The new SHSAT officially tests English and math. But the digital version introduces another unstated and unscored skill: navigating the new digital test itself.
That skill costs time and attention. On this test both can cost points.
Based on automated testing, human speed trials, and trials with students unfamiliar with the interface, we estimate that navigating the digital SHSAT could consume roughly 16 to 24 minutes of the 180-minute exam.
That's time spent clicking, advancing, entering responses, moving through directions and passage transitions, and figuring out what the interface requires. It is not time spent reading, reasoning, or solving.
That's roughly 9% to 13% of the entire test.
We measured it
We built a tool to take the available 2026 practice SHSAT, which includes the new navigation restrictions. It solves nothing. It matches each question to a predetermined answer, selects it, and tracks the experience.
Across dozens of runs, completion times clustered around nine minutes. More than two minutes accumulated just between submitting one answer and displaying the next question.
Then we raced through the test ourselves. No reading. No solving. Correctness didn't matter. Click, advance, repeat. Our fastest run was 6 minutes, 27 seconds.
An experienced adult racing through a familiar interface is not the same as a student encountering it for the first time. So we gave predetermined answers to ninth graders in Georgia who were unfamiliar with the SHSAT and told them to get through the test as quickly as possible. Just figure out what to click and go.
They took 16 to 24 minutes. On paper, flipping pages and filling in 114 bubbles with no concern for right or wrong takes under two minutes, with almost no cognitive load.
Different methods, same finding: operating the digital test takes meaningful time even when answering the questions requires none.
The interface keeps interrupting the test
The available practice tests indicate that the digital SHSAT won't be a continuous sequence of questions.
ELA includes seven passage sets plus stand-alone questions. Each passage set gets a start screen and an end screen. Two more interruptions surround the stand-alone section.
That's 16 interruption screens across 50 ELA questions, before counting warnings or pop-ups triggered by errors or omissions.
The consequences of navigation also change. Advancing can mean simply moving to the next question or permanently giving up the ability to return to previous questions.
Then there are the new "Technology Enhanced Items." Students may have to drag items, open menus, select multiple answers, or enter responses into fields.
Each interaction adds something else the student must know, recognize, and execute correctly. Students will vary in how quickly and accurately they do that, creating another source of performance variation beyond the English and math knowledge the SHSAT is supposed to test.
No single interaction is particularly difficult, but stack them across 100 questions with a three-hour countdown and they become a steady drain on time and concentration.
The interface repeatedly breaks the student's flow, then charges the interruption against the same 180-minute clock.
That's a testing problem
The testing profession has a name for this: construct-irrelevant variance.
The Standards for Educational and Psychological Testing defines it as the extent to which test scores are affected by processes outside what the test intends to measure.
The SHSAT is supposed to measure English and math. Navigating its interface isn't one of those stated skills. The issue isn’t whether students can learn these interactions. It’s whether learning them should affect a score intended to measure English and math.
Standard 6.4 says the testing environment should provide "minimal distractions to avoid construct-irrelevant variance." Standard 6.5 calls for appropriate instructions, practice, and support to reduce it, including practice with equipment unless operating that equipment is itself being assessed.
But familiarity with this particular digital environment clearly matters.
Repeated practice teaches students where to click, how each response type works, which screens can be moved through quickly, when answers become locked, and how to navigate without breaking concentration.
That knowledge saves time. Students with more access to practice arrive with more of it.
But the available practice tests create another problem. NYCPS says they mirror the actual exam in both form and content, while the actual SHSAT will be adaptive. Our repeated testing produced the same questions despite deliberately different answer patterns. The practice tests behaved like fixed, linear tests.
So students need familiarity with the practice tests to reduce the interface burden, but those tests can’t provide complete familiarity with the actual test.
The practice test can teach students where to click. They can't show students what taking an adaptive SHSAT actually feels like.
Fewer questions doesn't necessarily mean more working time
The paper SHSAT gave students 180 minutes for 114 questions, about 95 seconds per question. The new test gives them 180 minutes for 100 questions, or 108 seconds per question.
That looks like more time.
But it assumes all 180 minutes are available for answering questions.
If navigating the digital test consumes 20 minutes, students have 160 minutes left for 100 questions: 96 seconds per question. At 24 minutes, that falls to about 94 seconds.
At around 20 minutes of navigation, nearly all the apparent time gained by removing 14 questions disappears into operating the test.
Show the evidence
Usability is a known factor in testing and is important to investigate. We tested the available interface on a small scale and found a measurable burden.
For a high-stakes admissions test, NYCPS and Pearson should have much better evidence.
How much testing time does the interface consume? How much does familiarity matter? Do the new digital interactions affect performance beyond the English and math skills the SHSAT is supposed to measure?
If that research has been done, publish it. If it hasn’t, do it and fix what it finds.
Before 100 questions are the sole basis used to rank students for admission to NYC's most selective public high schools, NYCPS should be able to show how much of the test is testing English and math, and how much is testing whether students know how to operate the test.