Log in
A 150-year visual history of what happened to the digits of pi after they were calculated.
The recent national conversation about the “dumbing down” of AP Exams is missing the point. Most commentators have focused on what I have called “The Great Recalibration of AP Exams,” College Board’s multi-year project to raise the scores of millions of AP exams every year (see Figure 1). The Great Recalibration has led professors at Harvard University and the University of California to question the accuracy of AP scores and doubt the academic rigor of the AP program. It has even prompted Chester E. Finn Jr., one of the program’s most ardent defenders, to “suspect that the gold standard for academic achievement is losing its shine” (see “Are AP Exams Getting Easier?” July 19, 2024).
But emphasizing score inflation in this way is like focusing on the sack of Rome as the reason for its fall while ignoring the serious signs of decline across the entire empire that preceded it. Indeed, almost every aspect of the AP program has become easier than it was just a decade or two ago. Today’s AP exams require students to know less information, read shorter and simpler texts, answer more basic questions, and complete easier writing tasks with more time, more scaffolds, and more generous rubrics. While these changes are most obvious in humanities subjects like English and history, they are everywhere in the AP program. Through a thousand paper cuts over the past two decades, College Board has slowly bled its signature program of much of its rigor.
Tracking all these paper cuts is no easy task, especially because 42 AP courses are now on offer. The frameworks for these courses are created by College Board with the support of thousands of high school teachers and college professors who work on AP development committees and at AP exam readings. I’ve spoken to dozens of these insiders, and they assure me that no one just decided one day to make AP exams easier. They insist there is no nefarious plot at College Board to “dumb down” the program, nor any basis for the overtly racist claims that the exams have been made easier to accommodate non-white students. Instead, the AP development committees, which largely operate independently of one another, have tended toward lowering academic standards over the past two decades. The story is therefore one of mismanagement and neglect: it seems that no one at College Board had any incentive to maintain rigor in its signature academic program.
Whatever the causes for the lowering of standards, these committee members and readers have admitted to me that AP exams have become objectively easier than they were just a decade or two ago. College Board could have just conceded that AP exams have changed as academic standards have been lowered at the university level. That’s a defensible argument for maintaining the validity of AP exams as predictors of college grades, however inflated they may be. Instead, College Board has denied that the exams are easier and gone so far as to contend that AP exams have not changed at all: “AP Exam questions are not easier than in the past, nor are they harder. The exams themselves have not changed,” Trevor Packer, a College Board senior vice president, wrote in Education Next (see “AP Exams Are as Rigorous as Ever,” April 8, 2026).
In this article I catalogue the many ways AP exams have changed that make them measurably easier than they once were. The evidence tells a story more of gradual decline than sudden collapse—a series of individually defensible changes that, taken together, led the AP program on an inexorable trajectory away from its stated goal of providing a “rigorous academic curriculum” and exams that show that students are “ready for college-level coursework.” The result is that, rather than being an exception to the national decline in student achievement and the erosion of academic standards, the AP program is the perfect picture of it.
The Product of “Immoderate Greatness”?
Because AP exam volume has exploded by more than 700 percent over the past three decades, some commentators have portrayed the decline of rigor in Advanced Placement as if it were a “natural and inevitable effect of immoderate greatness,” to borrow an old phrase. As millions of new students flooded into the AP program, College Board faced persistent criticism that too many of them are struggling in AP classes and “failing” AP exams. And in an education environment where so many forces are conspiring to lower the bar, the AP program’s high academic standards never stood a chance.
In reality, the AP program could have maintained its rigor even as it became more inclusive of students of a wide range of abilities. It was always wise to remove irrelevant barriers and expand access to the program, but equity does not require students to know less content, read less complex texts, or produce less sophisticated arguments. The failure of College Board was not in expanding the program; it was that it could not sustain its lofty aspirations of bringing college-level work to high school students at scale.
How Do We Measure “Rigor”?
Determining whether one exam is “easier” than another is more complicated than it sounds, even for experts. “Rigor,” for example, is an overused buzzword that is notoriously difficult to define. Whatever “rigor” is, it should not be reduced merely to a description of changes in the format of AP exams. These changes are an important part of my analysis, but providing fewer answer choices or more time per question is not, in itself, proof of a decline in rigor. Some changes to exam format have improved AP exams by removing irrelevant obstacles and making questions clearer and fairer for everyone. Many of them are also defensible from decades of academic research about AP exams. Nevertheless, the collective effect of these changes is that they have made taking AP exams much easier for students over the past two decades.
The much more serious problem is that the underlying academic demands of AP courses have been greatly diminished hand in hand with the changes to the format of the exams. These academic demands are measurable in precise ways (that is, the amount of content knowledge an exam requires, the grade-level readability of reading passages, the requirements of scoring rubrics, and the methodology used to calculate scores). A rigorous AP exam that claims to be testing college-level knowledge should require students to know a substantial body of content, think about it in disciplined ways, apply it to the analysis of challenging texts, and produce written evidence of their mastery under consistently high standards. When AP exams cover less material, require less reasoning on more easily readable texts and questions, and reward weaker performances with the same scores, they have become less rigorous in their academic substance.
In this article I assess the evidence for the decline in rigor on today’s AP courses and exams across seven distinct categories. In each instance, the story is a variation on the same theme: Course and exam developers have made AP exams meaningfully and objectively easier as they attempted to improve the AP program. Crucially, they also have made the courses less rigorous, even if they may not have had a particular intention of doing so. Many of these changes could be defended in isolation, but nearly all of them have moved the program in the same direction away from the rigor that once defined AP exams as a “gold standard” for academic excellence.
1. Reduced Content Requirements
The lowering of standards in the AP program long predates the negative impact of Covid or the rise of artificial intelligence. The crucial turning point was a series of course “redesigns” that began in the early 2010s. This overhaul was intended to address persistent, legitimate criticisms that AP courses were overloaded with too much content and focused too much on mere factual recall. These new AP courses were supposed to emphasize “depth over breadth,” allowing AP teachers adequate time to cover material and focus on critical thinking skills beyond rote memorization. The redesigns also explicitly aligned AP courses to the Common Core standards that College Board’s then-incoming CEO, David Coleman, had helped create.
