The room was okay. Some students found the posters interesting, but the room was nothing special.

Over the past few years, my colleague and I have transformed our current room to a comfortable, student-personalized space. There are very few posters but lots of student-generated projects and art work.

Some of my favorite pieces are a former student’s AP Studio Art portfolio which was focused on physics.

The two pieces on the left are also part of her portfolio. We also highlight some bridges and towers that students have built. Sometimes inspiration strikes and a water bottle gets glued to the wall.

We also frame and hang photos from students who have entered the AAPT High School Photo Contest.

We display photos from the clubs that we mentor, like Physics Club and FIRST Robotics, to inspire students to participate.

The whole ceiling is covered with mobile projects (forces and torques in equilibrium). Almost every student leaves their project until next year’s class makes their own.

My least favorite part of the room is now the individual desks. The plan is to get hand-me-down tables and chairs from an adjoining physics room and be able to arrange students in groups of four next year!

I tried to explain what computational modeling is and how it is more than just programming. I then encouraged teachers to use computational modeling in their classroom and shared why I think it improves student learning.

We used VPython and the physutils package.

We started with John Burk’s 1-dMotionSimulation.py example. I then asked each teacher to modify the model in some way and observe the results.

I then presented [several starting examples](https://github.com/gcschmit/vpython-physics ) that I created for my AP Physics B class and shared how students built upon these examples to solve everything from homework problems to their projectile motion lab practicum.

I left lots of time for teachers to explore these starting examples and help each other and get help from me. I saw teachers unfamiliar with Python create some pretty cool models in very little time.

Here are the slides I used to introduce computational modeling:

Computational Modeling with VPython by

Here are the links to the resources that I displayed at the end of the session:

• Georgia Tech PER Group
• my GitHub
• John Burk’s blog
• Integrating Numerical Computation into the Modeling Instruction Curriculum by Caballero, Burk, et al.

I started with the basics: specifying names, short name, and units for the dependent and independent variables; titling the graph; setting the graph options. I showed how linear the fit uses the specified variables and units and how to specify measurement uncertainty and see its affect with error bars.

We then focused on using calculated columns to perform linearization manually and then using LoggerPro’s curve fit feature.

Near the end of the session, I demonstrated some of the more advanced graphing features of LoggerPro with examples from this year:

• multiple data sets on the same axis
• multiple y axes
• examine and tangent tools
• grouped graphs (position vs. time and velocity vs. time)
• histograms

Here is the tutorial handout I provided:

Logger Pro Graphing Tutorial by

I absolutely love Canvas’ ease of use. I use all the typical features like announcements, discussions, and file storage. More unique features like modules help my students find everything they need for each unit and pages allow me to easily share enrichment materials.

However, looking back at this first full semester with Canvas, I was surprised which feature had the greatest impact on student learning. It wasn’t any of the above. It was SpeedGrader. Specifically, the ease with which SpeedGrader enables me to provide rich feedback to students on their assignments. Sure, I provided feedback before Canvas by writing comments on lab reports, but it was time consuming (I write much slower than I type) and not always legible (my handwriting is poor). I always had more feedback to provide than what I took the time to write. SpeedGrader has changed all of this.

Here’s my workflow for AP Physics B. Students create an ePortfolio in Canvas that contains all of the labs for which they perform analysis and are assessed. I create an assignment in Canvas for each lab and they submit a link to their ePortfolio. (The ePortfolio part isn’t critical, you could create assignments and have students submit their work in any number of ways.) In SpeedGrader, I can view their ePortfolio in one pane while typing feedback in another. This feedback is what has had the greatest impact on student learning.

I don’t score labs by subtracting a bunch of points, I read them. For my AP class, they earn a score of 1-5 which is reported in the online grade book, but doesn’t show up anywhere in Canvas. In Canvas, I just mark the assignment as complete or incomplete. In Canvas, the focus is on learning; not grades. What students do get is my feedback which often starts a discussion about their lab. My feedback is usually questions of the type I would ask of them in person. Questions that help them make connections between different ideas, clarify a misunderstanding, or illustrate an inconsistency in their analysis. In addition, I can easily point out sections that are incomplete. Many students have their notifications configured so that they receive an email when I submit feedback and some respond back almost immediately.

The integrated discussions in SpeedGrader is a perfect example of the role that technology should play in education. Enhancing a sound educational practice (rich feedback and discussion) by making it more efficient and easier for all involved.

