The Generalizability of Measurements: An Application to ’Blind Refereeing’ Process

This paper introduces generalizability theory as a means of assessing and improving the dependability (reliability) of the refereeing process used in journal article publication decisions. Concepts and computational procedures are presented and the power and flexibility of generalizability analysis are illustrated. It is shown that the refereeing process may be "optimized" by trading off costs and desired levels of generalizability.