Computational Skills

Computational skills encompass computational physics skills (e.g., translating models into code, choosing scales and units, choosing appropriate algorithms and tools, extracting physical insight, understanding the limitations of computers and computer models), the use of a variety of computational tools (e.g., spreadsheets, structured programming languages, computer-based symbolic manipulations, modeling packages), and technical computing skills (e.g., analysis, visualization, and presentation of data). Computational skills can be introduced at numerous points throughout the curriculum through individual and group activities, and can become more sophisticated as students progress through their education.

Benefits

Computational tools and techniques are used ubiquitously in physics, are integral to how physics is currently practiced, and provide excellent preparation for careers. Knowledge and skills in programming, simulations, and modeling are needed by physics graduates in a variety of careers. Adding these skills to the curriculum addresses a common weakness that many physics graduates report in their undergraduate programs, improves and accelerates students’ ability to engage in research and solve research-like problems, and may assist in the recruiting of students. Computational skills allow students to answer questions not solvable through analytic techniques, including practical and applied problems, and are transferable across disciplines. The use of computational tools can deepen students’ understanding of fundamental concepts and principles.

Effective Practices

Thematic grouping (1, 2, 3, ...)

Actionable practice (A, B, C, ...)

Implementation strategy (i, ii, iii, ...)

1. Inventory software and hardware available on campus.
  Determine students’ access to technology (e.g., computers, phones, and software) and consider this access in plans involving computational activities given as homework (e.g., ensure that students have access to computing labs outside of class time, work with your instructional technology office to make site-licensed software available off campus, ensure that assignments use software that is accessible on the devices students have, and/or loan students laptops with software installed).
  Locate other computational resources on and off campus, e.g., coding or robotics clubs, local industry, or technology companies.
  Survey faculty on the software and computational skills they integrate into their courses and their expertise and interest in expanding computational skills into the physics curriculum.
  Inventory the expertise within your institution to identify potential collaborators who can either provide guidance on teaching computational skills or identify existing activities outside the department that could be integrated into and/or modified for use within the physics curriculum.
2. Survey the computational skill sets of incoming students, including computational physics skills, use of computational tools, and technical computing skills.
  Collaboratively determine the desired computational skill set of program graduates, using input from sources such as departmental alumni, potential employers of program graduates, etc.
  Collaboratively determine the desired computational skill set of students in non-majors physics courses, using input from sources such as student feedback, discussions with faculty in other departments, etc.
  Collaboratively determine program-level student learning outcomes for computational skills, including computational physics skills, use of computational tools, and technical computing skills, throughout the curriculum. Revise these outcomes as needed.
  Determine what can be done with existing resources.
  Determine what can be done with additional resources and investment of time.
  Reflect on and revise student learning outcomes as necessary, based on assessment feedback.
3. Pay special attention to ensuring that all students feel supported in participating in computational work, in light of pervasive stereotypes of computation as a white and male activity.
  Encourage faculty to engage students in the design of instructional materials, assessments, and other projects, and to ask students what they need to be able to learn and demonstrate their understanding. For example, students may be able to provide input on grading and assessment structures to better support their learning.
  Train your faculty and instructional staff to notice and respond to the marginalization of students during group work, recognizing that the nature of this marginalization may be different during computational activities than during other kinds of activities. For example, students with more experience in computer science may dominate computational group work, so instructional staff should be aware of students’ backgrounds and strategies and include all students.
  Use the results of incoming student surveys to intentionally design courses and experiences that are inclusive and promote equity within the program.
  Provide computational learning opportunities for and celebrate achievements of all students, particularly those from historically marginalized groups.
  Invite women and people of color whose work includes computational physics to share their experiences with students in classes or departmental seminars.
  Discuss historical contributions to computation by women and people of color, e.g., Katherine Johnson, Dorothy Vaughan, Grace Hopper, and Ada Lovelace.
  See the section on for other strategies for supporting students from marginalized groups.
4. Develop and implement a strategy to assess which faculty practices are most effective for supporting students in achieving your program-level student learning outcomes.
  Develop a curriculum map that identifies where each student learning outcome for computational skills is introduced and developed in the curriculum.
  Have regular meetings of faculty teaching computational skills to check whether all learning outcomes in your curriculum map are being addressed in the appropriate courses and how these outcomes are connected throughout the curriculum.
  Identify and resolve any issues with implementing the curriculum map, e.g., gaps, lack of faculty buy-in, or a need for faculty development.
  Communicate the outcomes of these activities at department meetings, identifying accomplishments as well as gaps for potential curricular revisions.
5. Lower barriers to using computation in courses by finding reliable technological solutions, sharing resources, and encouraging faculty collaboration, e.g., by pairing faculty members who have less experience with computation with another faculty member who can support their growth in this area.
  Encourage faculty who are teaching computational skills to use the many available resources for computational instruction. See Resources below for examples.
  Support faculty to take advantage of professional development opportunities around evidence-based teaching practices in computational instruction.
  Support faculty to present about their experiences both locally and nationally.
  When developing or reforming courses and curricula, use your curriculum map to include computational skills development in your course activities.
  Recognize and encourage faculty who implement computation in their courses, document the impacts on student learning and student attitudes, and use this information to improve instruction.
1. Implement active-learning approaches, particularly those suited for computational skills, such as pair programming or other group or project-based approaches. See the section on for details.
