SHSAT cutoff trends 2021–2026

Every March, the NYC DOE releases the year's SHSAT cutoffs, and every March, families try to read meaning into the small year-over-year movements. Some years a cutoff moves up by five points and there's speculation about increased competition. Some years it moves down by three and there's speculation about easier tests. Most of these narratives are pattern-matching on noise.

I wanted to look at the actual cutoff data across the past five admissions cycles — 2021–22 through 2025–26 — and try to identify what's real signal and what's noise. The dataset is too small for rigorous statistical claims, but the patterns are still informative for families planning future cycles.

The full data, all 8 schools, 5 years

School	21–22	22–23	23–24	24–25	25–26	5-yr range
Stuyvesant	555	557	561	558	556	7
Staten Island Tech	520	524	528	531	527	11
Queens Sciences	521	525	532	530	527	11
HSMSE	521	524	530	528	526	9
Bronx Science	512	518	524	521	518	12
HSAS Lehman	514	516	520	519	516	6
Brooklyn Tech	502	508	514	510	505	12
Brooklyn Latin	489	495	501	498	493	12

Pattern 1: Cutoffs peaked in 2023–24

This is the clearest signal in the data. Every single specialized high school had its highest cutoff of the five-year window in 2023–24. Stuyvesant reached 561 — its highest in years. Bronx Science peaked at 524. Brooklyn Tech reached 514. The peak wasn't marginal; it was system-wide.

What drove it? Most likely a combination of pandemic-era cohort effects and the testing environment's normalization. The 2023–24 SHSAT was administered to students who had completed most of their preparation in a more stable environment than the 2021–22 cohort (who had pandemic-era 6th and 7th grade disruption) but before the 2025 CAT format transition. By the time those students sat for the test, they'd had cleaner preparation runs and stronger families-driven prep markets than the immediately surrounding cohorts.

Pattern 2: Higher-cutoff schools are more stable

Stuyvesant's 5-year range is 7 points (555–561). HSAS Lehman is 6 points (514–520). Every other school shows 9–12 points of range.

This pattern is structural, not circumstantial. Schools with higher cutoffs are determined by the upper tail of the score distribution, which has relatively limited variance because perfect scores are rare and the top 5% of test-takers cluster tightly. Schools in the middle of the cutoff distribution depend on the marginal admit position, where applicant pool composition can move the cutoff more.

For families: if you're targeting Stuyvesant, treat the cutoff as 555–562 with high confidence. If you're targeting a middle-cutoff school like Bronx Science or HSMSE, prepare for 515–530 because the variance is real.

Pattern 3: The 2025–26 drop is consistent with format change, not declining standards

The most recent cutoff data (2025–26) shows every school down from its 2024–25 cutoff. Some of the early reporting framed this as "specialized high school standards declining," but I don't think that's the right read.

The 2025–26 SHSAT was the first administration of the new Computer Adaptive Test format. Format transitions in standardized testing routinely produce 3–10 point shifts in equating during the first calibration cycle, because the underlying psychometric model is being applied to new conditions. The fact that every school moved by 2–5 points in the same direction (down) is consistent with a slight equating shift, not with a real change in academic preparation.

For families planning the 2026–27 cycle, I'd expect smaller year-over-year movement than the 2024→2025 transition. The format-change adjustment is mostly behind us. Future cycles will move based on normal applicant pool variation.

What this means for prep targets

Combining the patterns above, here are honest buffer targets for the 2026–27 cycle:

• Stuyvesant: Target 565. The school has been remarkably stable; 565 provides 9 points of buffer over the recent 556.
• Staten Island Tech, Queens Sciences, HSMSE: Target 535. These schools cluster in the 526–532 range; 535 covers the upper bound of recent movement.
• Bronx Science, HSAS Lehman: Target 525. Both schools have been in the 516–524 range; 525 provides comfortable buffer.
• Brooklyn Tech: Target 515. The school has shown the largest swings (502 to 514); 515 covers the recent range.
• Brooklyn Latin: Target 500. The lowest cutoff in the system has been 489–501; 500 provides safe buffer.

These targets are practice-test goals, not real-test goals. If you consistently hit these on full-length timed practice tests under realistic conditions, you have a high probability of clearing the corresponding cutoffs.

What I won't predict

I want to be explicit about what the data doesn't tell us. It doesn't tell us whether 2026–27 cutoffs will continue to drop, stabilize, or rebound. It doesn't tell us how the digital format will affect the applicant pool composition over time. It doesn't tell us whether NYC DOE policy changes will shift seat allocations.

Anyone selling you a definitive prediction of next year's cutoffs is selling you a prediction the data doesn't support. The honest answer is that recent ranges are useful planning anchors, but year-to-year variation is real. Prepare to the buffer targets above and don't over-index on any single year's number.