Validity

From EvaluationWiki

A key property of, most commonly in the present context, a test, claim, argument, interpretation, or evaluation. Roughly speaking, this means the truth (or soundness) of these things or of some assertion about them, often built into their name or title. Hence the validity of a test is the truth of the claim that it is, for example, a(n appropriate) test of middle-school mathematical ability. There is a technical sense of the term in logic that we ignore here, staying with the useful informal notion of the validity of an argument as meaning that it is sound, at least in enough of its premises and its inferences to establish its conclusion(s). In experimental design, the study of investigation structures, there is a well-known technical distinction between internal and external validity. In test thory, a number of different types of validity have been distinguished: construct validity, predictive validity, face validity, etc. The current view Is that they are all best seen as variations on the same notion, adjusted for particular contexts, especially for particular tasks and levels of analysis. Validity is close in meaning to accuracy, although the latter tends to connote precision, which does not always apply to the former. It does always require relevance, though it does not always imply certainty or finality; one may say that a claim provides valid support for a certain conclusion, without committing to the view that there are no other arguments pointing towards a different conclusion. This point is often put by saying that a valid argument or evaluation at least provides a prima facie case for its conclusion. However, to be a valid summary of the arguments on a particular topic, or an evaluation, it is expected that all currently available considerations bearing on the conclusion have been considered.