Four of my five classes submit all of their assignments in Canvas. Guess what that fifth class will start doing this semester?

I had remembered reading about John Burk’s capstones and modeled mine after his. While I have done similar end-of-semester projects before, these were by far the most successful. Reflecting on the capstones, I think a couple of factors helped my students be so successful.

The first is that I provided several exemplars as sources of inspiration. I stressed that I didn’t want a traditional lab report or a PowerPoint presentation. I wanted something that other students would want to read or watch. I pointed students to the following YouTube channels and blogs:

• Veritasium
• Minute Physics
• Dot Physics
• xkcd’s What If?

Based on our current unit, I frequently share links to these folks; so, most students were familiar with their style. Since I wanted students to spend the final exam period engaged in each other’s capstones and I would be spending my winter break scoring these capstones, I wanted them to be as interesting as possible.

The second factor that contributed to my students’ success is the requirement that students submit an initial idea for their capstone and then are provided considerable time in class for work and feedback from me. I spent an entire two days moving from student to student and sharing feedback on their capstone. Students had two additional days to work in class, solicit additional feedback, and use lab equipment. This wasn’t enough feedback for every student. A few really needed another round of review and feedback as their capstones missed the mark. While they could have solicited this additional input on their own, their final capstone would have been better if I could have shared another round of feedback.

The third factor that contributed to success is that I provided access to a wide range of technology, but I didn’t mandate that students use any particular technology. This is how educational technology should be integrated into the classroom, it is a tool to enhance learning where appropriate and not a means to an end unto itself. Students chose to use a high-speed video camera, Tracker, Logger Pro, iPads, iPad document camera stand, Educreations app, Scribd, video screen capture, VPython, LabPro and sensors, Vimeo video hosting for embedding, and Canvas ePortfolios.

There were so many fantastic capstones. I selected several to share that represent the variety of engaging presentations:

• Energy Efficiency of a Piano Minute Physics-style video
• D-Rose and Gravity An xkcd What If-style capstone on how the position of celestial bodies affects Derrick Rose’s vertical
• Bubbles Vogue-style magazine about Secrets of the Soap Bubble: How to be Beautiful
• Yo-Yos video analysis of a yo-yo to calculate its moment of inertia
• The Physics of Space Travel and Orbital Manuevers analysis of Kerbal Space Program simulator
• No Stress about Stresses Minute Physics-inspired video
• Newton’s Law of Cooling Experiment Well done and polished experiment to explore Newton’s Law of Cooling for different liquids
• LCAC Hovercraft Analysis Minute Physics-inspired video
• Lava Lamps in the style of a science magazine

Thanks to John Burk for planting the capstone seed in my mind and Derek, Henry, Rhett, and Randall for providing amazing and engaging exemplars for my students to model!

In case you’re interested, here is the rubric. It isn’t perfect, but it worked okay:

Capstone Rubric

4,329 films were submitted to the 2012 Cannes Film Festival. This blog had 13,000 views in 2012. If each view were a film, this blog would power 3 Film Festivals In 2012, there were 22 new posts, growing the total archive of this blog to 73 posts. The busiest day of the year was January 5th with 160 views. The most popular post that day was Holometer.

The full report is available.

Differentiation

I have a very diverse collection of students in my AP Computer Science classes. I have students struggling in algebra 2 sitting next to students who are taking multi-variable calculus. I have students who debug race conditions on our FIRST Robotics programming team and students who are still struggling to distinguish types from variables. I welcome this diversity. I don’t think there should be a math prerequisite. Managing this diversity wasn’t nearly as challenging as I expected. In fact, I find it easier to differentiate in my computer science classes than I do in my physics classes. I do this is a couple of ways.

First, we spend most of our time in class programming. If a student isn’t working on their own program, they are most likely helping someone else with their program. Most of the programming activities that we do are learning activities and, therefore, not graded. I welcome and encourage their collaborative efforts. That’s how software is created in the real world. As a result, the students who need a lot of extra assistance get significant one-on-one time with me and students who need a little assistance get help from other students. This works quite well.

Second, I (and sometimes the students) create challenges for each programming lab. There is a certain base level of functionality everyone will achieve, but that will take some students two days and others twenty minutes. While the students that finish in twenty minutes can help others, I want to provide them a challenge that makes them stretch. I’ll illustrate with a couple of examples.

