Criterion Validity
Criterion validity considers whether scores on the instrument agree with a definitive, “gold standard” measurement of the same theme. This option for validating a measure typically occurs when a new instrument is being developed as a simpler, more convenient alternative to an accepted measurement: can a self-report of anxiety replicate what a psychiatrist would have diagnosed? The new and the established approaches are applied to a suitable sample of people and the results are compared using an appropriate indicator of agreement. The comparison may be used in a summative manner to indicate the validity of the measure as a whole, or it may be used in a formative manner during the development of the new measure to guide the selection of items by identifying those that correlate best with the criterion. The latter forms part of “item analysis” and involves cumulating evidence on the performance of each item from a range of validity analyses outlined subsequently in this chapter.