Do performance standards clearly articulate levels of achievement?
It is tempting but misleading to base performance standards solely on the percentage of items on a test that are answered correctly. While there is some logic to this approach, it may result in misleading information because the referent is not clearly defined. Good performance standards are defined by examples and benchmarks that illustrate qualitatively different levels of proficiency. It is not enough to simply say, for example, only 36% of students are reading at or above the basic level. From a policy perspective, to understand what this result actually means, there is a need to examine how basic was defined in terms of actual student performance, and what a basic reader is able to do that students who fail to meet this level are not able to do. • Are performance standards based on absolutes rather than normative comparisons? While normative comparative information is valuable to help set performance standards, the standards themselves must refer to a clearly defined level of prof
