With the rise of high school GPA inflation, Advanced Placement (AP) exams have become a go-to signal of college readiness for admissions officers.
But Harvard University Professor Paul Peterson argues the exams may not be as reliable a benchmark as they once were.
Writing in Education Next, Peterson points out what he sees as a clear shift in AP scoring after 2021, when the College Board introduced a new method for translating raw exam performance into the familiar 1-5 scale. The change affected nine of the most frequently taken AP exams: English Language and Composition, U.S. History, English Literature and Composition, World History, U.S. Government and Politics, Psychology, Biology, Human Geography, and Chemistry.
The results are hard to ignore.
Top score (5): The share of students earning a 5 increased from about 10 percent in 2021 to 17 percent in 2025, on average—a 61 percent increase. Under the old system, the share of 5s awarded in these subjects, on average, hardly changed over the previous six years.
Top two scores (4 or 5): In 2021, just 28 percent of test takers received a 4 or 5 on these nine exams. By 2025, that had jumped to 45 percent, a gain of 17 percentage points—or a roughly 63 percent increase.
Passing scores (3 or higher): The share of students receiving a 3 or better rose from roughly 52 percent to 71 percent over the same period, a 19 percentage-point increase—resulting in a 37 percent jump in passing rates.
At some point, a test that almost everyone passes is no longer a meaningful signal of achievement.
Meanwhile, less popular exams that didn’t adopt the new scoring approach, including Music Theory, Art History, Japanese, Italian, and Physics (Electricity and Magnetism), showed no such gains. Their results have remained flat or even declined, tracking more closely with national and international achievement data.
For high school counselors and college admissions committees, this means they “now need to look closely at which AP exams a student took, and in which years,” Peterson continues.
The College Board’s Senior Vice President of Advanced Placement and Instruction, Trevor Packer, rejects the inflation claims. He argues that the exams themselves haven’t changed but rather the “process of converting students’ specific AP Exam points into scores on the 1-5 scale.” In other words, Packer says the scores are now more accurate, not more generous.
Advances in technology have improved how exam data are collected and analyzed, allowing statisticians to draw on input from hundreds of college professors and better determine which skills individual students are demonstrating on their AP Exams.
This work is part of a rigorous, evidence-based standards setting process, overseen for AP by Dr. Amy Hendrickson, the current president of the National Council on Measurement in Education.
…
When evidence shows students are not meeting expectations, pass rates go down. When evidence shows more students are meeting them, pass rates go up.That’s exactly what we’ve seen.
From 2024 to 2025, pass rates rose significantly in subjects like AP English Language, AP Environmental Science, and AP Human Geography—because the student performance data showed more students were meeting the enduring learning objectives for these courses.
In other subjects, including AP Calculus BC, AP Statistics, AP Physics C: Mechanics, and both AP Government and Politics courses, pass rates declined—because fewer students met those enduring standards.
In both groups, the learning objectives for the courses didn’t change. The exams didn’t change. What changed is that students’ scores more accurately reflected whether they were meeting college-level expectations.
Peterson’s critique centers on the above named scoring process, evidence-based standards setting (EBSS). As he explains, EBSS (which is used by the nine popular AP courses with notably higher passing scores) replaces small panels of subject-matter experts with large surveys of instructors who recommend how scores should be distributed. In practice, Peterson continues, that shift tends to lower the bar because a broader group is more likely to produce more accommodating benchmarks than a smaller, more selective panel.
And because the exams themselves haven’t changed, adjusting the scale can boost scores without any real change in what students know.
Peterson also notes how this trend is hard to explain otherwise. The number of AP test takers has increased, so there’s no evidence the test-taking pool has become more selective, and there is little evidence from national or international assessments that student performance has improved dramatically in recent years.
Given the effects of grade inflation — not only on academic outcomes but also on college enrollment and lifetime earnings — what is actually gained by inflating AP scores?
Well, for one, the financial incentives are hard to ignore, Peterson argues. The College Board receives a significant portion of its revenue from AP exam fees. A system in which most students “succeed” is easier to market to schools, parents, and students.
At minimum, the shift makes AP scores harder to interpret than they once were. Admissions offices may have to pay more attention to which exams students took, and when.
That uncertainty has created space for alternatives. New programs like the Classical Learning Test’s Classical Baccalaureate (CB) program, developed with Arcadia Education, are positioning themselves as more rigorous benchmarks of student ability.
Serving 10th through 12th graders, the CB program includes both a diploma track and college credit exams designed to complement existing school curricula. Its pitch to colleges is straightforward:
Colleges and universities are increasingly looking for programs they can trust. With grade inflation, AI-generated essays, and even dual enrollment losing credibility, colleges need clear indicators and benchmarks that reflect academic mastery and intellectual formation.
Classical Baccalaureate is intentionally addressing where AP has lost rigor over the past twenty years. We’re restoring the depth that college-level courses used to have — especially in subjects like physics and calculus where critical content has been removed.
Whether colleges pursue alternatives such as the CB will depend on how willing they are to admit that current scores don’t tell the whole story and to do something about it.
