After months of speculation and projections, the official National Merit Semifinalist cutoffs for the class of 2027 were released September 9th. On average, these are the lowest cutoffs in National Merit history, with all 50 state cutoffs sitting lower than they did last year. Most cutoffs went down at least 3 points from the class of 2026’s figures. Alaska set the pace with a 7-point decline. New Jersey again finished with the highest mark, dropping from 225 to 223. Compass will explore:
- How the class of 2027 results stack up historically and how cutoffs are set
- The lower number of Semifinalists nationally
- Why cutoffs declined so precipitously. Hint: It’s not the students.
- The implications of score instability for students, schools, and states
- What the class of 2028 might expect after they take the PSAT/NMSQT in October 2026 and on what timeline
The junior-year PSAT is also the National Merit Scholarship Qualifying Test (NMSQT). Each year, National Merit recognizes the top 55,000 test takers. To be included in that group, students must be among the top 4% of the 1.3 million eligible students who take the PSAT. Of that elite group, approximately 16,500 are recognized as Semifinalists for having the highest scores within their states. Semifinalists have the opportunity to continue in the program to become Finalists and, potentially, scholarship recipients. Some colleges provide significant scholarships and even “full-rides” for Finalists. The remaining 39,000 students are recognized as Commended Students. More information can be found in our National Merit Explained post.
National Merit uses the Selection Index (SI), found on the PSAT score report, to rank students. The Selection Index is calculated by doubling the Reading & Writing score (RW), adding the Math score (M), and dividing that sum by 10. For example, a student with a 720 RW and 700 M score has a Selection Index of 214: (720 x 2 + 700) ÷ 10. A student with a 700 RW and 720 M has a Selection Index of 212. The highest SI a student can receive is 228 (760 RW and 760 M). The weighting of the index means that not all students with the same Total Score have the same SI. “Will a 1450 qualify?” is not the right question. Instead, look at whether a 750 RW and 700 M will qualify.
Summarizing the changes seen for the class of 2027
In October 2015 the PSAT moved to a 320-1520 scale, which was set to provide vertical alignment with the SAT’s 400-1600 scale. The class of 2017 was the first group of students taking the redesigned PSAT/NMSQT, so the fairest comparisons are in the last eleven class years.
This year saw the average Semifinalist cutoff hit a record low. It provided something of a mirror image to the class of 2026. Last year, 11 states set new record highs in the modern PSAT era, and 10 more tied their records. This year, 5 states established new lows, and 18 tied their record lows. Kansas and Minnesota both dropped 2 points below their previous lows, an unusual event for mid-sized states. Even more unusual was what happened in the largest dozen states. None of those has ever seen a drop of more than 3 points, but this year saw California, Illinois, and Michigan fall 4 points.
Results this year were closer to those in the class of 2024 than they were to the previous two years’ results. Recency is not always a predictor of future results. The Commended level, which is set nationally, declined 2 points this year to 208. The Semifinalist cutoffs, in contrast, fell 3 points.
Entry requirements for National Merit versus qualifying for National Merit.
Your PSAT/NMSQT score report tells you whether you meet the eligibility requirements for the NMSP. In general, juniors taking the October PSAT are eligible. If you have an asterisk next to your Selection Index, it means that your answers to the entrance questions have made you ineligible. Your answers are conveniently noted on your score report. If you think there is an error, you will also find instructions on how to contact NMSC. Meeting the eligibility requirements simply means that your score will be considered. Approximately 1.3 million students enter the competition each year. Only about 55,000 students will be named as Commended Students, Semifinalists, Finalists, or Scholars. See National Merit Explained for more information.
Why did the total number of Semifinalists decline?
After recognizing more than 17,000 Semifinalists in the prior two years, the class of 2027 saw only 16,230, a decline of 5%. The decline in Semifinalists is a separate issue from lower cutoffs. Setting cutoffs can be a blunt tool. In a large state, there may be more than 200-300 students at a single score. Small states might have multiple scores with zero students. NMSC can’t always hit its targets on the nose. Below are the largest swings in both directions.
How Semifinalists are allocated among states
It’s possible that NMSC recalibrated the total target number of Semifinalists this year. It helps to understand how National Merit sets cutoffs. NMSC allocates a target number of Semifinalists to states based on their population of high school students, not all of whom take the PSAT. If Texas is allocated 1,500 Semifinalists based on its population of high school graduates, then NMSC works down from a perfect 228 Selection Index until it gets as close as possible to that target. In the class of 2027, 1,462 students scored at or above the cutoff of 219. A cutoff of 220 would have produced too few Semifinalists. A cutoff of 218 would have gone over the allocation. As scores go lower, more and more students occupy each level, and where the crowd forms does not stay consistent from year to year. It is easy for cutoffs to move up or down a point even when there is minimal change in testing behavior or performance. All students at or above a cutoff qualify.
The changes in the above chart fit within the historical pattern, so students should expect similar swings to happen in the future. New Jersey had about the same number of Semifinalists in 2023 as it saw this year, and actually fell from a greater height compared to 2022. Similarly, Texas saw fewer than 1,462 Semifinalists in classes of 2020, 2021, and 2022. And there is little chance that NMSC calculated that Texas’ high school population was falling faster than, say, California’s. We won’t know until next September whether the 16,230 Semifinalists was a fluke or a new standard.
The Mystery of the Declining PSAT Scores
When you have eliminated the impossible, whatever remains, however improbable, must be the truth. National Merit cutoffs may not have the makings of a Sherlock Holmes novel. The guilty party is too easy to identify and the deduction not all that improbable. To borrow from another mystery trope: it’s College Board in the testing room with a faulty scale.
