States use achievement level descriptors (ALDs) to delineate what knowledge, skills, and abilities (KSAs) found in the standards that a student should possess depending upon the student's level of achievement (e.g. Basic, Proficient, Advanced; Perie, 2008; Lewis & Green, 1997). Bejar, Braun, and Tannenbaum (2007) posit, however, that ALDs should be developed as a conceptualization point that connects and drives a standards and assessment system. If ALDs are developed in conjunction with content standards (or shortly thereafter), then states can use the ALDs, themselves, as a tool to aid in the development of test blueprints, test items, and in the reporting of results.

Policymakers provide ALDs to test users to guide the user's interpretation of student performance based upon a test score. Because ALDs "drive" score interpretation the measurement community is moving to the position that ALDs should be developed and used to inform test design and item development. Bejar, Braun, and Tannenbaum (2007) trace the historical and often overlooked recommendation that achievement level descriptors (ALDs) should serve as a basis of item development (Kane, 1994; Mills and Jaeger, 1998; Pellegrino, Jones, and Mitchell, 1999; Hambleton et al., 2000). This practice enhances test designers' abilities to develop a test so that it supports intended test score interpretations (i.e., what students should know and be able to do in relation to the content standards) and intended test score uses.

Egan, Schneider, and Ferrara (in press) propose three stages of ALD development to correspond with the three intertwined uses of ALDs in the field: item development, standard setting, and score reporting. First, states develop target ALDs to specify a state's expectations for students at the threshold of the achievement level. These define the state's policy and content-based expectations (that is, what it means to be Proficient). They are the lower-bound descriptions of an achievement level, and they may guide the cut score recommendation workshop. These descriptors target the skills all Proficient students should have in common. Next, states should develop range ALDs, which are expansions of the target ALDs. Along with the target ALDs, range ALDs should be developed prior to item writing. The range ALDs reflect the knowledge and skills expected of examinees in the Proficient range. Because these descriptors target a range of student skills, not all Proficient students should be able to demonstrate all knowledge and skills in the descriptions. Range ALDs may guide item development. Finally, states should produce final reporting ALDs once final cut scores are adopted. These ALDs represent the reconciliation of the target ALDs with the final cut scores. The target ALDs reflect a state's expectations of minimal student performance within an achievement level while the final reporting ALDs reflect actual student performance based upon the final approved cut scores. The final reporting ALDs define the appropriate, validated inferences stakeholders may make about examinee knowledge and skills based upon the student's test score. States should clearly explicate in their final reporting ALDs whether the target student or the typical student is being described for reporting purposes.

Continue Reading

View .PDF of the complete research paper »