[Motivation and Goals] [Design and Implementation] [Produced Materials]
[Outcomes and Analysis] [Recommendations] [Presentations & Publications] [References] [Acknowledgements]

Motivation and Goals

Physics 1A is a 3-unit, lower-division Newtonian Mechanics course designed for life-science majors at UCSD. This course, which covers kinematics, forces, energy, momentum and rotation, uses calculus but focuses primarily on algebraic relationships for quantitative analysis. During the Fall term, this course has an average total enrollment of 800-1000 students, usually taught by 2-3 instructors and three half-time graduate TAs in three large lecture sections of 200-350 students each. The course is aligned with a separate 1AL lab course, and an optional problem solving session (typically 1-2 hours) is also generally scheduled.

In this format, student interaction with instructors is extremely constrained, with a student-instructor ratio of roughly 150:1 (if TAs are included); problem sessions are as heavily attended as lecture. The UCSD Physics department has compensated by investing in a Physics Tutorial Center staffed by graduate TAs (two at any time) 25-30 hours per week. However, as this center supports all lower-division courses - in excess of 2500 students in any given quarter - one-on-one help remains difficult to obtain. While interactive learning techniques have been deployed to increase student investment and performance (e.g., peer learning, in-class workbooks, in-class writing, metacognition), the scale of the course inhibits meaningful attention to students' individual needs, limits the development of quantitative analysis and problem-solving skills, and makes it difficult to identify and address persistent misconceptions.

In 2015, inspired by work being done by Thomas Gredig at CSU Long Beach, I experimented with a Cooperative Problem Solving model for Physics 1A with the goal of addressing what I saw were major shortcomings in several key problem solving and social/emotional goals for this course and for this population of students:

This model was based on Heller & Heller (2010), which is designed to engage small teams of students in real-world, complex physics problems through the development of and practice with concrete procedures for solving these problems. This is a so-called "flipped" teaching model, where lectures were provided through videos watched at home, while class time is focused on problem solving skills facilitated by multiple instructors, and team project-based learning. My implementation differed somewhat from the original, as described below.

back to top


In Fall 2015, I was assigned to teach all three lecture sections of Physics 1A, providing a baseline of nearly 1000 students to test the Cooperative Problem Solving model. I retained the largest of these sections (360 students) as a control case ("Lecture Course"), and distributed the remaining students into nine sections of 50-60 students each as the experimental cases ("Workshop Course"). The control group participated in an interactive lecture (e.g., peer instruction using clickers, in-class problems and writing), while the experimental group participated in a modified cooperative problem-solving model. For both courses, the 10-week quarter was divided into five two-week cycles covering Kinematics, Forces, Energy, Momentum and Rotation; and all students took the same biweekly exams and final exam. All students were allowed to participate in 5-6 optional problem solving sessions per week led by the graduate TAs and myself, in addition to the Physics Tutorial Center.

Our implementation of the Cooperative Problem Solving model assigned students to teams of five, with specifically assigned roles within those teams modeled from the categories of problem solving discussed in Heller & Heller (2010) and other sources:

Manager: leads planning on how to solve problem and sets schedule; keeps team members on task; contributes to solution process description section of final write-up and presentation
Researcher: performs background research to determine relevant quantities and identify similar problems in course; performs context research on how problem applies to real world; contributes to background section of final write-up and presentation
Solver: performs quantitative analysis of problem; contributes to solution section of final write-up and presentation
Communicator: leads write-up of problem and solution, incorporating elements from all team members; presents results to class in 3-minute presentation at end of cycle
Skeptic: checks solution and tests against extreme cases; fact-checks context research; suggests alternative approaches; contributes to validation section of final write-up and presentation

Team-based work was facilitated by the structure of the classroom, which was a flexible-furniture room with moveable tables and chairs, and video screens and chalk/white boards on all walls. During class, students would be given a brief topical lecture (generally to review core concepts), then worked with their teams on in-class training sets or team projects. Training set solutions were not collected, but review of solutions would either consist of individual or team presentations with peer discussion and feedback, or an instructor going over the problem, depending on the degree of completion and understanding. Not all problems in the worksheets were completed in-class, and students were expected to complete the remaining problems on their own. Solutions to the training sets were distributed at the end of each cycle as a study resource. The teams were changed every two weeks, using an algorithm that aimed to produce mixed-exam-performance teams (e.g., both high and low performers) based on the most recent biweekly exam. Toward the end of the quarter, when write-up and presentation elements were curtailed, the Researcher and Communicator positions were merged to form teams of four.

