AP students know that the CollegeBoard does not demand perfection to get a 5. In fact, if the famous Albert.io score calculators are to be believed, most tests only require a mere C — or around 70% correct — to receive that coveted score.
Allegedly the lowest threshold to a 5, AP Physics C: Mechanics requires only 55% correct, while the highest, AP Macroeconomics at 81%, still doesn’t seem impossible to achieve. However, when looking at the pass/fail rates for each AP exam, it turns out that most people either fail or get a 3, with 4s and 5s unattainable for the vast majority of students.
I’ve been personally reassured about AP exams I’ve taken through placating statements like “don’t worry, everyone else failed so the curve must be great.” On the contrary, the College Board pre-determines the threshold for each score using a variety of metrics that do not include actual students’ scores, meaning more than half of the test-takers can fail (the case for AP Environmental Science, AP US History, AP English Literature, AP U.S. Government and AP Physics 1).
One might argue that this failure simply reflects the lack of preparation students have for taking college courses. To simulate a true college exam, the College Board compares test-takers’ performance with college students’ performance every few years, and adjusts the threshold for passing according to how well the college students did. If AP test-takers are as “unqualified” for college as their average scores suggest, then the college students should outperform them.
Instead, in the years that the College Board does directly compare AP tests to college students (AP Biology was the most recent to receive a direct comparison, in 2022), the college students perform worse. The College Board then lowers the passing threshold to better represent a true college performance, improving AP scores for that year. As this clearly shows, more AP students have qualifications for college than the score distribution traditionally indicates, so why don’t their scores reflect that?
Rather than basing the threshold for each score on estimates/comparisons and not students’ actual performance on the exam, College Board should aim to create a normal distribution curve where most people get an average score (in this case, a 3), an equal number of people get a 2 and 4 and an equal number of people get a 1 and 5. Data tends to follow this trend anyway, so deviations from the standard suggest the test may have something “off” about it like being too difficult or too easy.
Allowing a standardized curve based on test-takers’ real performance rather than arbitrary numbers would mitigate the disparity between expectation and reality. Because AP exams are a one time deal, a curve would ensure a too difficult test doesn’t unfairly deprive a student of college credit, while also preventing a too easy test from giving credit to an unprepared student. AP tests should be hard, yes, but not so impossible that thousands of students flush away $100 for nothing.