requirements are carefully spelled out in a poster distributed to students and teachers; scoring rubrics are also widely distributed. Formative assessment is a critical part of the program as well. Students engage in evaluation of their own work and that of their peers, then use that feedback to inform next steps in building their portfolios. Thus while the AP Studio Art program is not directly based on cognitive research, it does reflect general cognitive principles, such as setting clear learning goals and providing students with opportunities for formative feedback, including evaluation of their own work.

Portfolios are scored quickly but fairly by trained raters. It is possible to assign reliable holistic scores to portfolios in a short amount of time. Numerous readings go into the scoring of each portfolio, enhancing the fairness of the assessment process (Mislevy, 1996). In this way, technically sound judgments are made, based on information collected through the learning process, that fulfill certification purposes. Thus by using a curriculum-embedded approach, the AP Studio Art program is able to collect rich and varied samples of student work that are tied to students’ instructional experiences over the course of the year, but can also be evaluated in a standardized way for the purposes of summative assessment.

It should be noted that some states attempting to implement large-scale portfolio assessment programs have encountered difficulties (Koretz and Barron, 1998). Therefore, while this is a good example of an alternative approach to on-demand testing, it should be recognized that there are many implementation challenges to be addressed.

We return to Minstrell and Hunt’s facets-based DIAGNOSER (Minstrell, 2000), described in some detail in Chapter 5, to illustrate another way of thinking about assessment of individuals’ summative achievement. The DIAGNOSER, developed for use at the classroom level to assist learning, does not fit the mold of traditional large-scale assessment. Various modules (each of which takes 15 to 20 minutes) cover small amounts of material fairly intensively. However, the DIAGNOSER could be used to certify individual attainment by noting the most advanced module a student had completed at a successful level of understanding in the course of instruction. For instance, the resulting assessment record would distinguish between students who had completed only Newtonian mechanics and those who had completed modules on the more advanced topics of waves or direct-circuit electricity. Because the assessment is part of instruction, there would be less concern about instructional time lost to testing.

Minstrell (2000) also speculates about how a facets approach could be applied to the development of external assessments designed to inform decisions at the program and policy levels. Expectations for learning, currently