Whatever the intentions, the result was that most AP courses were redesigned to reduce content, lower requirements, and simplify exam questions. In an era when factual recall has been de-emphasized in K–8 education, hundreds of thousands of high school students are now entering AP classes with weak content mastery and poor foundational skills. As one AP textbook author argued, “[T]he problem is that AP exams are increasingly de-emphasizing actual knowledge and focusing more on ‘holistic’ skills.” The choice to turn away from memorization and reduce content within AP courses has helped create a whole generation of AP students who know shockingly little of the content they have allegedly studied at a college level in AP courses.
None of this should be a surprise. This exact outcome was predicted in a 2007 report produced by the Thomas B. Fordham Institute, before the redesigns had even occurred:
Our fear is that actual content knowledge will be de-emphasized as “big ideas” eclipse facts in a potentially extreme tendency to favor depth over breadth. The College Board should take great care to make sure that doesn’t happen.
College Board did not, in fact, take great care, and the results are clearly visible on college campuses, where many professors have lamented that AP students are clearly unready for college-level work:
At the University of Texas at Austin, where I am a professor of history, we see the consequences of this trend firsthand. Students who pass the AP U.S. History exam receive credit only for the second half of our U.S. history survey course. To earn credit for the first half, they must pass a faculty-graded essay exam. A substantial majority of these students fail to demonstrate the content knowledge or writing skills that we expect from a university-level history student. This reality exposes the gap between what AP courses promise and what they actually deliver.
Similarly, according to one English professor,
At Colby College, where I teach . . . many students arrive here and elsewhere having taken A.P. classes. . . . [W]hen I teach a first-year writing course, or an introductory course in the English major, I inevitably spend the first few weeks undoing the damage the A.P. system does to how students understand both writing and the study of literature.
These complaints have been echoed by many university professors, most recently by Professor Paul E. Peterson of Harvard University. The concerns transcend disciplinary distinctions: One math professor has argued that AP courses are not truly college-level courses. “This equivalence is an illusion,” he says. “Calculus at my university, Johns Hopkins University, could not pass the audit to qualify for AP Calculus. This is equally true for many other universities and colleges.” Likewise, an economics professor has identified that the free-response questions of AP Macroeconomics “test mid-level understanding” rather than rigorous, college-level work. This crisis of standards has led some to call upon College Board to “increase the rigor of its advanced placement examinations to satisfy these colleges’ and universities’ expectations as to what constitutes genuinely collegiate-level instruction.”
It’s not just college professors beating the drum. After the redesigned courses were first launched, some AP teachers decried what they saw as a “dumbing down” of the program. When AP World History was stripped of thousands of years of the past, many AP teachers were “outraged” that their students would “lose a broad global understanding of history” as the course became more narrowly focused on modern and Western history. Across a range of subjects, critics began to question whether the redesigned AP courses were still at a college level. About the new AP U.S. Government and Politics course, one suggested “it is time to end the charade that an AP course is equivalent to a college course.” An AP Environmental Science reader observed that “over the years, the math has gotten easier on the FRQs,” while a veteran AP Chemistry teacher responded to College Board’s changes to AP Chemistry by maintaining the rigor in his class, even if he had seen “the level of difficulty plummet” on the exam. “[W]hen it came time to review for the test,” he said, “I told them that the AP exam will only have the ‘easy things’ on it, not the hard things.” Critics have pointed out that the new AP Statistics framework has reduced course content more than it did in the last redesign in 2018, which had already simplified the course. The changes led the South Carolina Department of Education to change its policies because College Board has lowered the rigor of the course so significantly:
College Board has removed mathematical content from the AP Statistics course, thereby lowering the rigor required for a fourth higher-level course.
Likewise, the new AP Latin course framework has slashed the required lines of Vergil’s Aeneid in half and left a major part of the syllabus up to “teacher choice units.” Similarly, AP Spanish Literature’s redesign reduced the required reading list by one third (from 56 to 38 titles). AP Computer Science has been criticized for a number of similarly “concerning revisions” over the past two decades, and AP Chemistry has been accused of being “dumbed down” for years. After College Board “cut 40 percent of the content out of” AP Biology, a veteran AP teacher observed that “the course now is a little bit easier on the students.” College Board continues to revise courses to remove content and lower expectations across most of the full scope of 42 AP courses. Taken together, these changes show that the noble goal of ensuring AP classes had “depth over breadth” has devolved into AP courses with consistently less depth and less breadth.
2. “Early Access” and “Career Prep” AP Courses
The AP program has also loosened its standards in developing new, less challenging AP classes. This process began with College Board’s strategy to bring its so-called college-level AP courses to high school freshmen and sophomores in subjects such as AP Human Geography, AP World History, and AP Seminar. These “Early Access” courses are intended to function as a gateway to the AP program. AP Computer Science Principles, for example, was intentionally created for students who needed, according to College Board, “more of a gentle introduction” to computer science. The result, according to one critic, is “a dumbed-down curriculum” that “probably does more harm than good.”
Recently, College Board launched its easiest AP math course ever: AP Precalculus. While AP Precalculus may be a well-intentioned way to bring the AP program to younger and more diverse students, the course is, in the words of one college math professor, “high-school-level material, not university-level material, and so offering AP credit for it makes no sense.” To make matters worse, College Board quickly slashed 25 percent of the AP Precalculus curriculum in order to appease some critics of the new course. The result is that today’s AP Precalculus course is 75 percent of a high-school course rebranded as “Advanced Placement” without a corresponding increase in difficulty.