When test scores change from year-to-year, College Board issues press releases discussing testing patterns and student achievement. What they never include is “That one’s on us.” When it comes to National Merit scores, however, it often is.
A history of change
Over the last seventeen years, there have been six classes that saw cutoffs change in more than forty states. The heat strip below shows how cutoffs shift from year to year.
Even if we eliminate “bounce back” years where scores come back into line after an unusual shift, a major disruption occurs approximately every 4 years. Interestingly, none of those scoring disruptions can be attached to the biggest behavioral disruption of all — the Covid cancellations that drove PSAT/NMSQT volumes below one million for the first time in decades. In addition to last year’s exam, the PSATs from 2011 (class of 2013), 2016 (class of 2018), and 2019 (class of 2021) stand out as problematic. The anomalous 2019 results could be traced back to a particularly mis-scaled form, which I wrote about at the time. The class of 2014 also saw significant changes, but those were more of a bounce-back from the previous year, similar to what we are seeing with the class of 2027.
When Semifinalist cutoffs reached vertiginous levels last year, several theories made the rounds:
- (A) Covid-related learning losses were vanquished
- (B) The integrated Desmos calculator on the digital PSAT had “broken” the Math section
- (C) Test preparation — and maybe AI-enabled test preparation — had made the test too easy
- (D) The test taking population shifted
Compass disputed each of these theories, but it wasn’t until this year’s free-fall that we could put each of them fully to rest. A key piece of evidence is which scores rose and fell.
Over the most recent three PSATs, the highest reported Math scores (College Board groups all 700-760 scores together) did go up on the October 2024 PSAT, but they settled back down on the October 2025 PSAT. If improving Math skills were the key driver, why did they not keep progressing upward? The fall PSAT also provides an inherent double-check, because sophomores and juniors both take the exam in similar numbers. Sophomores and juniors consistently tracked each other across the different exams. The class of 2025 and 2026 overlapped, then 2026 and 2027, and then 2027 and 2028. Every PSAT the movement has been similar. Yet sophomores do not do test preparation in sufficient force to shape score results. The sets of classes were in different points of math development during remote learning, and yet they showed the same rises and falls. Learning loss and test preparation do not provide coherent answers.
Reading & Writing scores were the true driver of the class of 2026 records and the 2027 declines
High RW scores are typically less common than high Math scores. But on the October 2024 PSAT, almost as many students scored 700-760 on RW as on Math. Top scores went up by 18%. The double-weight that RW scores provide in the Selection Index pushed cutoffs to new highs. The 210 Commended ranked only fourth when compared to the previous decade. Earlier class years saw cutoffs of 211 and 212 (twice). So why did last year’s cutoffs rise to levels that had seemed impossible? The highest cutoffs had never exceeded 223 (Maryland’s 224 in the class of 2022 was due to every PSAT in the state being cancelled and students qualifying via SAT scores). How did New Jersey and Massachusetts reach 225 and California, Maryland, Virginia, and Washington end up at 224? The best explanation is that the very top RW scores — 740-760 — increased at an even higher rate than 700-730s. A 225 cutoff is not possible without at least a 750 RW. A 224 requires at least a 740 RW. Students were acing the test in record numbers.
The plummet for the class of 2027
On the 2025 PSAT, the number of high RW scores dropped by 27%. In fact, the RW count for the class of 2027 is even lower than it was two years ago. The surfeit of 700+ scores dried up. Obtaining a 740-760 grew harder. As with Math scores, sophomore performance changes matched that of juniors.
Process of elimination
The rapid gains and collapse eliminate (A), (B), and (C). As sophomores, the class of 2027 was phenomenal at reading according to the PSAT. Post-Covid learning changes cannot explain why they suddenly struggled as juniors. Option (B) doesn’t fit, because Math scores were not the source of the changes. Why would test preparation have stopped working (C), and how can it explain that sophomores and juniors were similarly impacted across the years? This leaves (D) as a remote contender. Except that it fails to explain the 50-state reach of the changes and the magnitude of the changes even in the largest states.
Thirteen of the largest fifteen states dropped by 3 points or more. Pennsylvania and New Jersey saw 2-point declines. If the smaller states were pulling down the class of 2027’s average cutoffs, we might at least attribute some of the disruption to the instability of cutoffs when there are fewer than 100 Semifinalists. That’s not what happened. And if only a few large states saw inflows or outflows of top test takers, the impact would have been muted. Nor did overall test taker numbers change much. When the entire set of large states has this magnitude of a change, the test scale itself has to be at the heart of it.
What does a 3-point change represent and why is it so bad?
For simplicity, the change can be characterized as a 15-point miss in RW on the 160-760 scale (200-800 on the SAT).
- The typical annual growth in RW scores is about 40-50 points, so the scale change represents approximately one-third of an academic year.
- Systematic scaling error is additive to individual student fluctuations in scores — typically around +/-30 points on each side of the PSAT and SAT. Scaling error makes it harder for students to determine where they will score and how their score stacks up.
- Scaling error produces incorrect percentile ranks, because percentiles are based on the 3 prior years of results. Sudden inflation or deflation gives an inaccurate read.
- The PSAT is used for purposes beyond National Merit such as AP placement or state-mandated testing. Results may not align with prior years or academic growth expectations.
- When error happens primarily on one side of the PSAT, some students are unfairly disadvantaged.
