When assessing student performance, some educators choose to discard the worst tests written by the student. The mean of the remaining tests is taken as opposed to the mean of all tests written. Naturally, this process will result in some expected grade inflation. In our paper we provide a model for the expected grade inflation in the case when the student’s test-writing ability changes over time. Furthermore, we provide results on the grade inflation when the educator decides to use nested selective averaging processes. The Physics Laboratory post-tests' grades have been analyzed in our model.