In addition to these courses, College Board is in the process of launching new “Career Kickstart” courses as “a new entry point to AP for a broader range of students.” AP Business with Personal Finance, for example, was created with the express purpose of “career preparation” for “those who may not be immediately college-bound.” College Board’s CEO, David Coleman, indicated that this aligns with a new vision of Advanced Placement: “How do we make high school relevant, engaging, and purposeful?” he asked in an Education Week interview. “Bluntly, it takes [the] next generation of coursework. We are reconsidering the kinds of courses we offer.” Whatever the benefits of these courses in increasing access, they are expanding the Advanced Placement brand away from college-level work in favor of “career prep” and advanced high school work. In other words, they are making the AP program easier.
3. Simpler Multiple-Choice Questions
The revisions to multiple-choice questions provide an excellent example of how minor revisions, however defensible by psychometric research, have the cumulative effect of making AP exams easier to take. For example, College Board eliminated the “guessing penalty” in 2011, a change grounded in research that shows that such penalties reward risk-taking more than content mastery. Whatever the motive for the change, AP students can now earn more points by randomly guessing when they do not know the answer.
College Board is in the process of reducing the number of response options on multiple-choice questions from five to four on most exams. While this change is also justifiable, it will certainly “make the multiple choice portion of the exam less of a beast.” Many exams have been redesigned to reduce the total number of questions, allocate more time per question, or otherwise make the section measurably less challenging. Critics have noted that the AP Spanish Language exam became “easier to pass” and that all the foreign language courses have stripped away complex grammatical constructions in favor of “holistic” topics. As Erica L. Meltzer wrote in her blog, The Critical Reader:
[U]ntil a few years ago, most foreign-language AP exams required students to complete a series of grammar fill-ins testing verb tenses, prepositions, and idiomatic usage. . . . But fill-ins have now been eliminated entirely, replaced by more “holistic,” “higher-order” assignments.
Across a number of math and science subjects, students now have access to more formula sheets and graphing calculators on the multiple-choice section. These are probably all reasonable choices, but they also make the exams less challenging in the sense that they require less “mental lift” on the part of AP students.
More importantly, many multiple-choice questions have been changed in a way that requires less knowledge of academic content. On the AP English Language exam, College Board replaced almost half of the multiple-choice reading questions with much easier writing-revision questions. Regarding these changes, one AP English Language textbook author argued, “the reading passages will be shorter and have fewer questions, and the writing passages will be written at a considerably lower level than any of the multiple-choice passages on the current exam.”
To see how different multiple-choice questions have become, compare an AP U.S. History multiple-choice question from the 1994 exam with one from the most recent course framework.
It is true that the old format of multiple-choice questions required “rote memorization and recall under pressure” rather than deeper “academic and intellectual investigation in a college class,” in the words of one critic. Nevertheless, to prepare for these kinds of questions, AP students needed to learn a lot of specific information about all sorts of topics in U.S. History. Multiple-choice questions on today’s exam provide a “stimulus” in the form of a text or image, but they do not require a deep understanding of anything to find the correct answer. What prior knowledge would an AP student need in order to guess that a picture of women building an airplane in the early 1940s shows that World War II mobilized American society? You could be forgiven for advising today’s AP U.S. History students to focus on studying broad themes of history rather than the substantive facts that undergird it.
In trying to address criticism over too much rote memorization in AP courses, the AP program created a new problem by diluting the difficulty of multiple-choice questions, which now have no guessing penalty, fewer answer choices, more time per question, more formula sheets, more calculator access, a more generous scoring scale, and, most importantly, less academic substance. Individually, most of these changes could be considered improvements to the exam, but taken together, they make the multiple-choice sections of many AP exams less difficult.
4. More Basic Free-Response Prompts
On many AP exams, the free-response prompts are not only easier to answer, but they also test less academically substantive topics. To illustrate: Compare the Question 3 prompt I faced on the AP English Language exam back in high school to an equivalent prompt from the 2025 exam.
From Shakespeare to People magazine, from “the strong lance of justice hurtless breaks” to “trying to live in the moment and enjoy the journey,” the AP English Language exam has not merely been modernized or made relevant for high school students. It has been drained of the challenging academic material and serious scholarly questions that used to characterize the exam.
Even when a current test topic corresponds with one from the past, the content of questions is less complex than it once was. Compare the 2001 AP English prompt that asked students to analyze a philosophical meditation on how photography limits our understanding of the world with the 2024 prompt that asked students to respond to a brief quote from a newspaper about selfies.
AP English Language in 2001: analyze this philosophical meditation on how photography can limit our understanding of the world
AP English Language in 2024: respond to this quote from a newspaper about taking selfies https://t.co/gUa6vyN0sW
— John Moscatiello (@HookedonJohnics) April 22, 2026
This isn’t happening only on the AP English Language exam. On many science exams, the free-response questions may appear more difficult but are actually much less so. As part of the overall redesign of AP courses, the focus of these questions has shifted away from demonstrating mastery of a wide range of concepts to more application-based questions that require a narrower base of knowledge. The redesign led one AP Chemistry teacher to argue as far back as 2013 that the free-response questions on the AP Chemistry exam were “hopelessly dumbed down.” The same author has shown how the 2025 AP Chemistry exam is designed in a way that “hand-holds” students with prompting, clues, and step-by-step guides to help students earn points more easily. He concluded:
The single, significant take-away for me from the 2025 released FRQs is this; “Oh, that’s exactly how I would dumb down an AP problem to make it easier for my honors or even my regular classes to be able to answer it.”
If you suspect I am exaggerating here, you really need to read the free-response questions from the 2025 AP World History exam for yourself. The first question asks students to “identify one claim that the author makes in the first paragraph” of an excerpt from an “academic book.” To earn full credit on this part, AP students need only write one or two sentences that successfully identify any relevant claim in that paragraph. And the “academic book” in question? It’s actually a popular-press book written at a high-school reading level.
College Board: “AP Exam questions are not easier than in the past, nor are they harder. The exams themselves have not changed.”
2025 AP World History Exam: “Identify one likely audience” for a women’s newspaper. [Spoiler: a correct answer is “women.”]
