Survey of physics reasoning on uncertainty concepts in experiments: An assessment of measurement uncertainty for introductory physics labs

[This paper is part of the Focused Collection on Instructional labs: Improving traditions and new directions.] Measurement uncertainty is a critical feature of experimental research in the physical sciences, and the concepts and practices surrounding measurement uncertainty are important components of physics lab courses. However, there has not been a broadly applicable, research-based assessment tool that allows physics instructors to easily measure students’ knowledge of measurement uncertainty concepts and practices. To address this need, we employed evidence-centered design to create the Survey of Physics Reasoning on Uncertainty Concepts in Experiments (SPRUCE). SPRUCE is a pre-post assessment instrument intended for use in introductory (first and second year) physics lab courses to help instructors and researchers identify student strengths and challenges with measurement uncertainty. In this paper, we discuss the development of SPRUCE’s assessment items guided by evidence-centered design, focusing on how instructors’ and researchers’ assessment priorities were incorporated into the assessment items and how students’ reasoning from pilot testing informed decisions around item answer options. We also present an example of some of the feedback an instructor would receive after implementing SPRUCE in a pre-post fashion, along with a brief discussion of how that feedback could be interpreted and acted upon.

This article appears in the following collection:

Figure 1

Item 3.3 (with modified numbers) asks students a single MC question about reporting the value of a single period of oscillation of a mass on a spring based on a measurement of 20 oscillations. The scoring of responses to this item is depicted in Table 3.

Figure 2

Item 4.1 went through iterations informed by multiple rounds of pilot testing with students.

Figure 3

Histogram showing students’ reported uncertainty values for item 4.4 on beta 1. These responses show peaks at the correct value and at our planned distractors.

Figure 4

Item 1.1 asks students what they would do after taking a single measurement for time, then asks students to support that choice.

Figure 5

A portion of an instructor report showing preinstruction and postinstruction scores for the course and all historic data. Statistically significant shifts (as determined by a Mann-Whitney $U$ test) for each AO are shown with solid circles and nonsignificant shifts are shown with open circles. Effect sizes for the statistically significant items are calculated using Cohen’s $d$ and shown on the right side of the chart.