In decision analysis, assessing a decision basis element typically involves two steps: at the level of structure, it involves defining the distinction represented by that element and the relation of that distinction to other distinctions within the basis; at the levels of function and number, assessing a decision basis element involves determining its internal value or values. One speaks of assessing particular decision basis parameters (eg, probability distributions, risk-aversion coefficients, preference trade off parameters), of assessing nodes in a relevance diagram, and of assessing full relevance diagrams. In particular, a decision relevance diagram is typically assessed in a backward (or top down) fashion. This backward process starts by producing a minimal decision relevance diagram, producing from it an incipient decision relevance diagram, and continuing by directly or indirectly assessing the diagram's frontier nodes until there are no unassessed nodes.

See also: Background State of Information, Basic Relevance Diagram.