— John Moscatiello (@HookedonJohnics) April 20, 2026
The second short-answer question is even simpler. It presents students with an excerpt from a women’s newspaper and asks them to “identify one likely audience” for that text. To earn full credit on this part, AP students need to write just one or two sentences that identify women as one likely audience for a women’s newspaper. If these sound more like questions for middle school than for college-level exams, that’s because they are. You do not need a PhD in psychometrics to see the difference.
How much content knowledge do you need to earn points on each part of this question from the 2026 AP European History exam?
Can you answer this “college level” question from the 2026 AP European History exam? (If you can write answers like these sample ones, you could earn full credit on this part of the exam.) https://t.co/RQSh9dV6GE
— John Moscatiello (@HookedonJohnics) May 18, 2026
5. Less Complex Reading Passages
One of the most striking changes to the exams in the past two decades is the gradual replacement of dense, challenging texts with simpler texts by more contemporary authors. These simpler texts are often accompanied by definitions of straightforward vocabulary. Incredibly, a 2025 AP English prompt included footnote definitions for words like “optimism” and “pessimism.”
@robertschombs Look at the footnotes on 2025 AP Lang exam. Apparently, we shouldn’t expect students reading at a “college level” to know the difference between “optimism” and “pessimism.” https://t.co/zRCCzS8LRJ
— John Moscatiello (@HookedonJohnics) April 22, 2026
The decline of difficulty in reading passages is measurable. For example, consider the prose passages on free-response Question 2 of the AP English Language exam. When they are measured on traditional scales of readability (Flesch-Kincaid, Gunning Fog, and SMOG), it is clear that these passages have become easier over time.
They just gave up on pretending AP exams are “college level work” https://t.co/heTpV3BzqB
— John Moscatiello (@HookedonJohnics) May 17, 2026
These reading scales are just three of many metrics used to assess the complexity of texts, and this is just one question type on one exam. But I have run the data using many metrics and across different question types, and the story is always the same: AP English Language and AP English Literature passages are becoming simpler, with shorter sentences, more basic vocabulary, and less sophisticated ideas. While readability metrics are not a perfect science, and the readability of a passage in itself is not proof that the exams are easier, these numbers provide further verification that the apparent decline in rigor is supported by more than mere anecdotal evidence. How many more grade levels will AP reading passages sink before we begin doubting whether they represent college-level material? Will we ever see a passage like the 2006 AP English Language Question 2 on an AP exam again?
@tzenes Behold, the 2006 AP English Language free-response Question 2: https://t.co/j06zjCyUcp
— John Moscatiello (@HookedonJohnics) May 18, 2026
6. Reduced Rubric Requirements
Even though the prompts have become simpler and the passages easier to read, the free-response questions on AP exams could still challenge test takers if the scoring system compensated for the decrease in difficulty. But the system of scoring AP essays has been transformed from holistic rubrics that rewarded coherent, unified writing to analytic (or point-by-point) rubrics that reward fragmented, mechanical writing. The new analytic rubrics are more user-friendly and perhaps even fairer than the old rubrics, but they have the pernicious effect of encouraging AP students to do the bare minimum needed to “earn a point” rather than striving to write at a college level. The ability to earn points for individual sentences or fragments of ideas leaves almost no incentive for students to try to write elegant, unified prose. This is why influential AP history teachers on YouTube explicitly recommend that AP students not write a concluding paragraph on their essays. With no points awarded for writing a complete essay, they might as well write a partial essay instead.
All AP free-response essays are scored by human readers. In the past few years, many of these readers have reported that it is “obvious that College Board wanted us to find opportunities to give points.” A veteran AP English Literature reader recently revealed that “this year they’ve told us to be exceptionally generous with giving points.” Another reader for that exam observed that, whereas previous exams required students to select works of “literary merit” on free-response Question 3, “Nowadays readers are told to read and score the essay according to the benchmarks and scoring guide regardless of what the work is.” A university professor who has graded exams for many years claimed that College Board “changed the format and I have seen the scores higher but the quality of work much lower.” Another university professor who reads for AP history exams told me that “the default was to look for the minimum threshold and award the points.” The professor went on to say that “only very rarely did I read an answer and think, ‘I would gladly accept this in one of my gen ed courses.’” Even a decade ago, readers were reporting that they were being pressured to give away points more freely:
Over the past few years, [The College Board] has subtly pressured graders to increase student scores. This year the pressure is no longer subtle. Graders have been told to “give students the benefit of the doubt” and “set the bar very low.” (Those are actual quotes.)
A high school teacher who reads for AP Capstone exams wrote to me that she was “stunned” when a table leader instructed her to award points for work that was clearly below the standard of the published rubric. Similarly, an AP Environmental Science reader has exposed how the criteria for the rubrics can be modified during an AP reading so that “exams graded later in the week are more likely to receive higher scores.” While AP readers are deeply divided on this issue, a vocal subset of veteran AP readers has lamented the conversion to analytic rubrics and the consistent lowering of the bar.
Once the most important part of AP history exams, document-based questions (DBQs) have become demonstrably less challenging. Students used to be presented with about 10 documents and were encouraged to cite them all in their responses. Now, students in AP European History, AP U.S. History, and AP World History are presented with only seven documents and are only required to integrate four of them into their essays to earn all of the evidence points on the newly simplified rubrics. And because essays are scored on analytic rubrics rather than holistic rubrics, students can earn points for mere sentences or clusters of sentences rather than entire essays. The result is that exams have become “WAY EASIER!” according to Steve Heimler, publisher of the website Heimler’s History.
BREAKING: The College Board has updated the rubrics for AP History courses. In DBQs, students only need to cite 4 documents instead of 6 and source 2 documents instead of 3. To read the full set of changes, see the 2023 CED for your course and go to the “AP History Rubrics” section.
— Marco Learning (@marcolearning) September 12, 2023
One advocate for this lowering of the bar lamented that the old rubrics required students to weave a complex argument throughout their essays, an approach that, in his words, “unfairly benefited those who could write in a complex manner and, thus, those who came from privileged backgrounds.” Never mind the assumption that students from less privileged backgrounds are not really capable of writing “in a complex manner.” In the name of improving the program, College Board has devalued great writing in a program that ostensibly prepares students for the rigors of America’s colleges and universities.
