This explanation could have benefited from histograms showing the distribution of grades pre-and-post making the test optional, and then showing statistical tests that the null-hypothesis, that the grades are the same, could be rejected.