Development and uses of upper-division conceptual assessments

[This paper is part of the Focused Collection on Upper Division Physics Courses.] The use of validated conceptual assessments alongside conventional course exams to measure student learning in introductory courses has become standard practice in many physics departments. These assessments provide a more standard measure of certain learning goals, allowing for comparisons of student learning across instructors, semesters, institutions, and pedagogies. Researchers at the University of Colorado Boulder have developed several similar assessments designed to target the more advanced physics of upper-division classical mechanics, electrostatics, quantum mechanics, and electrodynamics courses. Here, we synthesize the existing research on our upper-division assessments and discuss some of the barriers and challenges associated with their development, validation, and implementation as well as some of the strategies we have used to overcome these barriers.

Figure 1

A schematic of the design process used to develop CU’s upper-division conceptual assessments.

Figure 2

An example question from the CUE along with the associated scoring rubric. This style of complex rubric was used for the CUE and QMAT and graders must undergo specific training to use these rubrics to produce reliable scores.

Figure 3

An example question from the CCMI along with the associated scoring rubric. This style of simple rubric was used for the CCMI and CURrENT and graders need minimal training to produce reliable scores.

Figure 4

Histogram of average course score on the CUE across 7 institutions demonstrating improved performance for courses using CU’s transformed materials (11 courses, 9 CU; 329 students, 312 CU) relative to courses using only traditional lecture (10 courses, 5 CU; 303 students, 191 CU). All transformed courses used some or all of CU’s transformed electrostatics materials.

Figure 5

Histogram of average course score on the CURrENT across 6 institutions demonstrating preliminary indications of improved performance for CU’s transformed ($N=124$) and PER ($N=172$) courses relative to courses using only traditional lecture ($N=50$). Overall averages by course are marked by the dashed lines.