EdNext in your inbox
Sign up for the EdNext Weekly newsletter, and stay up to date with the Daily Digest, delivered straight to your inbox.
7. Higher Scaled Scores
Over the past two years, we have been engaged in a national conversation about College Board’s recalibration of AP exams. College Board has defended these changes by appealing to a relatively new psychometric methodology called evidence-based standard setting (EBSS). Even though its explanations of this methodology are confusing and unconvincing, even to professors at Harvard University and the University of California, College Board has assuaged some concerns, perhaps in part because most people know so little about psychometrics and its limitations in the first place.
Regardless of these psychometric explanations, AP scores are not actually set by “science” in the abstract. They are set by people. And the College Board’s own research has shown that these very people, the hundreds of professors and teachers who set standards every year, have presided over massive grade inflation at their own institutions. How are these same people not importing their inflated grading standards into the AP program?
College Board continues to compare AP scores to college grades with no statistical adjustment for the college grade inflation it has criticized. It has argued:
Typically, AP Program’s high standards result in 60%–75% of AP Exams earning a score of 3 or higher, which is equivalent to a college C or better. By contrast, colleges in our standard-setting studies award grades of C or higher to approximately 75%–90% of their students in similar courses.
But if 60 to 75 percent of AP Exams “earn a score of 3 or higher,” it is because College Board has increased scores on exams it has simplified over the past two decades. Even if its new scoring method is a superior methodology, it cannot establish that today’s exams themselves demand as much knowledge and skill as earlier exams did.
College Board has also cited decreasing scores in multiple subjects to prove that its new methodology does not merely inflate scores:
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.
This claim is technically true in that very minor decreases in scores did occur in those five subjects in a single year (between 2024 and 2025). But as Figure 2 demonstrates, the claim is profoundly misleading because four of those five subjects now have higher “pass rates” than they did just a few years ago.
Using contemporary college performance as the benchmark for AP standards poses a problem because college standards themselves have declined. This makes College Board’s next claim especially difficult to believe:
Even as college grades have drifted upward over time, AP has maintained high standards for performance. In fact, the AP standard of achievement in 2025 is more rigorous than the standard of achievement in colleges in 1990—35 years ago! (average college GPA in 1990: 2.8. GPA equivalent of the average AP score in 2025: 2.5).
Again, the “GPA equivalent of the average AP score in 2025” is a number that College Board has intentionally increased as part of its stated goal to increase scores so that nearly all exams result in 60 to 75 percent of test takers earning “qualifying” scores. This high percentage is also based on exams that are objectively less difficult than they once were.
Yet the real problem lies not with the higher scores themselves nor with the methodology used to justify them. The old scores could have been artificially low, and the methodology could represent a major improvement to the AP program. The problem is that AP scores are soaring even as College Board has reduced content requirements, simplified question stems and answer choices, selected easier reading passages, lowered rubric requirements, and made many other changes that have diminished the academic substance of AP courses and exams.
Restoring Credibility to the AP Program
While some AP courses have retained robust academic standards and some exam questions are still challenging, College Board has lowered the bar in so many ways across so many subjects that the entire program has become an unrecognizable shadow of its former self. The AP program many of us remember from high school is long gone.
So, what do we do now? The AP program administers more than 6 million exams every year, and College Board has decided that around 4 million (that is, 60 to 75 percent) of those scores should be worthy of college credit. Do we trust that College Board and its many committees of professors and teachers have kept highly rigorous standards, even when all the evidence is mounting that they have not? Even when the official explanations for this scoring are clearly specious?
What we must not do is rush to change policies that offer college credits for AP scores. Doing so would deprive students of a meaningful way of saving money on college tuition. It would also punish students for the failure of College Board to maintain high academic standards in the AP program. This is College Board’s problem, and College Board needs to offer a solution that goes beyond telling us not to believe our lying eyes about what these exams have become. Instead of rushing to change credit policies, universities should follow the University of California’s lead and demand more information from College Board about its opaque scoring methodology. They should also demand that College Board explain how it will restore rigor to the AP program over the next few years.
Whatever happens, College Board needs to be transparent and honest in a way it was not during the redesign of the AP African American Studies course in 2023. It needs to stop making easily falsifiable claims like “AP Exam questions are not easier than in the past, nor are they harder. The exams themselves have not changed.” It needs to allow for independent, third-party research to verify its incredible claim that “the AP standard of achievement in 2025 is more rigorous than the standard of achievement in colleges in 1990.” It should also conduct studies that compare student performance on today’s AP exams with student performance under equivalent standards from 10 or 20 years ago. How would today’s students perform on intact 1995, 2005, or 2010 AP exams under the old scoring standards? How would earlier cohorts perform on today’s exams? If AP exams are as difficult or even more difficult than they once were, then let equivalent student performance data prove the claim.
College Board must undergo an honest reckoning with what has happened in the past two decades and a meaningful restoration of college-level rigor to the AP program over the next few years. The organization should follow its own research that shows that scoring as low as a 2 on an AP exam still has benefits. If so, then we should just let students score a 2 and reap those benefits. Isn’t it better to score a 2 on a real AP exam than to be offered a participation trophy of 3 on a simplified AP exam that does not really prepare them for college-level work? As educators, we all need to accept the fact that not everyone deserves an A or a qualifying score on an AP exam, even if we face enormous social pressure everywhere to lower the bar.
It is neither elitist nor racist to advocate for the highest academic standards in the AP program. We do not need to long for the “good old days” of an exclusive and exclusionary AP program, and we do not need to revert to the old problems of AP exams from the 1990s or early 2000s. Reclaiming academic rigor in the AP program will be challenging (and probably even painful), but it will be good for everyone. It may be the AP program’s best chance of recovering its credibility and saving itself from what may very well be “the natural and inevitable effect of immoderate greatness.”
