2011 Computer Science AP Reading
When I told my students I was going to be grading AP tests this year, they responded with “you’re leaving us for a week to sit in a room and grade all day?!” I was getting ready to leave just as students were putting finishing touches on their incredible final projects and by then I was sure I’d made a terrible mistake. How could flying across country to spend 8 hours a day grading possibly be anything but exhausting and mind-numbing?
Luckily, what I had heard from other teachers is true — the AP reading is an opportunity to get to know like-minded people, get new teaching ideas and learn a ton about the AP test. It was exhausting and at times mind-numbing but it was worth it. The week-long event brings together high school and college instructors who are so passionate about what they do and want to discuss it. Being around them was energizing and gave me hope. Though the days were grading-filled, we had time in the evenings to learn a lot from each other.
I graded question 3 which involved working with several objects and writing ArrayList algorithms. Each free response has a few question leaders in charge of creating a grading rubric for it and making sure that scoring is done consistently across the dozens of teachers assigned to the question. I was very impressed by both the question leaders and table leaders — the rubric was thorough and easy to understand and our training was robust.
Over the 7 days we were there, we graded over 22,000 exams. Grading on that scale really emphasized the importance of a well-designed rubric that accounts for a broad range of student (non-)solutions. In my own teaching, I’ve found APCS-style rubrics (example here) very useful in providing students with feedback they can immediately use. It leaves all guesswork out of figuring out what went wrong and students can make sure to avoid making similar mistakes in the future. I have no doubt that the fact that I grade both assignments and exams this way has increased student preparedness while also ensuring grading consistency.
Looking at thousands of responses to one question was eye-opening. Most notably, I was blown away by just how many students turned in solutions that were either blank or worth no points. I already knew this happened from reading student performance Q&A written by the chief reader every year — for example, in 2010 over a third of students got no points on the GridWorld question — but I wasn’t quite ready for just what that looks like. My second day, I happened to have a pack of 25 in which only six or eight had any points at all and that was really demoralizing. Many of those students left the entire test blank and several wrote long passages explaining that they didn’t know any computer science for various reasons. Why do they take the test? Do their schools make them? Why do they take the class? Does the AP designation mean more than actual learning? Why do teachers accept this? Do they know programming themselves? Whatever the cause is, I don’t think it’s very healthy to have so many students with no clue.
I was also particularly struck by how difficult it seems to be for students to understand which classes contain which methods and which class they are writing methods for. This certainly reinforced the notion that object-oriented programming is conceptually challenging and that it needs to be approached very deliberately.
Those who had been grading for years were able to put some of these kinds of thoughts in perspective and it was very interesting to hear their stories. Connections to these wiser, more experienced instructors are invaluable and I look forward to continuing to learn from them.