Lectures for the experimental groups were delivered as online videos, organized into "concept" and "problem solving" formats (see Lecture Videos below). Students could watch the videos at their own pace, but had to complete 3-4 lecture quizzes during each two-week cycle at regular intervals to provide a reference for where they should be in their viewing.

The learning activities of the Lecture (control) and Workshop (experimental) groups were as follows:

Lecture courseWorkshop Course
  • Lecture participation (20 lectures, 30 hours total)
  • Eighteen (18) pre-lecture reading quizzes
  • Eight (8) weekly on-line homework assignments
  • Five (5) bi-weekly book problem sets (optional: not graded)
  • Nine (9) weekly metacognition write-ups online
  • Four (4) bi-weekly cycle exams
  • One (1) final course exam
  • Watch online lectures (70 lectures, 25 hours total)
  • Class participation, including training set work (20 classes, 30 hours total)
  • Eighteen (18) pre-lecture reading quizzes
  • Eight (8) weekly on-line homework assignments (reduced)
  • Five (5) bi-weekly book problem sets (optional: not graded)
  • Nine (9) weekly metacognition write-ups online (optional: bonus)
  • Four (4) bi-weekly cycle exams
  • One (1) final course exam

The graded elements of the course also differed slightly in their weighting between the student groups, in order to keep exam grades consistent:

Lecture courseWorkshop Course
  • Video lecture quizzes (10%)
  • On-line homework via WebAssign (10%)
  • Class attendance (10%)
  • Weekly Metacognition (5% bonus)
  • Team Projects (15%)
  • Bi-weekly exams (30%)
  • Final exam (25%)
  • Pre-lecture online quiz (10%)
  • On-line homework via WebAssign (20%)
  • In-class participation (10%)
  • Weekly Metacognition (5%)
  • Bi-weekly exams (30%)
  • Final exam (25%)

In addition to these graded elements, students took the Force Concept Inventory(Hestenes et al. 1992) and Colorado Learning Attitudes about Science Survey (CLASS; Adams et al. 2006) both pre- and post-course to assess changes in basic Physics concept knowledge and attitude evolution. These surveys were given online and students were given extra credit for completing them by their respective deadlines.

back to top


Lecture Videos

Example frame from Learning Glass format (problem solving video).

Example frame from Lecture Slide format (concept video) with annotation.

70 Lecture videos totaling 25 hours were produced in the summer preceding the course in collaboration with UCSD's Educational Technology Services (ETS). The video production took advantage of ETS's Learning Glass technology, allowing viewers to see both board writing and the instructor in the same frame. Both Learning Glass and lecture slide formats were used to cover problem solving and concept skills, respectively.

The videos used in the course are currently being formatted for general distribution; example videos are linked below.

Procedures for Solving

Students were provided and guided through step-by-step procedures for approaching various classes of mechanics problems. These procedure guides were aimed to be comprehensive, so individual parts were broken down in the worksheet exercises in class, and students were encouraged to follow the procedures as they solved homework problems on their own.

In-class Training Sets

To practice problem solving skills and build up to more complex problems, students worked as a team on problem-solving worksheets, or "training sets", which aimed to be scaffolds for their learning. The first three worksheets are based on worksheets developed by Michael Anderson, which probed both problem solving and concept learning, but we branched off to primarily problem solving in subsequent worksheets.

Solutions for these will provided on an individual request basis; if you find errors in these worksheets, please do not hesitate to contact me.

Team Projects

To align content learning with student's research and career interests, we developed a series of biweekly projects based on the material of a given two-week cycle, but which must be solved cooperatively as a team. By design, these projects were aimed to be more complex and open-ended than standard problems, required background research, and required analysis of data or pseudo-data. Initially these projects were presented to the rest of the class, but because excessive effort was being put in my students in the class overall, the last two projects were graded based on a solution writeup only.

