This report seeks to make large-scale assessments of student achievement more transparent and provide guidance on item development and quality. The report provides an in-depth look at what is assessed in a selection of current large-scale assessment items to illustrate what kinds of student learning can be assessed using different item formats (e.g., multiple choice vs. constructed-response items or performance tasks). Through a close examination of sample assessment items, the study shows that some assessment items do a better job than others at tapping into the key disciplinary knowledge, understandings, and skills that students are expected to learn in their academic studies.
The study concludes that a balance of well-designed assessment items of varying formats provides greater potential for valid measurement of student learning (i.e., the test measures what it intends to measure) than a test that consists of only one item format. In addition, the study finds that assessments with a variety of items formats are capable of tapping into a broader range of key disciplinary knowledge, understandings, and skills that are central to the disciplines.
