist, they ordinarily carry greater weight than commentary describing them.
Source independence matters.
Several publications copied from one origin may still represent one evidential source.
Contradictions remain visible.
Material contrary evidence is not removed to make a report cleaner.
Uncertainty is part of the result.
Professional research distinguishes what is known, supported, disputed, unsupported and unknown.
Absence of evidence is handled carefully.
Not finding a record does not automatically prove an event never occurred.