ERIC Number: ED415238
Record Type: Non-Journal
Publication Date: 1997
Reference Count: N/A
Measuring Achievement: What Does It Mean To Test for Robust Understanding? William H. Angoff Memorial Lecture Series.
Shepard, Lorrie A.
In testing, educators have two competing intentions. They want to be fair in allowing students to demonstrate what they know without creating testing conventions and formats that let students pretend to know. Teaching-the-test literature shows that test scores can be inflated, meaning that they can go up without a generalized increase in knowledge. Students can appear to know what they really do not know, as can be illustrated by comparing results from the National Assessment of Educational Progress with those from more familiar standardized tests. Performance assessments are intended to overcome this problem. They are also intended to overcome the format effects of multiple-choice tests that have distorted instruction and allowed students to pretend to know more than they actually do. Even with performance assessments, students may rely on familiar, rote routines, and so pretend to know more than they really do. As a videotape presented as part of the discussion illustrates, asking in different ways is the way to assure that students really know what they are doing, and that their understandings generalize across contexts. Those who are concerned with test bias explore the opposite side of the coin--that students really know, but are not able to show their knowledge and abilities because of some aspect of the test. These two perspectives can be reconciled by careful and thoughtful assessment that approaches student knowledge in different contexts. (Contains 16 figures and 12 references.) (SLD)
Publication Type: Opinion Papers; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: Educational Testing Service, Princeton, NJ. Policy Information Center.