- In a competition such as National Merit where relative standing matters, a uniform increase or decrease might be tolerable. A rising tide lifts all students (sort of). Except that there is no way of knowing whether the error applied equally to all students. The digital PSAT is not a single form or even a limited set of forms. Exams across the PSAT testing window can vary, and even on the same test dates, students often see differing questions. It’s one of the security advantages provided by digital testing, but it can also be a curse. College Board provides no reporting on the items or scaling, so students (and Compass) have no way of assessing the impact on individuals. Schools have no way of knowing if their students were hit harder than at other schools.
- This keeps happening. Is College Board having difficulty keeping the digital PSAT scale consistent or will things stay more consistent over time? The historical record says that they will not.
Do these sorts of systematic scaling errors occur on the SAT?
They do. They are harder to spot, but likely even more common. Noted above is that the largest disruptions on the PSAT/NMSQT occur, on average, every 4 years. The SAT is given on (a) 8 national dates every year, (b) during a 2-month SAT School Day window in the spring, and (c) during a 1-month window in the fall. This accelerates the chance for errors. It creates a near certainty that some students will encounter systematic error in their favor and others will see it working against them.
The PSAT and the National Merit Scholarship Program provide a unique laboratory. The PSAT/NMSQT represents one slice in time. The population of PSAT takers is relatively stable from year to year, and the sophomore and junior PSATs provide a sort of longitudinal double-check. The 50 state cutoffs and their varying levels give information about the extent of scaling errors. Comparing ever-shifting SAT dates is difficult, especially because College Board restricts access to the data. It is unfortunate that the primary protection students have is taking the SAT multiple times. In combination with superscoring, it represents the best way for students to lock-in positive error.
How do errors happen? Shouldn’t scores stay consistent?
When new tests are designed, not every problem can be tested on previous populations of students. Instead, a sampling of problems is tested with a sampling of students, perhaps 4-5,000. In theory, that’s a significant enough size that random error will make little difference. The sample group should be comparable to the entire population. Random error is an acceptable part of test creation, because it largely cancels out at scale.
But what if the error is not random? What if the test design is flawed? For example, if the methodology continually overestimates the ability of the sample groups, future questions will be predicted to be harder than they really are. And then over-prediction may happen with the next group. This systematic bias can creep into a test production and equating process. Normally it is spotted. As we see from the October 2024 PSAT and similar examples, sometimes it is not. The National Merit program provides a challenge to test creators, because everything important happens in a narrow part of the scale. On the digital PSAT, only 50 questions count toward a student’s score. Differentiating between a 720 student, for example, and a 730 student, and a 740 student rests on small differences in performance. If College Board has the statistics even slightly off, scores will not align well across tests. We saw the occasional wild ride on the paper PSAT, but the rides may get wilder and more frequent on the digital PSAT.
Predictions and timeline for the class of 2028 and beyond
Compass has always warned students not to be too fixated on a single cutoff target. The crazy events of the last two years have made it obvious that cutoffs can move in large and unexpected ways. It’s better to think in terms of score ranges rather than a pass/fail score. Prior to the PSAT, we can only estimate based on historical figures, without knowing what kind of surprises College Board has in store. After the PSAT, Compass analyzes the national results in the broad buckets that College Board provides. This usually gives a clear idea of which way the overall trend is moving. We should see how RW and Math scores behave, what the 1400-1520 bucket looks like, and develop a tight estimate of the Commended cutoff. For example, here is the estimate from last year.
Unfortunately, it does not provide insight into individual states. In April, the Commended level usually leaks, and we get confirmation of the national trend predictions. It’s not until Semifinalist notifications start being handed out in late August that state-by-state information starts coming into focus.
Below is a pre-PSAT summary. Find the range of potential cutoffs in your state, and how it has stacked up in recent class years. Below the table is more information about the National Merit program. You may also want to consult our post on National Merit Explained, which goes more into the steps of the process, what different honors mean, and what sort of scholarships are available.
State | Class of 2028 (Est Range) | Class of 2027 (Actual) | Class of 2026 (Actual) | Class of 2025 (Actual) | Class of 2024 (Actual) | Avg SF Count |
|---|---|---|---|---|---|---|
| Alabama | 209-216 | 210 | 214 | 212 | 210 | 239 |
| Alaska | 207-217 | 208 | 215 | 214 | 209 | 36 |
| Arizona | 214-220 | 215 | 218 | 217 | 216 | 359 |
| Arkansas | 209-215 | 210 | 215 | 213 | 210 | 145 |
| California | 219-224 | 220 | 224 | 221 | 221 | 2,125 |
| Colorado | 216-221 | 217 | 219 | 218 | 216 | 280 |
| Connecticut | 219-223 | 220 | 223 | 221 | 221 | 182 |
| Delaware | 218-222 | 219 | 220 | 219 | 219 | 43 |
