Validating Assessment for Learning: Consequential and Systems Approaches

Abstract

In order to adequately evaluate assessment for learning, expanded approaches to validity need to be considered. The purpose of this manuscript is to explore what is necessary to evaluate claims that assessment facilitates learning. Messick’s (1994) concept of consequential validity provides one lens for determining the learning consequences of assessment. His approach suggests that learning consequences are a special case of the general concept of consequential validity and should be evaluated from that perspective. The systems approach developed by Frederiksen and Collins (1989) provides another perspective of how assessments can be designed with learning consequences in mind. Their model provides a transparent way to link assessments to learning by making teaching to the test a valid activity. The adequacy of both models for evaluating claims of learning from assessment is explored. Based on the analysis, a model for evaluating the validity of evidence for assessment for learning is outlined.