  Integrate computational activities throughout the standard physics course offerings in a deliberate and coherent way.
  Design laboratory activities that use computational approaches for data acquisition and organization, analysis and modeling, and visualization. See the section on for details.
  Scaffold students’ development of career-relevant computational skills throughout the curriculum.
2. Teach students about analytical models for which there are computational analogs and extensions, e.g., projectiles with and without air drag.
  Have students reflect on the similarities and differences between using analytics versus computation to model a physical system, e.g., obtaining an expression whose parameters can be tuned versus having to choose fixed values for parameters, or solving an ordinary differential equation numerically using a “step” algorithm versus analytically by guessing a functional form and showing that it is a solution.
  Have students explore the affordances and constraints of analytical and numerical modeling, e.g., the computational time necessary for accurate numerical results or the lack of analytic solutions in some limits or regions of parameter space for models such as the simple pendulum.
  Have students explore the connection between posing problems and solutions in terms of continuous (analytic) variables and posing them in terms of discrete (computational) variables, e.g., discrete versus continuous location and time information for a projectile.
  Have students reflect on the advantages and disadvantages of analytical versus computational analyses based on the activities above.
3. Support students in learning to plan and document their code.
  Support students in analyzing their programs for efficiency and accuracy, and iteratively improving them.
  Support students in learning the theoretical basis of canonical algorithms and data analysis techniques used in physics.
  Familiarize students with commercial computational packages.
  Familiarize students with artificial intelligence and machine learning methods.
  Familiarize students with ethical issues in computing, with respect to, e.g., model choice, algorithm choice, uncertainty quantification, artificial intelligence, energy usage, and the goals of projects that computational work can be used to support.
4. Establish working relationships with appropriate related departments (e.g., computer science, data science, and engineering) to design or co-design computational courses and degree tracks and to provide physics credit for those courses.
  Discuss and collaboratively align courses, languages, and resources with other departments to avoid duplication of efforts, extension of graduation timelines, or excess costs to students.
  Consider cross-listing computational physics courses with other departments.
  Encourage students to take appropriate computer science courses if no computational physics courses exist.
1. Develop program-level student learning outcomes for a core set of computational concepts and teach them within a physics context.
  Incorporate activities that integrate foundational techniques such as iterative numerical integration methods and data visualization, fitting, and analysis.
  Design exercises using video analysis, data acquisition interfaces, and computer simulations, e.g., video analysis of the center-of-mass motion of an object when tossed between students, acquiring and analyzing position, velocity, and acceleration data for a mass-spring system, or quantification of damping of the same system using a PhET simulation.
  Include activities that require students to create, explore, and refine models of physical systems through programming and simulations, e.g., exploring complex multi-body systems using Newton’s Laws or analyzing the motion of a projectile without and with air drag.
  Include activities in which students compare experimental data to the behavior of a simulation/computational model, e.g., comparing experimental data to the analytical model of a simple pendulum or harmonic oscillator.
  Include activities in which students explore how varying initial conditions may produce interesting, surprising, or complex behavior, e.g., periodic versus chaotic motion of a double pendulum.
  Seek recommendations for computational examples and activities from other departments these courses serve.
3. Choose a programming language or computational tool that aligns with the resources and computational experience of your faculty. Consider the experience of faculty, learning resources, other computational courses, relevance of computational courses to physics, and cost of computational tools for the classroom.
  Explore more sophisticated problems from a variety of physics subfields.
  Integrate advanced computational techniques such as machine learning or big-data analysis.
4. Identify problems and/or projects in advanced courses that are appropriate for computer modeling and computational solutions.
  Identify advanced lab experiments (on, e.g., chaotic and stochastic systems) that require computer modeling to predict and understand the experimental system’s behavior.
  Continue and, when appropriate, enhance the use of data acquisition, visualization, and analysis in advanced laboratory activities.
5. See the sections on Undergraduate Research and for details.
  Identify research opportunities within your department, outside your department, and/or off campus through which your students can get involved in computer modeling and analysis.
  Identify research within and outside your department that would benefit from computer modeling and analysis, and find ways to involve undergraduate students in doing computational work to support that research, e.g., through independent study projects.
  Integrate advanced computation and visualization techniques into theoretical and experimental research experiences.
  Encourage computational independent studies if research experiences are lacking.
1. Communicate the value of computational skills in a variety of careers. See the section on Career Preparation for details.
  Promote the computational skills used by students and alumni when talking with current and prospective students.
  Encourage presenters at department seminars to highlight use of computational skills, e.g., how they choose which type of package (analytical, coding, or computational modeling) to use to solve a problem.
  Invite alumni to speak with students and faculty (in, e.g., classes, colloquia, and small group meetings) regarding their use of computational skills.
  Encourage faculty and students to highlight their computational skills in research talks, posters, and departmental promotional materials.
  Encourage faculty to highlight information from AIP (see Resources below) and other professional societies in science and engineering on the use and importance of computation in careers.
2. Encourage students to discuss computational skills with employers at internship and career fairs.
  Invite speakers from non-academic careers who use computation to meet with students and discuss their experiences.
  Develop a list of local and regional employers looking for students with computational skills; include how employees at those workplaces use such skills.
  Connect with curricular partners (e.g., engineering programs and computer science programs) to establish contacts between students and potential employers. Contact these programs’ internship site liaisons, members of their advisory boards, and/or their representatives at internal job fairs.