This essay is adapted from an article originally posted at X.
John Moscatiello is a teacher and administrator at The Wilberforce School in Princeton, New Jersey. He is the author of “The Great Recalibration of AP Exams.”
The post The Decline and Fall of Rigor in Advanced Placement appeared first on Education Next.
From Catriona Agg’s many-splendored Twitter feed:
Two congruent rectangles are packed into a square. What’s the angle?
All models are wrong, but some are useful.
-George Box
A basic observation in education is that learning is easier for some students than others. Some students seem to soak up everything around them. Others, despite putting in effort, learn much more slowly.
These variations exist at the school level as well. I’ve worked in a few schools, with varying levels of student achievement. My experience, and the experience of lots of other teachers, is that teaching is harder in some schools than others. At every school I’ve worked at, I’ve worked with lots of passionate and hardworking teachers. Yet in some contexts that effort seems to translate into success much more easily than in others. This correlates strongly with poverty: schools with more students living in poverty struggle more than those in affluent areas. There are plenty of other variables as well, but the basic trend is obvious to anyone who has worked in a few different contexts.
There are two narratives that I don’t think are helpful in understanding this question. The first is that teaching in some schools is harder because of the circumstances the students in them are born into, and schools cannot overcome those circumstances and deliver a quality education for students in poverty. I don’t think that’s true. There are plenty of counterexamples: schools that do help students rise above their circumstances.1 They are far less common than we would like, but that evidence exists. The second narrative is that failing schools are failed by callous or incompetent adults. I haven’t been in every school, and I don’t doubt that there are examples of incompetence in some schools. But there are also incompetent teachers and leaders at affluent, successful schools. My experience is that the vast majority of teachers are hardworking and care about their students. The differences in achievement between schools are large; I don’t think they can be explained by the quality of the teachers in them.
“All models are wrong, but some are useful.” What I’m going to describe is a mental model for why some schools are harder to teach in than others. It’s a model. I’m not arguing this is a perfect description or captures everything important about schools. But I do think it captures a decent chunk of the truth, and it’s a useful model for understanding schools and what those who work in challenging schools can do to improve outcomes for students.
Cognitive Resources
Humans vary in random ways. Two examples of that variation are working memory capacity and processing speed. Working memory capacity refers to how many ideas can be held in mind and manipulated at once. Working memory is where we think, and when learning something new it’s useful to be able to juggle multiple ideas. Processing speed refers to how rapidly we can process and reason about new information.
Working memory capacity and processing speed are two major components of what researchers call “fluid intelligence.” There’s plenty of research establishing that humans vary in these two capacities. Beyond that research, these sources of variation match the experience of many teachers: we can see that some students are quicker thinkers, and some students can hold more in their mind at once. It’s also my experience that most teachers avoid judging students on these capacities: we strive to teach all of our students, and we don’t think a student deserves a better or worse education because they are a quicker thinker.
To be concise, I’ll refer to varying levels of working memory capacity and processing speed in this piece as “cognitive resources.” My premise here is that students vary in the cognitive resources they are endowed with, and that variation is random.2
Learning
More cognitive resources make learning easier. Students with fewer cognitive resources are still capable of learning, but they require better teaching. The primary consequence of differences in cognitive resources is how information comes in. Working memory is the bottleneck between the learning environment and long-term memory.
Cognitive resources don’t say anything about what students can or cannot learn, but they do influence the way that information needs to be delivered in order for learning to happen.3 Peps McCrea calls this “instructional sensitivity.” Students with more cognitive resources are not very sensitive to the quality of instruction: even with mediocre instruction, they are still likely to learn. Students with fewer cognitive resources are much more sensitive to the quality of instruction: with excellent instruction they can achieve at high levels, but without it they are likely to struggle.
The key to this type of instruction is managing cognitive resources: teaching in a way that reduces the load on working memory. Here is a rough list of teaching practices that are important for students with fewer cognitive resources:
Provide plenty of practice, including spaced and interleaved practice, and continue practicing until students reach a high success rate. Practice moves knowledge securely into long-term memory and reduces the load on working memory.
Break concepts into small steps and provide practice after each step. Focusing on one step at a time reduces cognitive load and ensures each step is secure before moving on.
Manage attention to focus student thinking on key ideas. Memory is the residue of thought; students learn what they think about. Instructional design should be laser-focused on the exact concepts we want students to think hard about, leaving as little to chance as possible.
Make explicit connections between ideas. The best way to learn something new is to connect it to something already in long-term memory. Those connections make learning durable. Students with more cognitive resources are more likely to make connections spontaneously; students with fewer benefit from more guidance in making connections.
Sequence ideas logically. Learning is cumulative, and learning works best when it builds logically on what students already know. Careful sequencing reduces cognitive load by relying more on prior knowledge and less on processing large amounts of new information in working memory.
Model thinking and provide plenty of examples. Reduce the burden of figuring things out independently by giving students clear models of what we want them to be able to do.
Provide plenty of review. Students with fewer cognitive resources benefit from plenty of practice and making connections between ideas, and this means providing lots of review of previously-learned concepts.
Provide scaffolds for difficult tasks. Ask students to do challenging things. Don’t lower the bar. But also recognize that challenging tasks will inevitably involve increased cognitive load. Provide scaffolds to help reduce the cognitive load and focus students on essential concepts, then fade those scaffolds over time.4
Compensating Factors
Cognitive resources aren’t the only thing that matters for learning. There are a number of compensating factors: factors that can help a student with fewer cognitive resources succeed even without excellent instruction.
Outside instruction. This seems obvious: with additional instruction, students will do better. This can come in lots of forms. Private tutoring. Supplemental instruction at programs like Kumon or Russian School of Math, which are common in some parts of the US. Several Asian countries have built reputations for educational excellence on a huge system of out-of-school supplemental instruction. And then there’s mom at the kitchen table, helping with homework or reteaching something their kid is confused about.