| Florida | 215-219 | 216 | 219 | 217 | 216 | 959 |
| Georgia | 216-220 | 217 | 220 | 218 | 217 | 554 |
| Hawaii | 215-220 | 218 | 219 | 217 | 217 | 62 |
| Idaho | 210-216 | 211 | 215 | 213 | 211 | 90 |
| Illinois | 217-222 | 218 | 222 | 220 | 219 | 693 |
| Indiana | 214-219 | 215 | 218 | 217 | 216 | 320 |
| Iowa | 209-216 | 210 | 214 | 212 | 210 | 152 |
| Kansas | 211-219 | 212 | 216 | 215 | 214 | 147 |
| Kentucky | 210-218 | 211 | 214 | 213 | 211 | 208 |
| Louisiana | 210-217 | 211 | 216 | 214 | 214 | 227 |
| Maine | 211-217 | 214 | 217 | 214 | 213 | 61 |
| Maryland | 220-224 | 221 | 224 | 222 | 221 | 313 |
| Massachusetts | 220-225 | 221 | 225 | 223 | 222 | 322 |
| Michigan | 215-220 | 216 | 220 | 218 | 217 | 509 |
| Minnesota | 213-220 | 214 | 219 | 217 | 216 | 284 |
| Mississippi | 209-215 | 211 | 213 | 212 | 209 | 149 |
| Missouri | 213-217 | 214 | 217 | 215 | 214 | 299 |
| Montana | 207-214 | 208 | 213 | 209 | 209 | 46 |
| Nebraska | 210-216 | 212 | 214 | 211 | 210 | 105 |
| Nevada | 210-218 | 211 | 214 | 214 | 211 | 148 |
| New Hampshire | 213-219 | 215 | 219 | 217 | 215 | 64 |
| New Jersey | 222-225 | 223 | 225 | 223 | 223 | 486 |
| New Mexico | 207-215 | 209 | 210 | 211 | 207 | 99 |
| New York | 219-223 | 220 | 223 | 220 | 220 | 998 |
| North Carolina | 216-220 | 217 | 220 | 218 | 217 | 502 |
| North Dakota | 207-212 | 208 | 210 | 210 | 207 | 28 |
| Ohio | 215-219 | 216 | 219 | 217 | 216 | 588 |
| Oklahoma | 208-216 | 210 | 212 | 211 | 208 | 195 |
| Oregon | 215-221 | 216 | 219 | 216 | 216 | 185 |
| Pennsylvania | 217-221 | 219 | 221 | 219 | 219 | 623 |
| Rhode Island | 212-220 | 213 | 219 | 217 | 215 | 50 |
| South Carolina | 209-217 | 212 | 215 | 214 | 209 | 225 |
| South Dakota | 208-215 | 210 | 211 | 208 | 209 | 39 |
| Tennessee | 213-219 | 214 | 219 | 217 | 217 | 324 |
| Texas | 218-222 | 219 | 222 | 219 | 219 | 1,508 |
| Utah | 208-216 | 209 | 213 | 211 | 209 | 181 |
| Vermont | 211-217 | 212 | 216 | 215 | 212 | 30 |
| Virginia | 219-224 | 221 | 224 | 222 | 219 | 412 |
| Washington | 220-224 | 221 | 224 | 222 | 220 | 353 |
| West Virginia | 207-212 | 208 | 210 | 209 | 207 | 64 |
| Wisconsin | 211-217 | 212 | 215 | 214 | 213 | 295 |
| Wyoming | 207-213 | 208 | 210 | 209 | 207 | 21 |
| District of Columbia | 222-225 | 223 | 225 | 223 | 223 | 36 |
| Studying Abroad | 222-225 | 223 | 225 | 223 | 223 | 87 |
| U.S. Territories | 207-212 | 208 | 210 | 208 | 207 | 39 |
| Commended | 207-212 | 208 | 210 | 208 | 207 |
How cutoffs are determined
Qualifying scores (“cutoffs”) are not based on the total score for the PSAT (320-1520) but on the Selection Index, which is calculated by doubling the RW score, adding the Math score, and then dividing the sum by 10. The maximum Selection Index is 228. Students can find a historical set of cutoff data here or see how Semifinalist and Commended counts have changed state by state. Students still ahead of the October PSAT can prepare with small group PSAT classes built around the Selection Index.
Why does each state have its own Semifinalist cutoff if the program is NATIONAL Merit?
This is always a hot button question. NMSC allocates the approximately 16,500 Semifinalists among states based on the number of high school graduates. That way, students across the nation are represented. It also means that there are very different qualifying standards from state to state. A Massachusetts student with a 220 might miss out on being a Semifinalist. If she lived 10 miles away in New Hampshire, she would qualify.
No Semifinalist cutoff can be lower than the national Commended level. Cutoffs for the District of Columbia and for U.S. students studying abroad are set at the highest state cutoff (typically New Jersey). The cutoff for students in U.S. territories and possessions falls at the Commended level each year. Boarding schools are grouped by region. The cutoff for a given region is the highest state cutoff within the region.
When are National Merit Semifinalists announced?
The lists of Semifinalists will not be distributed to high schools until the end of August 2027. With the exception of homeschoolers, students do not receive direct notification. NMSC asks that schools not share the results publicly until the end of the press embargo in mid-September, but schools are allowed to notify students privately before that date. NMSC does not send Commended Student letters to high schools until mid-September. Compass will keep students updated on developments as the dates approach.
Do state and national percentiles indicate whether a student will be a National Merit Semifinalist?
No! Approximately 1% of test takers qualify as Semifinalists each year, so it is tempting to view a 99th percentile score as indicating a high enough score — especially now that College Board provides students with percentiles by state. There are any number of flaws that rule out using percentiles as a quick way of determining National Merit status.
- Percentiles are based on section scores or total score, not Selection Index
- Percentiles are rounded. There is a large difference, from a National Merit perspective, between the top 0.51% and the top 1.49%
- Percentiles reveal the percentage of students at or below a certain score, but the “at” part is important when NMSC is determining cutoffs.
- The number of Semifinalists is based on the number of high school graduates in a state, not the number of PSAT takers. Percentiles are based on PSAT takers. States have widely varying participation rates.
- Most definitive of all: Percentiles do not reflect the current year’s scores! They are based on the prior 3 years’ performance. They are set even before the test is given. And if you are going to use prior history, why not use the record of prior National Merit cutoffs rather than the highly suspect percentiles?
Entry requirements for National Merit versus qualifying for National Merit.