Solutions for these will provided on an individual request basis; if you find errors in these worksheets, please do not hesitate to contact me.

Project 1: Trapping Fruit Flies with Physics (biomechanics)

Teams are given 2D time sequence measurements (synthetic data) of fruit fly velocities from which they must construct a path and acceleration vectors, to determine if a given chemical attracts or repels fruit flies. Tests construction and interpretation of position/velocity/acceleration diagrams, using kinematic relations for numerical integration/differentiation, and interpretation of experimental data. (Rubrics: [writeup] [presentation])

Project 2: Physics and Lawsuits (physical science in law)

Teams must assess the true circumstances of a collision between two cars in the vicinity of a blind curve, and assess whether the driver or the city (or both) are at fault. Each team has different parameters for the dimensions, speed limit and road conditions, and must formulate a logical argument as to the probability that the driver may be lying about his or her testimony. Tests force diagrams, accelerated linear and circular motion, friction force, logical arguments.

(Rubrics: [writeup] [presentation])

Project 3: Powering a Rural Hospital (physics and medical policymaking)

Teams must determine whether a hospital planned for construction in a rural area, and expected to be powered on-site by a hydroelectric plant, is feasibly designed. Each team has different parameters for the waterfall (flow rate, height) and hospital size, and must determine if energy consumption is match to energy needs. Tests work-energy theorem, dissipative energy loss, energy transformation, sustainable energy sources, estimation, and quantitative reasoning.

(Rubrics: [writeup])

Project 4: Forensic Investigation of JFK's Assassination (forensic science)

Teams review the Zapruder film of JFK's 1963 assassination, and read about the various theories as to why JFK's head snaps backwards in the final shot (e.g., second shooter theory, Alvarez jet model). Applying two models (ballistic pendulum and ejecta jet) with various model parameters, and 1D momentum conservation, must determine if a second shooter is necessary to explain JFK's head motion. Tests application of momentum conservation with and without energy loss, estimation, use of models, quantitative reasoning and logical argumentation, skepticism and weighing evidence.

(Rubrics: [writeup])

Associated Software

Much of the organizational work for this project required custom code, which is linked below. Note that these scripts are given as-is, and may require dependencies not included here.

back to top

Student Outcomes

The results are preliminary; we are still analyzing student outcomes

Video Viewing

We analyzed the video viewing habits of the students to understand when and why they watched the lecture videos. The data available track which video files were opened by which students, but does not provide information on total viewing time or video completion.

Viewing statistics followed expected patterns based on students' study habits. Specifically, video views peaked at the start of each cycle and just before an exam, with overall viewership generally declining with time over the quarter. On the day-to-week scale, viewing peaked in the hours of 3pm-8pm Sunday through Friday, with little or no views between 1am and 9am and all day Saturday.

Number of video views per day over the course of the quarter, with major events highlighted

Number of video views per day of week (vertical) and per hour of day (horizontal). Note that bi-weekly exams were scheduled 7-9pm on Fridays.

Very few of the students watched all of the videos. Fewer than half of the students watched more than half of the videos, with an average of XXX unique videos viewed per student (only three students watched all three videos). Total video views - including repeat views - averaged around XXX views, implying an average video repeat rate of XXXX. Only one person (perhaps) watching an incredible 338 videos over the course of the quarter.

Distribution of unique video views among students in the Physics 1A course. A total of 70 videos were available for viewing.

Distribution of total video views among students in the Physics 1A course. One outlier at 338 views is not shown in this distribution.

The figure below shows the total number of video views per video, separated into concept (yellow) and problem-solving (green) videos. In general, concept videos were viewed more often than problem solving videos. Some of the latter were labeled "optional", and consequently had single-digit views. There is a dramatic change in viewing behavior at the end of the second cycle, which coincides with a reduction in the work effort required for the bi-weekly team projects. We also see a steady decline in video viewership, particularly in the last cycle just before the final exam.

Distribution of all views by video, distinguished by concept (yellow) and problem-solving (green) formats.

Exam Performance


Force Concepts Inventory Performance


Evolution of Physics Attitudes


Analysis of Metacognition Responses


Student Feedback


back to top



back to top

Presentations and Publications

Here is a list of presentations and publications stemming from this project.




back to top