Our school has about six grading reporting periods – progress reports and semester grades. We use a piece of software (School Loop) to enter tests, quizzes and homework grades. IMHO, the software is not sufficiently flexible to handle variations in grading schemes. For example, last year I calculated letter grades based on a normal distribution and a curve that maintained the same z-scores for each student. I had to do this in Excel since School Loop could not handle these calculations. However, as teachers, we have a legal obligation to use School Loop. (Interesting challenge: define the dividing line between “we are all on the same page” and “assembly- line teaching”).
This sets up a conflict between the philosophy of SBG, which is based on learning and assessing separate objectives and the “one-grade-tells-all” scores that teachers must enter for the progress reports and the semester grades.
In addition, as an AP teacher, I feel that I need to give my kids a taste of the AP exam as often and as early as I can. Traditionally, this has meant chapter tests which were structured as 10 multiple choice questions, two free responses and something akin to the investigative task – all in about 60 minutes.
Therefore, the challenge is to extract from the SBG approach some sort of a single measure that can be put on a report card or that summarizes a unit test. A straight average would be the best choice if the test would consist of all different learning objectives (LOs). However, not all LOs are created equal. For example, I would think that knowing how to use z-scores should carry more weight than labeling a distribution as being right- or left- skewed. A further complication is that AP test questions are not necessarily limited to one LO per question.
Right now, my inclination is to divorce SBG assessments from unit exams. The unit exams can be 40% of the overall grade and the SBG based work 60%. This has the immediate appeal of making it easy to calculate a student’s grade for the report card. It signals to the student the importance of achieving mastery in all the LOs, but also weighs sufficiently heavy the tests that mimic the AP exam to signal their importance.
Traditionally I have not checked the students’ homework in AP Statistics. Partially this was because I did not have time to go thoroughly through all the problems assigned. In addition, I wanted to make the students take responsibility for their education. SBG offers a way of looking at homework as a separate, specific LO so the students have an additional incentive to do it.
General Helmuth von Moltke famously said “No [battle] plan survives contact with the enemy”. We can plan our SBG experience all we want, but only contact with the students in September will show how good our planning was.
Confidently Limited 
I really appreciate that you are documenting your thought process on SBG!