Your PSAT/NMSQT score report tells you whether you meet the eligibility requirements for the NMSP. In general, juniors taking the October PSAT are eligible. If you have an asterisk next to your Selection Index, it means that your answers to the entrance questions have made you ineligible. Your answers are conveniently noted on your score report. If you think there is an error, you will also find instructions on how to contact NMSC. Meeting the eligibility requirements simply means that your score will be considered. Approximately 1.3 million students enter the competition each year. Only about 55,000 students will be named as Commended Students, Semifinalists, Finalists, or Scholars. See National Merit Explained for more information.
Curious about your thoughts on alternate entry? A kid who was not on the official state list but was called down and given a letter along with my kid and the others said he did alternate entry using his SAT after taking it EIGHT times to get to a high enough score to submit. Eight. That just feels wrong to me. Why does the NMC allow this? There has to be a way to offer alternate entry without allowing exploitation.
Momof4,
Congratulations to your student!
First, I should mention that the scores of alternate entrants do not impact the cutoff and do not — at least historically — count against a state’s allocation of Semifinalists. In other words, they don’t squeeze out any PSAT takers (things tip a bit when you consider that there is additional competition at the Finalist stage).
You’re correct, though, that there is still an issue of fairness. One problem NMSC faces right off the bat is that College Board allows Score Choice when reporting scores to colleges or scholarship programs. NMSC could, I suppose, have some strong language about receiving all or particular scores, but there isn’t really an enforcement mechanism. The other problem is that any fix creates its own problems. Students take the SAT in very different patterns. Let’s say a particular student takes the SAT early such as March of sophomore year and then again in October of junior year and May of junior year. Which score should be used? It might seem like the closest to the PSAT. What if the student was not able to take the PSAT because they were ill throughout October and underperformed on the SAT for that reason? What about a similar student who took the SAT only in March of sophomore year and May of junior year. Which test is used? Eight tests is clearly an extreme. What is too many? Five? Three? I suppose NMSC could say, “Every student that wants to enter via alternate entry must take the March SAT.” I just don’t see an appetite for change, and NMSC prefers to consider students in a favorable light — as does pretty much every college, and admission is arguably a bigger deal than National Merit.
There have been changes to alternate entry this year, and I believe these changes were intended to minimize abuse. No AE applications were approved until after October, and likewise, no SAT scores from October 2025 or prior will count. This prevents students from having a score that they know will qualify them for NMSF and just sitting on the sidelines. The system is not perfect, but I think that NMSC is trying their best to make it fair for everyone.
Kat,
Thank you for the update! That’s very useful information.
Hi Art,
Thank you for putting together these very informative guides. I’ve seen you mention the cutoffs become known (unofficially) in April. Is this for both Commended + Semifinalist? Where can one find these around that time?
Thanks a lot!
Sam
Sam,
Only the Commended cutoff becomes known in April.
“Where can one find these around that time?” Right here, of course!
Great, thanks!
Has the commended cutoff been officially announced as of 10/16/2025?
Robert,
If you mean the class of 2026 cutoff (2024 PSAT), then it became official in September. It’s 210 and can be found in this publication.
If you are referring this fall’s PSAT, then you’ve got a long wait. Once scores are out for all tests, Compass will estimate the cutoff. In April, we’ll get unofficial word on the cutoff (unofficial only because it is not publicly released). In September 2026, NMSC will publish it.
My biggest qualm with alternate entry is how it affects those performing at the top percentile or even fraction of a percentile. While PSAT and SAT scores generally correlate very well at the middle of the testing range, this falls apart at the top end of the spectrum – at some point, how many points you lose on the PSAT becomes a more representative measure of how many points you will lose on the SAT. For example, anecdotally, I scored a 1500 (740R, 760M) on my PSAT and a 1570 (770R, 800M) on my SAT. In both of these scenarios, I likely got 1-2 reading and writing questions wrong and performed at the same level of ability. However, with my PSAT score, I failed to qualify in NJ. On the other hand, my SAT, taken shortly before my PSAT, would have been significantly higher than necessary to achieve the maximum index score of 228, guaranteeing semifinalist status. From my personal perspective, this is especially apparent at a competitive public high school (we had around 20 semifinalists) where many savvy test takers were aware of this and decided to cancel scores or just submit an “excuse” for alternate entry. I believe requiring alternate entry applicants to sit for a PSAT instead of an SAT would remedy this fault, but I am sure there are logistical issues to doing so. Just my two cents 🙂
I don’t disagree. At those extreme score levels, you are at the mercy of random error, and that’s not a great basis upon which to build a scholarship program. Another option would be to allow all students to use the PSAT, SAT, or ACT as an entry point. I don’t think either solution will come to pass.
Sadly, it seems to have become a strategy in my area for experienced parents. Having your kid be “sick” on PSAT day and then take the SAT as many times as possible between Oct-March to submit by the April 1 date has become somewhat of a norm for those in the know. In Florida, being NMF means a full ride state scholarship or over $100k if you agree to attend an in state school. My opinion is only one single SAT sitting after missing the PSAT should be allowed within say 45-60 days so that all students are on equal footing. Further SATs can of course be taken to increase college admission chances, but only the first one should be accepted by NMSC.
Hi Art, my daughter just got her 222 for class of 2027 in CA. Are we thinking CA will maintain it’s elevated level or any chance of it dropping 2 points for next year. Thanks as always!
Christie,
College Board is releasing scores over 3 tranches, so it will be a couple more weeks before we see how students did overall. We won’t know what is happening at the state level, but at least we should be able to get a read on the number of high scores nationally. There are more years than not where 222 would have been sufficient in CA, so let’s hope for a return to normalcy!
