Meanings and Consequences in Standard Setting

Standard setting is defined as the identification of certain points on a mark scale with particular performance standards, with the intention of enhancing the inferences that are warranted from the test scores. It is argued that the selection of both the points on the mark‐scales and the performance standards with which they are equated are arbitrary and are driven by a set of values (which are often implicit). In ‘high‐stakes’ settings, it is shown how the values implicit in the standard can come to dominate the values inherent in the domain they represent. The validation of standards must therefore include consideration of their consequences as well as their meanings. It is then argued that standards, where they exist, cannot be accounted for purely in terms of norm‐referenced or criterion‐referenced interpretations, but exist rather by virtue of a shared construct in a community of practice. These theoretical positions are then developed to classify standard‐setting methods along two dimensions, the first relating to the role of performance data in the setting of standards and the second relating to the extent to which the meanings or the consequences of the assessment are emphasised in the process.