Prior knowledge. The more you know the easier it is to learn. The best way to bypass the limitations of working memory and processing speed is to know a lot about a topic: that knowledge reduces the strain on working memory. This is one big reason why small differences early in education snowball and become much larger over the years. Students who fall a bit behind find learning harder and fall further behind; students who get ahead find learning easier and continue to accelerate.
Motivation. Many teachers will have experience with students for whom learning is hard, yet for whatever reason they are determined to learn. That motivation can compensate for fewer cognitive resources. The important part of motivation for this argument is that motivation is contagious. If everyone around you is putting in a lot of effort, you are likely to do the same. If everyone around you is slacking off, you’re likely to follow suit.
These compensating factors can, to an extent, mitigate the challenges of fewer cognitive resources. They can’t work miracles, but they can help around the edges, shifting the distribution to the right.
Deprivation
While cognitive resources may be randomly distributed at birth, that’s not the outcome teachers see in schools. Unfortunately, there are lots of ways those cognitive resources can be sapped. The simplest way to conceptualize this is to focus on poverty.
Poverty is correlated with all sorts of factors that sap cognitive resources: students are worried about where their next meal is coming from, or whether their housing is secure. Chronic stress takes a toll on attention and executive functioning. Trauma leaves behind scars. Poor sleep, inadequate nutrition, and exposure to environmental hazards can all affect cognitive development. Over time, these deprivations erode cognitive resources and make learning harder. Poverty is a scourge. While its effects vary and the exact mechanisms are tricky to pin down, on average a student living in poverty is likely to have a tougher time in school.
These aren’t exclusive to students living in poverty. Plenty of affluent families have a hard time, and many poor kids are unaffected by challenging circumstances around them. There’s a distribution. If we conceptualize cognitive resources as a spectrum, poverty shifts that distribution to the left. There are still well-off students who struggle, and there are plenty of students living in poverty who excel, but the distribution is shifted. The exact amount will depend on the details of the community.
Why Is Learning Harder In Some Schools Than Others?
Now we can put together an explanation for why some schools are harder to work in than others. In an affluent district, you have that random distribution of cognitive resources. Fewer students have their cognitive resources sapped by the effects of trauma and poverty. Many students who might otherwise struggle get outside instruction in ways large and small to compensate. A critical mass of students are successful, and this creates a culture of learning and a sense of order in the school. Some students will still have a hard time, but it will be easier to support those students because there are fewer of them, and precious resources can be directed to helping anyone who struggles. Test scores will be good, which will continue the virtuous cycle. There is plenty of variation here: in some affluent schools, more students will fall through the cracks. This isn’t a story where affluent schools are perfect. The averages often hide a lot of failures at the left end of the distribution. But the averages look good on the surface, so we tell ourselves these are the good schools.
In a high-poverty context, the basic distribution starts in the same place. But more students experience deprivations that sap precious cognitive resources. Fewer students who struggle get outside instruction to compensate. The difference is small at first, but over time the gaps in prior knowledge compound. More students feel dumb and frustrated because they know they aren’t learning the way they are supposed to. It becomes harder to keep order in the school. Low motivation can spread and become the norm. There’s a lot of variation here: some schools will do a better job creating a strong culture of learning. Some will provide the type of effective instruction that the student population is more sensitive to. Others will not. Some will make a kind of devil’s bargain, lowering expectations because it seems like the only way forward, but end up in a spiral of lower rigor that erodes the education of even the students who have more cognitive resources.
When I say some students are less sensitive to instruction, that’s a bit of a simplification. Students with more cognitive resources can still learn from mediocre instruction, but they aren’t likely to learn much if they aren’t taught at all. If a student doesn’t take algebra, they aren’t likely to learn algebra. Here the devil’s bargain starts to hurt stronger students: the school lowers expectations, and all students suffer because they aren’t exposed to the same level of rigor.5
Where Does This Leave Us?
There are no easy conclusions from this mental model.
In this story, even if the teachers and staff aren’t any better or worse in one school or another, you end up with very different outcomes. From my perspective, the basic state of education in the US is that we know how to teach students with more cognitive resources, but classrooms where students with fewer cognitive resources consistently succeed are few and far between. They do exist, in pockets here and there, but it’s far rarer than it should be.
There’s also a problem in finding these pockets: it’s easy to look at a school with relatively high achievement and assume they must be doing something right. In most cases, that high achievement is a result of demographics and not teaching quality. Pretty much any form of teaching will be successful with some students. These facts make it hard to improve education, because a practice can seem effective on the surface while not actually working for a significant fraction of students.
One tough consequence of this mental model is that schools that are successful working with a high-poverty student population look different. There is more structure. The teaching is intense. This makes some people squirm; they want to help students from challenging circumstances succeed, but they also want that teaching to look like teaching in affluent areas.
This model is about distributions and averages. Teachers work with individuals. It’s easy to caricature rich kids as having everything easy, but that’s not the case. Plenty of well-off students struggle and don’t get the instruction they need. Similarly, one can fall into the trap of stereotyping poor kids, but there is a ton of variation and plenty of poor students excel in school. Aggregated together the differences become clear, but we have to be careful not to paint with too broad a brush.
The real answer, to me, is teaching. Doing our best to give every kid great instruction, instruction that manages cognitive load and gives access to as many students as possible.
Effective Teaching Is Harder Than It Sounds
This is where I find cognitive science so appealing. The basic science of teaching students with fewer cognitive resources is to understand the simple model of the mind. The goal is the same for all students: to move knowledge into long-term memory. Managing cognitive load, making sure precious cognitive resources are focused on the essential ideas we want students to learn, is the heart of teaching so that more students can learn.
The basic steps sound simple: practice, small chunks, manage attention, connections, sequencing, model, review, scaffolds.
But that’s harder than it sounds. The hard part is that all of those things depend on the content that is being taught. Breaking ideas into small steps is very different when teaching reading fluency than when teaching algebra. Providing scaffolds is different for writing a paragraph than adding fractions. Generally speaking, teachers have some level of expertise in what they teach. That expertise can blind us to the best ways to break concepts down. The small pieces have become invisible, the connections between concepts so automatic that we cannot perceive them.