Just an observation. Student was getting 1500s on bluebook SAT practice tests and ended up with a low 1300s PSAT. Wondering if anyone else is experiencing something like this. I suspect this may be due to not reaching the more difficult section 2.
Barry,
The Bluebook tests are a bit suspect, but it’s also true that crazy things can happen on test day. While it’s possible that the student was routed to the easier section 2, that’s not usually the case for students who end up in the 1300s.
My junior got perfect scores on all their bluebook tests and scored a 1470 in the actual PSAT (way too low in our state to qualify.) My college-age child’s practice scores were more similar to their actual score (they got a 1500 and did qualify their year.) The Bluebook practice tests didn’t feel similar to my current junior at all. She said the math was similar (she got a 760 still,) but the verbal section was much much harder for her on the actual exam.
Art
Was wondering if you have a target date in November 2025 for releasing your first predictions on the Oct 2025 PSAT commended and semi-finalist scores. Son has a 1490/222 in GA so we are feeling fairly optimistic. It would take a big jump for the GA cut-off to go to 223 as it has not been above 220 historically. I am assuming the cut-off score is included (220 and above versus only above 220) when the numbers are posted.
Thanks
OD,
Well, the target date that I actually hit was today! I’ll have state-by-state analysis on the 17th, but the national numbers are up and they point downward. Hard. Your son looks good at 222. Yes, all students scoring at or above the official cutoff qualify as Semifinalists.
Hello Art,
Do we know if we see different scores based on the environment where the test was taken (in school during school day vs. elsewhere)? My daughter just got her score back (219 index for wi) and it was the same as her Sophomore score. The day she finished the test she complained about passing bells being on and a kid was screaming in the hall during the hard math section for minutes. She did not finished the last few math questions because of this.
Thanks,
Will
Will,
I’m sorry that your daughter had that experience. Testing is stressful enough without getting hit with unwelcome challenges. I have not seen any studies on performance versus test setting. With a 219 in Wisconsin, I think your daughter is in great shape despite the challenging environment!
I appreciate the stats based observations you provide. My son is a junior in PA with a 222 index so watching closely since Class of 2026 cutoff here was 221. When do you foresee updating your chart to include the first half of reported scores to show percentages and also the full set? These are definitely “anything can happen” times and states so as the parent of a kid possibly on the bubble I’m following closely. Thank you.
PAmom,
I’ve finally had a chance to do a first pass at the numbers (I decided against analyzing the partial batches), and you can find the draft analysis on the site. While I haven’t published my state-by-state ranges, yet, I can safely say that almost all of my “most likely” scores will be lower than last year’s. Basically, the exact factors that pushed things so high last year have shifted into reverse this year. I like your son’s odds a lot more now that I’ve had a chance to sink my teeth into the national numbers.
Just heard. In at 222 in PA.
Congratulations to your student! Thank you for reporting.
Hello Art,
Some state flagships (such as University of Wisconsin-Madison with the Wisconsin Guarantee program) have recently created a policy where being a national merit finalist is among the methods to receive auto entry into the university (as well as being in the top 5% of your class etc). Have you seen these policies increase the number of test takers in these states, and would you expect these polices to change cutoff scores?
Thanks,
Tom
Tom,
These programs can be excellent pathways to admission for students who perform well in school or on tests, but I haven’t seen them move the needle on National Merit cutoffs. To notice any impact, it usually requires a state mandate to take the PSAT. I’m going to go out on a limb and guess that you are in Wisconsin, so that’s what I’ll reference.
It’s not easy to change the testing behavior of NMSF-capable students, especially since most students take the PSAT only because their high school requires it (or at least provides it). It is a daunting process to take the PSAT as a free agent. There are already a number of strong reasons for students to want to be a National Merit Finalist — scholarships and admission at dozens (hundreds?) of schools stand out. Each year sees some colleges expand NM outreach and others retrench. Would the Wisconsin Guarantee make enough of a difference on its own? Also, becoming a Finalist requires “a record of consistently very high academic performance in all of grades 9 through 12.” That sounds a lot like a student who is already in good shape for U-W admission either through the 5% standard or an old-fashioned application. The Wisconsin Guarantee was announced in March 2024, so I was curious to see if it had an impact on the Oct 2024 PSAT. No, it did not. The number of PSAT-takers actually dropped from 8,375 to 7,678. As you probably know, WI is an ACT state. Only about 10-12% of juniors take the PSAT (and only 2% take the SAT!). If WI started requiring the PSAT, then I can give you my own Wisconsin guarantee that the state’s cutoff would go up. Short of that, it will be interesting to see if there is any slow, long-term trend toward the PSAT.
Thanks for asking today’s most intriguing question!
Art – thanks for all you do for students and parents alike! I’m sorry if I missed this, but when do you expect to release your first projections for PSAT cutoffs for the class of 2027? Blessings to you!
Roger,
I just hit “send” on my initial analysis of scores. I’ll have further updates on Monday, but the core of the results are now online. I expect most cutoffs to go down. Big changes!
Thank you, Art! Fantastic analysis as always! I am fascinated by the large drop in RW scores above 700, even significantly lower in an absolute sense than the class of 2025. That makes me think (hope?) there is even more room for the selection index to drop given the double weighting of that score.
Are you able to post/provide the Math and RW ranges from prior years, or are they available elsewhere on your site?
I’ve got a 220 selection index kid in Texas hoping for a reversion back to that level or lower! 🙏
Thanks again!
