Abstract
A self-contained account is given of the implications among the sufficiency, conditionality and likelihood axioms of statistical evidence for the discrete case. These include a previously unpublished derivation of sufficiency from conditionality. The nondiscrete case is discussed with reference to the same relation, and to the significance of nonunique determination of density and likelihood functions. The writer's current views on this problem area are indicated briefly.

This publication has 0 references indexed in Scilit: