Section-Adaptive, Not Question-Adaptive: How GRE® Scoring Really Works
If you've read anything about GRE® scoring, you've run into the folklore. Forums warn about "the algorithm," about how the test punishes a rough start, about hidden sections that don't count, about treating every question as if your whole score rides on it. The warnings sound authoritative, and the instinct behind them makes sense: you can't see the scoring machinery from inside the test, so any confident-sounding explanation fills the gap.
The explanations tend to contradict each other, though, and ETS publishes the mechanism. The GRE® adapts between sections, not between questions. That distinction helps you tell which strategies fit this test and which ones come from a different adaptive model. This post walks through what ETS's own pages say about the design, how a raw score becomes a 130-170 scaled score, and where the folklore diverges from the published record.
What "section-level adaptive" means
ETS describes the Verbal Reasoning and Quantitative Reasoning measures as section-level adaptive. The first section of each measure is set at average difficulty. When you finish it, the computer selects your second section based on how you performed on the first. ETS's own example: do very well on the first Quantitative Reasoning section, and the second Quant section comes in at a higher level of difficulty.
Two details in that description matter for strategy:
- The adaptation happens once, at the section boundary. Your answer to question 7 doesn't change question 8. The questions in front of you were chosen before the section started, and they don't shift mid-section.
- It happens twice, independently. Verbal section 1 sets Verbal section 2, and Quant section 1 sets Quant section 2. A rough first Quant section has no effect on your Verbal sections.
For the counts behind those sections: each Verbal measure runs 12 questions in section 1 and 15 in section 2, timed at 18 and 23 minutes, and Quant runs the same question counts in 21 and 26 minutes. Four scored sections plus the essay is the whole test, about 1 hour 58 minutes.
That design is different from a question-adaptive test, where each answer feeds the difficulty of the next question. If you've prepared for the GMAT®, you've worked with a question-adaptive model, so some familiar advice may not transfer. The current GMAT® does let you bookmark questions and change up to three answers per section, and it's still a strong test; the difference that matters here is how each exam adapts, question by question on the GMAT® and section by section on the GRE®. We compare the two in full in our GRE® vs GMAT® guide.
Within a section, every question counts the same
ETS's scoring page includes a sentence the folklore tends to skip: "Within each section, all questions contribute equally to the final score." Question 1 weighs the same as question 12. There's no difficulty weighting inside a section, no bonus for solving the questions in an impressive order, and no penalty for answering them out of order.
And answering them out of order is legal by design. The test's published design features include preview and review capabilities within a section, "Mark" and "Review" tools for tagging questions, the ability to change answers while time remains, and an on-screen calculator for Quant. ETS states outright that the design lets you choose which questions within a section you want to answer first.
The scoring consequence follows from that freedom: an unanswered question scores nothing, while a guess has a chance of being right. On test day, the number to manage is your blank count, not your count of hard questions.
From raw score to 130-170
For each measure, ETS computes a raw score: the number of questions you answered correctly across the two operational sections. That raw score is converted to a scaled score through equating, and ETS says the equating process accounts for minor variations in difficulty among test editions as well as the difficulty differences introduced by section-level adaptation.
Equating matters because two test-takers rarely see the same second section. Suppose you and a friend test on the same day, you land the harder Verbal section 2, and your friend lands the easier one. Identical raw scores wouldn't mean identical performances, since you answered harder questions along the way. Equating adjusts for that, and ETS attaches a guarantee to it: a given scaled score reflects the same level of performance regardless of which second section was selected and when the test was taken. A 158 Verbal is a 158 Verbal whether you saw the harder second section or the easier one.
The scale itself runs from 130 to 170 for each measure, in one-point increments, per ETS's score-interpretation guide. Added together, the two measures give the 260-340 range. Analytical Writing is scored separately on a 0-6 scale in half-point increments, with ETS's e-rater® engine among the scoring inputs.
What the first section decides, and what it doesn't
The first section does real work. Its difficulty is fixed at average, it's where the test decides which second section to hand you, and the difficulty level of the sections factors into scoring. Treating it as the foundation of your pacing plan is reasonable, and our pacing guide builds its budget around exactly that.
The folklore version goes further and claims the first section decides your score. That's not what ETS publishes. Scoring takes into consideration the total number of questions answered correctly across both sections along with the difficulty levels, and every question in your second section contributes equally within it, so a strong second section can still improve your score on either path. A harder second section typically means the first one went well, since the test routes you upward when first-section performance supports it and equating accounts for the difficulty difference. The tougher clock isn't a signal that anything went wrong.
One honest gap: ETS publishes the mechanism, not the routing table. How many first-section correct answers send you to the harder path, or what a given path's scoring range looks like, aren't public numbers. Sources that quote an exact routing table are estimating, and it's worth knowing that before you build a plan around a specific quote.
Three forum myths, checked against ETS's pages
Myth: The test hides an unscored experimental section.
That was true before September 2023 and isn't true of this format. The old test ran about 3 hours 45 minutes across six sections, one of which was an unscored section (in some administrations, an identified research section appeared in its place), plus a 10-minute break after the third section. The current test has five sections in about 1 hour and 58 minutes, and the published section times account for all of it: 30 minutes for the Analyze an Issue task, 41 for Verbal across two sections, and 47 for Quant across two more. There's no hidden section on the current format, and all five sections are scored: the four Verbal and Quant sections build your 130-170 measures, and the essay gets its own 0-6.
Myth: The first section decides everything.
ETS's scoring takes into consideration the total number of questions answered correctly across both sections, as well as the difficulty level of the sections. Both sections count, and the second one counts equally within itself.
Myth: Skipping questions is risky.
The scoring risk runs the other direction. Skipping, marking, and returning are published design features of the test, there's no penalty for a wrong answer, and an unanswered question scores nothing. The costly version is a section that ends with time on the clock and blanks on the screen. Our timing guide shows the budget math that makes the return trip possible.
Three moves that fit the mechanism
Build section 1 for accuracy. Your first-section performance decides which second section you see, so steady accuracy there keeps the stronger scoring paths in reach. Rushing question 8 to bank seconds for question 20 can cost more than it saves.
Use the navigation you're given. Skip, mark, return to, and change answers within the section. The freedom is a design feature, not a loophole.
Leave no blanks. Put a best guess on anything unresolved before the clock expires. A blank scores nothing; a guess might score.
FAQ
Is the GRE® adaptive?
Yes. ETS describes the Verbal Reasoning and Quantitative Reasoning measures as section-level adaptive: the computer selects the second section of each measure based on your performance on the first section.
What's the difference between section-adaptive and question-adaptive?
On a question-adaptive test, each response influences the difficulty of the next question, so the test tracks you answer by answer. The GRE® adapts once per measure, between sections: the questions within each section are fixed in advance, all count equally, and only the second section's difficulty depends on your first-section performance. The GMAT® works on the question-adaptive model.
Does the first GRE® section matter more than the second?
It matters differently. The first section is fixed at average difficulty and decides which second section you get. Scoring then weighs correct answers across both sections together with the difficulty levels, and ETS's equating means a scaled score represents the same performance regardless of which second section you saw. Your second-section answers count.
Can you skip questions and go back on the GRE®?
Yes. Published design features include preview and review within a section, Mark and Review tools for tagging questions, and the ability to change answers while time remains in the section. That freedom applies within a section; once a section ends, you can't return to it.
Does the GRE® have an unscored experimental section?
Not on the current format. The test has five sections (one Analyze an Issue task, two Verbal, two Quant) in about 1 hour 58 minutes, and the published section times account for all of them. An unscored section was part of the pre-September 2023 format.
Is a harder second section a good sign?
Usually, yes. ETS routes you to a harder second section when your first-section performance supports it, and equating accounts for the difficulty difference, so a tougher second section typically reflects a strong first one rather than a problem.
How is the GRE® raw score converted to the 130-170 scale?
Through equating. The raw score is the number of questions answered correctly, and equating converts it to a scaled score while accounting for difficulty differences across test editions and from section-level adaptation. That's why a given scaled score reflects the same level of performance regardless of which second section was selected or when you tested.
What happens if I don't answer any questions in a measure?
ETS reports an NS (No Score) for that measure. At least one answered question per measure is required for a score to exist at all.
Want to learn even more?
For the complete prep structure this scoring system sits inside:
For the pacing budgets that make skip-and-return work:
For how the adaptive design shapes difficulty, and why the test can feel harder than its content:
For the question-adaptive contrast in full:
And for what your scaled score means in percentiles:
We also talk through prep strategy on our podcast.
If you want help reading your own score report and building the plan from here, you're welcome to book a complimentary strategy session. We'll look at where your points are and figure out the path to where they need to be.