Roger,
“That makes me think (hope?) there is even more room for the selection index to drop given the double weighting of that score.”
You are not wrong. I tend to be conservative when calling large swings. I’d like to see where we end up with Commended.
No, I don’t haven’t been able to track down older RW/M splits other than the percentages College Board lists in its annual summaries. For example, in earlier years all we know is something like 3% RW and 4% M. There is a big difference between 2.6% and 3.49%, but they both round to 3%. I may throw them up nonetheless.
Hi Art,
Since last batch of PSAT scores came out today, when will you let us know a read on the number of high scores nationally and your prediction for cut off?
Sam,
I just hit “send” on my initial analysis of scores. I’ll have further updates on Monday, but the core of the results are now online. I expect most cutoffs to go down. Big changes!
Thanks for this info. How does the decline in test takers figure into the number of commended and NMSQ students? I see the drop and percentage of high students is still very high. With that said, I calculated the high reading percentage (results over test takers) and even with the population drop it’s a much lower percentage than last year and the previous one. I did the same for math and saw only a marginal decrease making me think that Desmos has now raised those for the near future and wondering if perfect math scores will become more common, putting even more emphasis on reading on both PSAT and SAT.
Gail,
The decline is test takers is modest this year. One of the reasons that I look at the number of students in the 1400-1520 range is that those are the students who matter most for National Merit. It’s possible that Desmos has contributed to higher Math scores. I remain a tad skeptical, maybe because I’ve been waiting for calculators to break the SAT since they were first allowed in 1994. I think College Board did a poor job scaling the RW last year.
Hi Art. Thank you so much for your analyses, which are incredibly informative. It’s also super interesting for those who like to geek out on this type of stuff. Speaking of which, I was curious what kind of data you use for your score cutoff analyses/predictions. E.g., does the College Board release or sell some of this data? Surveys of your clients? Etc. Thanks again!
Jeff,
Thank you for the kind words. Unless noted otherwise, I analyze data from College Board or National Merit. I cannot think of any reason — short of potential embarrassment — for College Board not to make all these data widely available. And, yet, it does not.
Hi Art,
Once again, thank you so much for your detailed insight. I love reading your analysis and have some renewed hope for my daughter’s 222 in CA. I was wondering if the data you analyze from college board is publicly available. In particular the split between R/W and math. Also, just as an anecdote to the current data, my daughter got a 760 on math and 730 on R/W whereas math is her lower ACT and SAT score.
Christie,
Yes, I’ve got my most likely at 223 as something of a placeholder until we see where the Commended cutoff falls in April. I think you could make a solid argument for 221, 222, or 223 as the “most likely” this year. There are just so many unknowns. Your daughter’s scores are the sort of variabilities that make things interesting.
It seems logical for College Board to post the data publicly. Alas, it does not.
Hi Art , is there a chance that NMS cut offs will go down for IL for the class of 2027 ? Thank you!
Vamsi,
I think declines are likely in just about every state (which is not the same thing as saying that they will come to pass in every state). I think Illinois is likely to go down to 220 or 221 based on the national numbers. Good luck!
Based on the RW ratio and the number of total scores over 1400 it would seem like cutoffs should be broadly lower than 2025. Are you being conservatively high, or do you have access to more information that changes this assumption?
Liam,
Yes, whether it is recency bias or conservatism, I have some hesitation in putting Most Likely’s at “broadly lower than 2025?” Do I disagree with your assessment? No. Without a bit more evidence, I’m going to keep my estimates conservative while encouraging independent thought among readers! That’s why I share what I know.
In Maryland, with a score of 222, I unfortunately missed the qualifying cutoff according to your analysis and prediction. But I still appreciate your professional assessment.
Daniel,
Thank you. Don’t count yourself out at all! Heck, I can make a case that 221 has a decent shot, since it qualified three times in the last decade, twice with a Commended cutoff of 209 and once with a 207. I think your odds are over 50%. My “Most Likely” doesn’t mean that all other things are unlikely. We just don’t have a good handle on how much the lower RW count is going to drop those high Class of 2026 cutoffs. It could be a LOT.
What would you say your level of confidence is in the range estimates for a state? We’re in Missouri and my son got a 218; while confident, he’s sweating bullets as it’s one point above last year and at the top end of the range estimate. Any reason to think (or perhaps “expect” is a better word) that it would spike higher than that? Thanks!
There seems to be strong downward pressure on scores this year. Missouri would need to go up 2 points beyond its current record of 217. No reason to expect that to happen. Not impossible, but stretches credulity.
Hi Art,
Thanks for the detailed analysis. It really helps parents and students as we wait for final results. Any chance you know how many scores in the 1400-1520 range by state? Specifically, I’m looking for Georgia.
Thank you!
AK,
I’d love to have that information, and I’d share it if I did. The lack of the state data is why I emphasize that those estimates can be unreliable. Good luck!
Hi Art, you mentioned that there is a downward pressure on the SI index cutoff for 2027. We are from Mississippi. Your cutoff prediction for Mississippi remains unchanged from the 2026 cutoff. Would that indicate that students in Mississippi performed better than or the same as they did last year? Though my total score is good and almost perfect in Math, the SI index is low due to a low score in RW, and my SI index is exactly at your “Most Likely” prediction. It seems a little unfair to me to give a boost to the RW score but not to Math; however, it’s not within our control. Do you expect the cutoff to be at “Most Likely” or below, or can it be in the expected range?