This type of teaching is basically an unsolved problem across education. Lots of experienced teachers know how to do this for random stuff they’ve taught for years, but for most domains taught in K-12 schools, there isn’t a well-accepted way of managing cognitive load to teach that topic. I can’t google, “I’m teaching adding and subtracting decimals and I want to know how to teach it while managing cognitive load.” There’s no easy answer, no expert consensus. There are lots of ideas out there, but unfortunately many of them are nonsense.6
It might seem crazy that this is an unsolved problem after all of the time and money that has been invested in improving education. I would argue that much of that time and money has paid little attention to the basic phenomena I’m describing. I’m an optimist: I’m optimistic that somewhere down the road we will know more about how to build an education system where more students succeed. I think the path is a better understanding of how humans learn, how that cognitive architecture suggests certain teaching practices, and then the slow, empirical work of putting that knowledge into practice with all of the content we want students to learn in school.
1
For a few examples of where students rise above their circumstances: Steubenville, Ohio, several southern US states, and England. It’s also worth noting that child poverty is down in the US, yet that hasn’t translated into learning gains.
2
I’m emphasizing working memory capacity and processing speed as the heart of cognitive resources because there is evidence to support these two constructs and they are relatively simple to conceptualize as part of learning. Here is a good paper to explore the topic, though there’s plenty of disagreement among researchers in this area. The paper also looks at a third major area, what they call executive processes, which often have to do with attention. Other researchers would mention different areas. I would emphasize here that first, this is a mental model. It’s not perfect, but I do still think it’s useful. And second, when I read about these types of cognitive processes, they largely have to do with the bottlenecks between the learning environment and long-term memory, and I think my basic thesis holds even under a different conceptualization of cognitive resources.
3
There’s an interesting question here that I don’t want to sidetrack the main post but is worth looking at. I’m claiming that students with fewer cognitive resources need better teaching, but that they can, generally speaking, achieve similar end goals to other students. Cognitive resources don’t limit a student’s potential, but they constrain the type of teaching needed to get them there.
When I’ve made this claim in the past it really bugs some people. I want to address a few possible responses. First, some people will reference students with profound intellectual disabilities. Sure, though that number is very small and can become an excuse for students who are still capable. One of my education heroes is Samuel Kirk. Kirk was a pioneer in special education who gained a reputation for working with students labeled “uneducable” and teaching them to read. Are there exceptions? Absolutely. But let’s be careful with the labels we throw around.
Another response I get is the visceral experience from some people, often in math, that they learned up to a certain point and then hit a wall they couldn’t surpass. My hypothesis here is pretty simple: in that situation, the instruction was not sufficient to keep you moving forward. With better instruction, you could have learned more. Does this mean every student has the potential to be the next Einstein? No. First, graduate-level physics programs aren’t exactly paragons of excellent pedagogy. In any intellectual endeavor, the quality of instruction decreases as you climb the ladder. That will always create a barrier. I’m also arguing in this post that the type of instruction I’m talking about is poorly-defined in most domains. More often than not we just don’t know how to do it yet, and it is much harder at the higher levels of a domain where content is more complex. So no, I’m not arguing that anyone can do anything they want. There are limitations. But generally speaking, I think our ceilings are higher than we think, and excellent instruction has remarkable power to improve outcomes.
A related point here is that, if we improve education outcomes, won’t we raise standards and the cycle will continue indefinitely? Maybe! Look, in the meantime, I’m teaching 5th graders who struggle to read fluently and to figure out 4 x 6. Let’s nail the core skills that schools are charged with teaching, and then we can have all sorts of debates about what to do next.
4
This list of effective teaching practices is heavily influenced by Rosenshine’s principles.
5
This mental model can lead to a few different strands of reasoning around improving schools. One can focus on motivation and try to find as many ways as possible to get students more motivated. One can focus on outside instruction — one example of this is advocating for families to read with their children. One can focus on increasing rigor, ensuring all students are exposed to rigorous content to mitigate the spiral of low expectations and help stronger students learn through exposure. All of these work, to an extent. And I’m in favor of all of these, to an extent. Building a school culture to maximize motivation is valuable. Families should read to their kids. Schools shouldn’t lower expectations. But all of these can also become distractions from high-quality instruction. We tell students to advocate for themselves and persevere rather than improving instruction so less perseverance is necessary. We start a read-at-home initiative to cover for the failures of early reading instruction. We raise the level of rigor without providing corresponding support. From my perspective, instruction should be the first priority, and everything else should play a supporting role.
6
One reason I find the science of reading compelling is that it’s the one area of education where we do have some decent evidence for ways to reduce cognitive load, in this case for reducing cognitive load while teaching students to decode text. We also have some suggestive work on other areas of the reading process, though that understanding is far from complete. I find this area of research interesting because it is so different from many other areas of education, where research is a long way from recommending specific, concrete classroom practices.
While Google is trying to move beyond Chromebooks, the ubiquity of these devices has made the company's education products huge in schools. Now, Google says it will let educators choose whether to allow its AI tools in the classroom. However, blocking AI tools also means giving up search, leaving schools with a difficult decision.
Google's new overarching approach to AI in K-12 classrooms sounds good on the surface. After consulting educators, researchers, and lawmakers, Google has decided that schools using its education products will be opted out of AI tools by default. It will be up to administrators to manually enable them when these changes roll out in the coming months.
When you look at what Google considers to be an AI tool, the nature of the Faustian bargain becomes clear. The affected products are Gemini, Gemini Notebook, Gemini in Classroom, and Google Search. There's no picking and choosing—you either allow all the AI or none of it.
Arkivism is a search and visual discovery engine for millions of public domain works from the collections of the Smithsonian, The Met, NASA, Library of Congress, National Gallery of Art, Rijksmuseum, and more.
Next Page of Stories