How we grade the evidence.

The six rungs

Each rung is scored from 0 to 3: 0 absent, 1 weak or indirect, 2 partial, 3 strong.

1. 1
   ObservationWas the phenomenon itself directly observed or measured, or only its supposed results?
2. 2
   HypothesisIs the claim stated precisely enough that it could turn out to be wrong?
3. 3
   PredictionDid it make a risky prediction before the data came in, and did the prediction succeed?
4. 4
   ExperimentHas it been tested under controlled conditions, with something to compare against?
5. 5
   ReplicationHave independent groups reproduced the result?
6. 6
   FalsifiabilityWere competing explanations considered and ruled out? Is there any result that would count against it?

The five grades

The letter grade is an editorial judgment guided by the six scores. It is not a sum. A claim with a strong observation but no test of its key step can't score above C, however well the rest is argued.

• ADemonstrated
  DemonstratedDirectly observed, experimentally tested and independently replicated, with alternatives ruled out.
• BStrongly supported
  Strongly supportedObserved and tested, with minor gaps, such as limited prediction or alternatives only partly addressed.
• CPlausible inference
  Plausible inferenceBuilt on real observations, but the key step is inferred rather than observed or tested.
• DExtrapolation
  ExtrapolationExtends an observed process far beyond the range in which it was actually observed.
• EAssumption
  AssumptionNo direct observation; the conclusion follows from the framework assumed at the start.

The reasoning check

The ladder grades the evidence. A second check looks at the reasoning: the step from the evidence to the conclusion. Each claim card names the logical fallacies it finds, in the claim as taught and in the alternative reading alike. The same labels appear on the Map, on each pillar, on individual pieces of evidence and sources, and in the Evaluator. Each label is marked with a red query mark; click one for its definition.

?Equivocation?Reification?Affirming the consequent?Begging the question?False dilemma?Hidden theological premise?Unwarranted extrapolation?Cherry-picking?Just-so story?Ad hoc rescue?Argument from ignorance?Argument from incredulity?Appeal to consensus?Genetic fallacy?Straw man

7 families, 15 fallacies. The full catalogue, with examples from both sides →

Principles

• Steelman firstEach claim is stated in its strongest form, from the textbooks that teach it, before it is examined.
• Separate the observation from the extrapolationMany claim cards contain two claims in one: a modest one that is observed, and a sweeping one that is inferred from it. They are graded separately.
• Say what it would takeEvery grade below A comes with the observation or experiment that would raise it.
• Where the evidence is strong, say soObserved adaptation, selection and variation often grade A or B. The cards say so plainly.

Every claim card is a draft until reviewed by a specialist. Corrections are welcome and will be logged openly.

See the rubric at work: Pillar 4 →