Methods of setting standards for complex performance-based assessments

This study investigates some procedures for setting performance standards. Both standard-setting methods for unidimensional, dichotomously scored tests (e.g., multiple-choice credentialing tests) and multidimensional, polytomously scored complex performance-based assessments are considered. Two procedures are distinguished for unidimensional tests: judgemental and empirical-judgemental methods. Nedelsky|s and Angoff|s methods are addressed in the first type of procedure, whereas the borderline group and contrasting-groups methods are dealt with in the second type of procedure. Three recently proposed procedures are reviewed for multidimensional complex assessments: the judgemental capturing policy, the extended Angoff procedure, and the dominant profile method.