Programmatic Assessments

Identify which of the above Effective Practices you are implementing and track the results of implementing them.
Conduct an inventory of where computational skills are introduced and developed in your curriculum and through extra-curricular mechanisms (e.g., internships and clubs), and how well these mechanisms address your program-level student learning outcomes for computational skills. Determine where there are implicit or explicit expectations about using and/or developing computational skills.
Survey alumni and current and potential employers of your graduates about the computational skills they need, and compare these needs to your program-level student learning outcomes for computational skills.
Evaluate whether students have access to adequate technology for computational work.
Collect formal feedback on students’ knowledge of available curricular and extra-curricular experiences to further computational skills, attitudes towards computational skills, experiences with computational skill development in your courses, perceptions of departmental support for computational skill development, and perceptions of support from collaborating departments. Use, e.g., course feedback forms, exit surveys of graduating seniors, and alumni surveys.
Analyze the demographics of your majors, likely in partnership with your office of institutional research, along with majors’ backgrounds and experiences in computation and access to technology. Correlate this analysis with survey responses to determine how well your program is providing support for computational skills development for students in various demographic categories.
Gather information via surveys or discussions about faculty perceptions of the importance of computational skills for undergraduates.
Review the expertise of your faculty with respect to computational skills and identify gaps that could be addressed through on- or off-campus professional development.
Use exams, homework, and projects to assess students’ achievement of program-level student learning outcomes for computational skills.
Track the number of physics graduates who obtain jobs that require computational skills.
Collect contact information for all graduates before they leave, and maintain a database of alumni contact information and positions. Survey alumni about their perceptions of the computational skills they developed while in the program, how the program prepared them for their careers, their use of computational skills in their careers, and what recommendations they have based on their experiences.
Survey employers of your alumni about how well prepared your graduates are in terms of computational skills.

