Quellmalz suggests that rapid adoption and implementation of performance assessment may cause test developers to overlook the importance of detailing adequate evaluative criteria that are able to measure growth according to clear target performances. Standards based on professional consensus must be maintained in order to motivate decisions about common performances. Quellmalz proposes a set of characteristics that should be met in order to ensure the validity of assessment criteria.
(1) Significance ensures that criteria involve a representative sample of knowledge and strategies from the real-world target domain; cognitive, metacognitive, and dispositional components of a performance task must be considered in evaluating an examinee's development.
(2) Fidelity of the performance criteria must be maintained by replicating as much as possible the tasks, conditions, expectations, and quality levels of real-world performance situations.
(3) Generalizability of the performance criteria is based on the representativeness of scoring rubrics across domains and within domains, and the rubrics should be in line with general educational practice (or vice versa); students, teachers, and raters should all have a common understanding of what criteria are valued in what way.
(4) Developmental appropriateness of performance assessment should cause criteria to fall in line with theoretically motivated milestones in learner development; bands of developmental abilities should be descriptively represented in performance criteria and should focus on accomplishments at each level (as opposed to weaknesses).
(5) Accessibility of the criteria must be ensured by using language that is clear to all of the audiences involved (students, teachers, parents, etc.); this should enable some kind of evaluative process to be engaged in by all interested parties.
(6) Utility of the criteria should revolve around addressing performance features that can be effectively addressed through instruction and practice within a reasonable amount of time.
Creation of performance assessment criteria should begin early on in the test development process, and it should incorporate professional and expert advice, information from former assessments, and comparison of actual examinee work for calibration of performance levels.
