Summary

This chapter began with discussion of the two perspectives that have traditionally been used to study the validity of Criterion-referenced tests: content validity (including the theoretical arguments approach and expert judgments approach) and construct validity (including intervention studies, differential groups studies, and hierarchical-structure studies). These five validity strategies were defined and discussed and examples were provided to show how they can be applied in real-life language programs. A section on the importance of combining several validity strategies was also included. Next, Messick’s (1988) expanded view of validity was covered in terms of the evidential and consequential bases of test interpretation and use. Cronbach’s (1988) view that studies of validity ought to raise questions about the functional, political, economic, and explanatory perspectives on testing was also explored. The chapter ended with a discussion of strategies for making valid decisions with Criterion-referenced tests, including topics such as traditional cut-points and grading on a curve, the arbitrariness of cut-points, definitions of standards setting, alternative approaches to the setting of standards, as well as the relationships between standards setting and Criterion-referenced decision making, between standards and dependability, and between validity and Criterion-referenced decision making.