Evidence

There is an extensive peer-reviewed literature in computational physics education, much of which is highlighted in a special issue of the American Journal of Physics (Reference 1). Reference 2 documents current practices in computational physics, and Reference 3 and references therein provide evidence for the need for computational skills development in physics programs.

W. Christian and B. Ambrose (editors), Theme Double Issue on Incorporating Computation into the Physics Curriculum, American Journal of Physics, 76(4&5) (2008).
M. D. Caballero and L. Merner, “Prevalence and nature of computational instruction in undergraduate physics programs across the United States,” Physical Review Physics Education Research, 14(2), 020129 (2018).
P. Heron, L. McNeil, et al. (editors) “Phys21: Preparing Physics Students for 21st-Century Careers,” American Physical Society (2016).

Resources

AAPT Undergraduate Curriculum Task Force, “AAPT Recommendations for Computational Physics in the Undergraduate Physics Curriculum,” American Association of Physics Teachers (2016): a report providing recommendations for how to include computation in the physics curriculum
“Advancing Interdisciplinary Integration of Computational Thinking in Science, Conference Report,” American Association of Physics Teachers (2020): a report that provides recommendations on integrating computational thinking in science courses, including supporting pedagogical computational knowledge for instructors
Partnership for Integration of Computation into Undergraduate Physics (PICUP): a community of instructors sharing resources including peer-reviewed instructional materials, faculty development workshops, and support for instructors working to include computation in their courses and programs at all levels
AIP Statistical Research Center: a group that regularly collects and analyzes data on education, careers, and diversity in physics, astronomy and other physical sciences. Several reports have findings support the integration of computational skills in the curriculum, e.g., Field of Employment for New Physics Bachelors and Skills Used Regularly: New Physics Bachelors Employed in STEM Fields.

Computational Skills

How to Use This Guide

Benefits

Effective Practices

Effective Practices

Establish goals and a plan for providing students with computational skills

Determine existing resources and expertise of faculty

Establish program-level student learning outcomes for your students’ computational skills

Include strategies for supporting students from marginalized groups, while not singling out such students

Create an implementation plan and a corresponding assessment strategy

Support faculty development around computational skills and generation and adaptation of instructional materials

Integrate opportunities to develop computational skills into the curriculum

Encourage the department to embed computational learning opportunities across the curriculum

Encourage faculty to provide opportunities for students to learn about the relationship between analytical and numerical models

Promote opportunities for students to learn the broader computational practices necessary for doing computational physics

Develop and/or collaborate with other departments to offer computational curricular opportunities

Provide students early and continuing opportunities to learn and apply computational skills

Develop technical computing skills and introduce computational physics skills in introductory courses for majors and non-majors

Introduce the use of computational tools and develop computational physics skills in second-year courses

Introduce the use of computational tools and develop computational physics skills through a stand-alone course

Integrate a variety of computational skills into advanced courses and labs

Develop students’ computational skills through research or similar experiences

Communicate the value of computation in physics and for a broad range of careers

Promote the applicability of computational skills

Promote connections with employers for jobs that use computational skills

Programmatic Assessments

Programmatic Assessments

What are the mechanisms through which your degree program includes computational skills? What are the additional mechanisms through which it could do so?

Are all of your students, including those from marginalized groups, participating in, aware of, and valuing computational learning opportunities?

What expertise and perceptions do your faculty have with respect to computational skills? What expertise do they need?

How successful is your program at preparing students to use computational skills? How has its success changed over time?

Evidence

Resources

Related Sections