One activity was to write a palindrome tester. The challenge was to only examine letters and ignore punctuation. Students who met the challenge found some really long palindromes to test.

Another assignment was to use arrays to implement a stack of integers and then use that stack to implement a base-ten integer to binary converter. I defined a couple of challenges: implement a peek method and optimize memory usage by shrinking the array when it is way too big. However, the most popular challenge was found by a couple of students; they extended the program to support converting decimals from base-ten to base-two!

Choice

When practicing looping structures, I provided three different programming activities. One was mathematical: the program calculated multiples of a specified number and the challenge was to write another program that calculated square roots using Newton’s method. The second was printing based: the program printed a triangle of asterisks and the challenge was to print a diamond instead. The third was graphical: the program printed various sized boxes and the challenge was to make the graphics look like a Mondrian painting.

While providing students these challenges is good and helps keep all of the students challenged and engaged, the choice of how to apply looping structures in different contexts seems even more powerful and important. I’ve been reading a bit about the disproportionate representation of women and minority groups in computer science classes and these types of choices seem to align with some of the suggestions of how to interest everyone in computer science.

I spent the past couple of days skimming Guzdial and Ericson’s Introduction to Computing & Programming: A Multimedia Approach. Their approach is fantastic; I’m going to start next semester by applying our new knowledge of arrays to image and sound processing.

Looking forward to next year, I plan to interweave various application strands (multimedia, computational modeling, numerical methods, graphics, databases) as we learn the various concepts. I hope to expose students a bit to all these strands while permitting them to spend additional time and effort on those that appeal most to them. I may field test some of these ideas next semester with capstone projects if time permits.

The particular methodology that I’m trying is best described by Stephanie Chasteen in her post “FTEP Effective facilitation of clickers workshop. She referenced a paper either on her blog or her podcast which described how a specific peer instruction technique was the most effective. The key elements are:

• students register their answer individually
• instructor shows distribution of answers but not which answer is correct
• students discuss, debate, and defend their answer; preferably with students who choose a different answer
• students register their answer again
• instruction show the distribution of answers and explains why that answer is correct

The research showed that this final step was critical.

In another post, Stephanie shared several resources for these types of questions for physics. I have found that Paul Hewitt’s Next-Time Questions also make excellent peer instruction questions as long as I post them in advance for students to consider outside of class with sufficient time for consideration.

I started the semester using Turning Technologies clickers, which I already use in some of my classes when administering exams. These worked fine. A colleague of mine showed me the NearPod app on the iPads, and I decided to try that for a change. While creating a presentation on the NearPod web site takes more effort on my part, I and my students have found using NearPod for peer instruction is better than the clickers. Because each student views the question on their own device, I can show more content-dense questions than I can when projecting on a screen in the classroom. In addition, NearPod allows me to create questions where students respond by drawing or annotating. This is perfect for graphical and diagrammatic answers. The only feature I miss when using NearPod is the lack of a countdown timer to remind students they need to submit their answers.

I have continuously been impressed with the level of engagement and the quality discussions that I observe during peer instruction. I’ve heard students devise novel and clear explanations when justifying their answer to other students. I’ve seen students leap into the air when they were in the minority but their answer is the correct one. I consistently receive very positive feedback about this element of class.

I hope to find a way to incorporate peer instruction into my Honors Physics class next year.

When reviewing kinematics, I took the opportunity to introduce computational modeling using VPython and the physutils package. I started with John Burk’s Computational Modeling Introduction and extended it with my experiences at Fermilab where computational modeling plays a role in everything from the optics of interferometers to the distribution of dark matter in the galaxy. I then provided students with a working example of a typical projectile motion model and let them explore. I encouraged them to extend the model to have the projectile launched with an initial vertical displacement.

Later that unit, I introduced the lab practicum which was based on a lab shared by my counterpart at our neighboring high school. The goal of the lab was to characterize the projectile launcher such that when the launcher is placed on a lab table, the projectile will hit a constant velocity buggy driving on the floor, away from the launcher, at the specified location. The location would not be specified until the day of the lab practicum. No procedure was specified and students decided what they needed to measure and how they wanted to measure it. I also used this as practice for writing clear and concise lab procedures like those required on the free response section of the AP exam.

