Standards in examinations: a matter of trust?

Publisher Website
Google Scholar
Abstract
The notion of a ‘standard’ in an assessment system is defined as the attachment of specific meanings to specific test or examination scores. When these meanings are in terms of the performance of a group of individuals, the standard can be described as norm‐referenced (when the individual is not a member of the reference group) or cohort‐referenced (when the individual is a member of the reference group), and when the meanings are in terms of explicit criteria, the standard is described as criterion‐referenced. It is then argued that none of these approaches adequately describes the way that standards are set in high‐stakes assessment systems, which are best described as construct‐referenced. Building on Cresswell's (1996) definition of examination comparability as being the extent to which the assessments are equivalently valued by users of test results, standard setting is regarded as neither objective nor subjective, but rather intersubjective, where maintenance of standards requires that those responsible for setting standards are full participants in a community of practice, and are trusted by users of assessment results.