Chris,
This is going to be a bad year for Most Likely’s to hit things on the nose. Mississippi has had a lot of volatility (as low as 209 and as high as 215). I don’t know that any of those scores is impossible, although I currently think that 209 is out of range. I considered 212 and 213 to be a coin toss as to which one to list as my ML, but I wouldn’t give that much weight. If your score is within the range, you have a shot! If it’s above the range, then you have more than a good shot!
Yes, NMSC decided to go with the double weighting 60-70 years ago, so I don’t foresee any changes coming soon.
How confident do you feel about the “most likely” number for Arkansas?
Cindy,
I don’t put much stock in it, honestly. The most heavily weighted variable plunged by what is, as far as I can tell, an unprecedented amount. But I don’t have enough history on the RW scores to put this year onto a continuum. Here is what I feel pretty good about saying: I think 80% or more states will see lower cutoffs, so there is a very good chance that Arkansas will be below 215. 212 vs 213 vs 214? I’d grade them similarly. 25% chance for each?
Hello. Amazing analysis!! This is why I look forward to coming to this site every year for my kids. My son got a 220 in Texas. If you had to give some kind of percentage estimate, what are the chances it goes to 220 here in Texas? Thank you!!
AA,
Thank you! I have a certain dread — not your fault — for these sorts of questions this year because of how much uncertainty is involved. I feel pretty good that it will be less than 222. Maybe a small chance of a 218. For 219, 220, and 221, I don’t know that there is a lot to distinguish among them. 25% – 30% chance each? Your son would only miss out at 221 and 222, so I would handicap that as a 70% chance of qualifying.
Art- Daughter has a 220 in Florida. Per your analysis, she looks to be in decent shape. Obviously, anything is possible. How much of a chance that the Florida number would increase from last year? Thanks so much for your insight!
Mac,
I don’t see a large state like Florida moving against the tide this year. There might be a low single-digit chance of an increase, but your daughter is quite safe at 220. Congratulations!
I appreciate your thorough analysis!
Am I correct in understanding that you have access to data regarding the overall PSAT scores at the national level only? If so, without score data per state how are you creating estimates of individual state ranges?
My child has a 214 SI in Oklahoma – we are feeling pretty good about national merit chances considering that this would be a qualifying score for 8 out of the last 10 years, but we would feel a tiny bit more hopeful if your range estimates do include per state data.
TTA,
I’d feel better with state data, too. Since those are not available, I can only look at the overall trend (“downward pressure”) and historical performance, especially in years that seem most similar to this year’s figures. The 2 years you point out as over 214 also had far more top scorers than this year, and a higher Commended cutoff than last year’s (which I expect to go down). So I like your student’s chances. I think we’d need to go back at least a dozen years to find Oklahoma at 5 points over the Commended level, and it has never hit 6.
Thank you for that additional insight into the top scorers in 2015/2016 and pointing out that commended delta from those years as another point of comparison. Feeling fairly confident, but still have 10 months of waiting
Hello Art. Amazing analysis! My son got a 220 in Illinois. If you had to give a percentage estimate, what are the chances it goes to 220 here in Illinois? Thank you!!
Vamsi,
We can probably rule out a 222 repeat, so it just becomes a question about the chances for a 221. The years where IL has hit 221 saw far more 1400-1520 scores nationally. I’d give it a 20% chance because of sheer uncertainty over the scaling this year. That still leaves your son with an 80% chance. Feel free to round that up!
Thank you so much for your input and for the help you are providing to the community. Happy Thanksgiving! God bless you!
This is incredibly helpful – love this kind of insight and would love to see more detail on your estimates if you have the time and inclination. My question however is not directly related to the NMS cutoff although your analysis has put my mind at ease after last year’s high index cutoffs (NY-224 for my class of 2027 son).
My question relates to the digital SAT and how stable it is. I am hearing about kids taking it 6, 7 or 8 times. Is there any evidence that this helps meaningfully even for the superscores? I understand taking it twice if you had a bad day, but 6 times! Also my sense is that kids think they can game the dynamics of the multistage adaptive test modules. They seem to believe that if they can only get to the harder second module their scores will meaningfully increase. My guess is that even if you do a little better on the first module your relatively poor performance on the harder second module will bring your score down, so no real difference. Like everything, it has gotten really gamified.
Thank you, Anton. My time and inclination is likely going to be driven by whether or not I can find any more useful data.
There is a decreasing return when repeating the SAT that many times. I haven’t seen studies that reflect gains from that 7th attempt, and College Board’s older research on repeat testing ignores superscoring. It has become more common for students to test 3-4 times. A number of factors are involved. There is, as you point out, the variability of the test and the variability of a student’s own performance. There is also the fact that students continue to improve their skills over time. We’ve seen a trend to earlier testing. There is a desire to get it out of the way or to provide a baseline for future preparation. That makes sense. But students are often not at their peak in sophomore year or fall of junior year, so there is usually a reason to repeat. Testing opportunities have also expanded with the addition of August, September, and school day options.
I hope students aren’t repeating the test in order to game the section adaptive structure, because that is simply not going to work. For students in the middle band of scores, routing has almost no impact. Students at a 500, for example, fall in what SAT calls the “area of indifference.” Flip a coin, choose a route, and, as you say, end up in the same place. Counterintuitively, modeling shows that a higher ability student should choose the lower difficulty section (choice is not an option in the real world!). Consider the extreme example of a 750M student wanting a perfect score. If that student aces the first section, knows it, and can choose the second section, they should choose the easier one. They can more easily answer ever question correctly. The unavoidable problem with that strategy is that a student does not get to choose the section difficulty. The only way to end up with the easier section is to miss a lot of problems. And if you miss a lot of problems…there goes the 800 (or even the 750).