All groups figured out that they needed to determine the velocity of the car (which some had done the previous year) and the initial velocity of the projectile. Some groups used a technique very similar to the previous year’s projectile motion lab where a marble is rolled down a ramp and launched horizontally. These groups fired the projectile horizontally from atop the table and measured the horizontal displacement. Groups that calculated the flight time based on the vertical height were more accurate than those that timed the flight with a stopwatch. Another group fired the projectile straight up, measured the maximum height, and calculated the initial velocity. This group was particularly successful. Another group attempted to use a motion sensor to measure the initial velocity of the ball as they fired it straight up. The motion sensor had trouble picking up the projectile and this group’s data was suspect. A couple of other groups fired the projectile at a variety of angles, timed the flight, and measured the horizontal displacement. Some of these groups later realized that they didn’t really need to perform measurements at a variety of angles. After gathering data and calculating the initial velocity of the projectile as a group, I asked the students to practice calculating their launch angle based on a sample target distance. I hadn’t really thought this lab through and didn’t appreciate how challenging it would be to derive an equation for the launch angle as a function of horizontal displacement when the projectile is launched with an initial vertical displacement. It wasn’t until that night that I appreciated the magnitude of this challenge and then realized how this challenge could be used to dramatically improve the value of this lab.

Most students returned the next day a bit frustrated but with an appreciation of how hard it is to derive this equation. One student, who is concurrently taking AP Physics B and AP Physics C, used the function from his AP Physics C text successfully. Another student amazed me by completing pages of trig and algebra to derive the equation. No one tried to use the range equation in the text, which pleased me greatly (the found candy discussion must have made an impact on them). As we discussed how challenging it was to solve this problem, I dramatically lamented, “if only there was another approach that would allow us to solve this complex scenario…” The connection clicked and students realized that they could apply the computational model for projectile motion to this lab. Almost all of the groups chose to use the computational model. One student wrote his own model in Matlab since he was more familiar with that than Python. With assistance, all groups were able to modify the computational model and most were successful in hitting the CV buggy. One group dressed for the occasion:

Students’ reflections on this lab were very positive. They remarked how they appreciated learning that there are some physics problems that are not easily solved algebraically (they are accustomed to only being given problems that they can solve). They also remarked that, while they didn’t appreciate the value of computational modeling at first, using their computational model in the lab practicum showed them its value. I saw evidence of their appreciation for computational modeling a couple of weeks later when a few of the students tried to model an after-school Physics Club challenge with VPython. For me, I was pleased that an oversight on my part resulted in a much more effective unit than what I had originally planned.

One of the first labs that we did was the ActorBox lab published as part of the A+ Computer Science curriculum materials. The students knew almost nothing about GridWorld and nothing about object-oriented programming, but they managed to figure out how to add Actors the the World. I introduced this lab as their first day on the job: “welcome to your new software job, here’s a large code base that you didn’t write and you don’t yet understand, figure out how to …”

Since then we have been learning the basics of Java, and now that we understand conditionals and loops, we can finally write interesting software. For the summative lab, I wanted to revisit GridWorld, emphasize the various phases of software development, and have them write a program that would be memorable. Implementing Conway’s Game of Life in GridWorld seemed to be the perfect fit.

I introduced Conway’s Game of Life last week and presented them with some initial requirements. However, I left a lot of the requirements unspecified. Students enumerated their own requirements on Friday and discussed them with me. I wanted them to realize that they needed to specify such requirements as the grid size, how to represent alive and dead cells, when the program finishes, and the time between generations. Students came up with a variety of ways of specifying these requirements with various levels of user specification. Students then started working on their design (flow charts or pseudocode; we haven’t touched on UML yet). This weekend, the were to enumerate their test cases and they will start implementation this week.

Through all of this, I continuously stressed that their requirements and design should be treated as living artifacts and will evolve as they progress through this lab. (I’m a huge proponent of Agile Software Development Methodologies, but that is a subject for another post.)

Since, at this point in the semester, my students are not that familiar with object oriented design, their solutions will be a functional approach and, most likely, inefficient and implemented with brute force. They will get to apply their newly acquired conditional and loop structures. I also think they will better appreciate the upcoming units related to object-oriented design as a result of this experience. In fact, I’m planning on a future lab to be going back and refactoring their Game of Life program.

Here’s the assignment that I posted:

Game of Life

I think this may be my first computer science, related post… cool.