I'm thrilled that the College Board has created a brand new course and exam description for all AP physics courses. Now it is abundantly clear how much material each topic contributes to the exam - that wasn't clear previously for AP Physics 1.
Part and parcel with the revised CED, the College Board will now provide access to an electronic platform called "AP Classroom" with a vast test bank. Teachers will be able to electronically assign "personal progress checks" relating to each unit - these will consist of authentic multiple choice questions. Furthermore, I've been told that through AP classroom we can assign authentic free response items, organized by unit. This is wonderful.
The new CED does encourage categorization by unit. I know that's far more useful than previously, when the exam was organized by "big ideas" which overlapped multiple physics topics. Teachers - and students - prefer to think sequentially. Learning might be nonlinear, but time is certainly linear. We need to decide what topics and problems to work on each day of class, each week, each month... Yes, we absolutely must scaffold our content such that the same topics reappear in multiple contexts throughout the year. But we also have to start somewhere, and progress somewhere else. Unit guides and "personal progress checks" are outstanding places to begin.
So how do you set up your tests (or in education language, your summative assessments)? Do you test every unit? Every two units? Do you use the personal progress checks as unit tests?
No. I don't suggest using "unit tests" at all.
I still recommend that you set test DATES far in advance. Choose days that are unlikely to be interrupted by state championship meets, by regional honors band auditions, by the Ohio State-Michigan game, or by the first day of deer hunting season.
Don't move these dates for anything other than the apocalypse. These are not "unit tests"! They are, simply, tests. Just as the football team plays their scheduled game whether or not they've mastered the zone blitz, your class takes the scheduled test whether or not they're fully comfortable with Newton's Third Law. Part of the whole point of the test - or the game - is to find out just how comfortable the students are with the material.
I recommend AGAINST advertising the topics of each test, or even stating the topics to yourself. The first test will include, say, kinematics and dynamics, to the extent that your class has discussed each. The second test will include kinematics and dynamics and, say, circular motion and gravitation and energy, to the extent that your class has discussed each. The third test... well, you get the idea.
A "unit test" is a signal that what came before can be forgotten. A "unit test" encourages your students - and you - to either rush to get material on the test, or to postpone the test because you're not ready.
Just schedule test dates. Then put questions on each test that are cumulative until that point. That way you're doing six to eight AP practice exams throughout the year. That way you're demanding learning in context. That way you've that much more incentive to scaffold. That way our students have that much more incentive to correct each test, knowing that the material isn't going away. (Just as the football team has considerable incentive to watch film.) That way someone who is slow, who didn't really get kinematics or dynamics as it was presented, can show an understanding of that material later.