Analyzing interviews on computational thinking for introductory physics students: Toward a generalized assessment

Computational thinking in physics has many different forms, definitions, and implementations depending on the level of physics or the institution it is presented in. To better integrate computational thinking in introductory physics, we need to understand what physicists find important about computational thinking in introductory physics. We present a qualitative analysis of 26 interviews asking academic ($N\_a=18$) and industrial ($N\_i=8$) physicists about the teaching and learning of computational thinking in introductory physics courses. These interviews are part of a long-term project toward developing an assessment protocol for computational thinking in introductory physics. We find that academic and industrial physicists value students’ ability to read code and that python (or vpython) and spreadsheets were the preferred computational language or environment used. Additionally, the interviewees mentioned that identifying the core physics concepts within a program, explaining code to others, and good program hygiene (i.e., commenting and using meaningful variable names) are important skills for introductory students to acquire. We also find that while a handful of interviewees note that the experience and skills gained from computation are quite useful for student’s future careers, they also describe multiple limiting factors of teaching computation in introductory physics, such as curricular overhaul, not having “space” for computation’, and student rejection. The interviews show that while adding computational thinking to physics students’ repertoire is important, the importance really comes from using computational thinking to learn and understand physics better. This informs us that the assessment we develop should only include the basics of computational thinking needed to assess introductory physics knowledge.

Figure 1

General outline of our results, in the order we present them. Of all the 35 interview questions, 5 are highlighted in this paper (represented in blue boxes). For each of the five questions, themes were extracted based on the participants’ responses (represented in green boxes). The orange box represents individual response ${\mathrm{codes}}^{\mathrm{q}}$ that fit into each theme. Responses could fall under multiple themes within a specific question.