Extension: A theory and methodology to quantify knowledge
Fanelli extends Kaal's identification limit for null evidence. A null or negative result can rule out a tested hypothesis under the stated design. It does not thereby select a unique explanation. Fanelli models the knowledge value of a negative result as declining with the number of hypotheses, methods, explanations, or outcomes that remain untested. Real experiments retain uncertainty because several alternative hypotheses can remain compatible with the same outcome. The problem persists even with greater statistical power when auxiliary assumptions remain uncontrolled. The correspondence is narrow. Fanelli supplies a general meta-scientific account of negative-result informativeness. The article does not test Kaal's computational economy, reconstruct his empirical condition, or identify the mechanism behind any particular null. It supports the distinction between disciplining a condition and establishing a mechanism. A null can remove or weaken one candidate explanation. Mechanism identification requires additional designs that discriminate among the remaining alternatives.
research-methodsscholarly-growth-coveragescholarly-literaturenull-resultsmechanism-identificationscientific-inferenceevidence-provenance