Measurement and characteristics of evidence-based instruction in physics

Project No.
PI Name
David Meltzer
Arizona State University

Abstract 1

Measurement and characteristics of evidence-based instruction in physics

Presentation Type
David E. Meltzer Yun Kang Michelle Zandieh Arizona State University


Ongoing work in physics education has established various methods for assessing the effectiveness of physics instruction, primarily utilizing various types of diagnostic tests administered both pre- and post-instruction. In addition, a variety of common characteristics of evidence-based instructional approaches has been identified. In this work, we have examined the evidentiary basis in more detail, revealing various gaps that need to be filled. We have pointed out some of the uncertainties inherent in attempts to disentangle the impact of various components of instructional methods. We have also probed the relationship between pre-instruction measures of student knowledge and student outcomes. This work will inform discipline-based education researchers and help illuminate pathways for future work that ultimately can strengthen the effectiveness of instruction.


We are attempting to better understand current measures of instructional effectiveness in STEM instruction and explore avenues for improvement. We are also attempting to probe the manner in which various elements of evidence-based instruction interact to produce specific outcomes. We are analyzing assessment and performance data to probe for common patterns, and are investigating historical data from the early 1900s to add perspective to contemporary work.


We administer diagnostic tests pre- and post-instruction, examine course performance data (grades, etc.), observe classes, examine course syllabi, and interview course instructors. We examine the historical literature to uncover previous efforts at developing evidence-based STEM instruction and compare historical data and instructional methods to those used today. We look for common characteristics of evidence-based instruction and search for evidence of effectiveness or interaction of specific instructional elements.


We have documented specific relationships between students' pre-instructional performance on diagnostic tests and their course performance. We have uncovered long-overlooked instructional experiments and associated data that provide insight into methods used and outcomes observed today. Based on analysis of measurement issues that are common to various eras and instructional approaches, we have pointed out specific questions that need to be addressed through future investigations. We hope to address some of these issues by developing more deeply-probing diagnostic methods that can reveal underlying pedagogical dynamics more effectively than current methods.

Broader Impacts

Our investigations have focused on physics and mathematics, but our approach and our findings are relevant to all STEM fields. On our own campus, students in introductory physics and mathematics courses have benefited from instructional methods that utilize evidence gathered on student learning both from our own students and from those reported in the literature. We have delivered numerous contributed and invited presentations both in the U.S. and internationally that have disseminated our findings to science educators in various STEM fields.

Unexpected Challenges

Due in part to the unusual nature of the project, we were not able to recruit a graduate student research assistant to work on it. Therefore, we have largely relied on undergraduate students to help with much of the project work.