GMAT Data Sufficiency: the strategy guide for 2026.

Data Sufficiency is the question type students either love or avoid. There's no in-between. The format is unusual, the answer choices are the same five every single time, and once you've internalised the logic the questions become unusually fast. Until then, every DS question feels like a trap. This guide covers the structure, the AD/BCE process I used to standardise my approach, the single mistake that costs more DS points than any other, and a study plan that actually moves the score.

One bit of context first: in the GMAT Focus Edition, Data Sufficiency lives inside the Data Insights section, not the Quant section. The format and logic are unchanged from the old GMAT — the move just shifted which section's score it influences.

DS isn't asking you to solve. It's asking whether the answer is determined. The students who internalise that distinction stop making the most expensive DS mistake there is.

The format

Every DS question has the same shape. A question stem, two numbered statements, and the same five answer choices in the same order:

1. Statement (1) ALONE is sufficient; (2) alone is not.
2. Statement (2) ALONE is sufficient; (1) alone is not.
3. BOTH statements TOGETHER are sufficient; neither alone is.
4. EACH statement ALONE is sufficient.
5. Statements (1) and (2) TOGETHER are NOT sufficient.

Memorise these. The answer-choice text is identical on every DS question; reading it on the exam is wasted time. By the time you sit for the test, you should be able to map any DS outcome to A/B/C/D/E without looking at the choices.

The AD/BCE process

The AD/BCE process is the single most useful operational shortcut for DS. It works like this:

1. Read the question stem.Translate it into a single concrete question. “What is the value of x?” or “Is x positive?”
2. Evaluate Statement (1) alone.Ignore Statement (2) completely. If (1) determines the answer, the answer is A or D — eliminate B, C, E. If (1) does not determine the answer, the answer is B, C, or E — eliminate A, D.
3. Evaluate Statement (2) alone.Ignore whatever you concluded from (1). Same logic: if (2) alone works, you're between D (if A also worked) and B (if A didn't). If (2) alone doesn't work, you're between C and E.
4. Combine only if you have to. If neither statement alone is sufficient, ask: do they together pin the answer down? If yes, C. If no, E.

The whole logic compresses to a 2×2 grid:

• (1) yes, (2) no → A
• (1) no, (2) yes → B
• (1) no, (2) no, together yes → C
• (1) yes, (2) yes → D
• (1) no, (2) no, together no → E

Five outcomes, five answer letters. Once this is automatic, DS becomes a different test. You stop second-guessing the answer-choice wording and start working the math.

The single most expensive DS mistake

Here it is: students treat DS like a Problem Solving question and try to compute the answer. DS isn't asking for the value. It's asking whether the value is determined. Those are completely different operations.

Computing means doing the algebra to find x.
Determiningmeans asking: “Given this information, can x take more than one value? If yes, not sufficient. If no, sufficient.”

On a value question (“What is x?”), determining sufficiency means proving x has exactly one value. On a yes/no question (“Is x positive?”), it means proving the answer is always yes or always no — never sometimes-yes-sometimes-no.

Students who try to compute the value lose 30-90 seconds per question that they didn't need to spend, and they often arrive at C when the right answer was D (because computing both statements together feels like “more evidence” even when each alone was already enough).

The DS question is “is it pinned down,” not “what is it.” Compute only when you have to.

The four DS sub-types

1. Value questions

Asks for a specific number. “What is the value of x?” / “How many people attended?”

Sufficiency means: the statement narrows x to exactly one value. Two values fail. A range fails. “x is between 5 and 10” is not sufficient even though it sounds specific.

2. Yes/No questions

Asks for a true/false answer. “Is x positive?” / “Does the function pass through the origin?”

Sufficiency means: the answer is consistently yes orconsistently no across every case the statement allows. If the statement allows both yes and no scenarios, not sufficient.

Yes/no questions are where most students lose extra points, because “the answer is sometimes yes” and “the answer is no” feel similar. They aren't. “Sometimes yes” means not sufficient.

3. Geometry value questions

Common DS variant. “What is the area of triangle ABC?” with statements about side lengths or angles.

Sufficiency requires that the triangle is uniquely determined — not just that one is consistent with the statement. Two triangles can have the same two sides and different areas, so “sides AB and BC are both 5” is not sufficient by itself even though it constrains the shape significantly.

4. Conditional / system questions

Statements are equations or systems. “What is x?” with statements like 2x + y = 10 and x — y = 2.

Two equations, two variables, generally enough — but check that the system isn't degenerate (parallel lines, for instance). The classic trap is the system that looks sufficient but actually has infinite solutions.

The trap families

Trap 1 — The C trap

The most famous DS trap. The question feels like it requires both statements. You combine, get a single answer, and pick C. But each statement alone was actually sufficient — the right answer was D.

Defence: evaluate each statement in isolation, with discipline, before considering them together. The AD/BCE process is the defence.

Trap 2 — The forgotten case

On yes/no questions especially: you find a case where the answer is yes and stop. But there's another case the statement allows where the answer is no. You called it sufficient when it wasn't.

Defence: on yes/no questions, actively try to find a counter-example before declaring sufficiency. Push the variable to extremes, plug in negatives, plug in fractions, plug in zero.

Trap 3 — The hidden constraint

The stem says “x and y are integers.” You forget that constraint when working through the statements. Your cases include x = 1.5 and you conclude not sufficient when with the integer constraint it actually was.

Defence:after reading the stem, write the constraints down separately before touching either statement. The constraints are part of the problem; they aren't optional flavour.

Trap 4 — Sign errors on inequalities

DS loves inequalities. The classic error: multiplying both sides by a variable whose sign is unknown. If x could be negative, you have to flip the inequality — and you have to consider both possibilities.

Defence:before manipulating an inequality with a variable, ask “what do I know about its sign?” If the answer is “nothing,” split the analysis into cases.

Timing

The DI section is 20 questions in 45 minutes — an average of 2:15 per question. DS questions land below average, around 1:45 to 2:00 each. Bank time on DS so you have the headroom for the longer Multi-Source Reasoning sets.

If a DS question crosses three minutes, guess and move on. Sticking with one DS question past the three-minute mark costs you points elsewhere — the section's clock is unforgiving.

How to practice DS specifically

1. Drill the AD/BCE process to muscle memory.First two weeks: every DS question, write A/B/C/D/E in your scratch space, eliminate as you go. The point isn't accuracy, it's building the workflow into your hands.
2. One sub-type per session.Don't mix DS with PS until your DS workflow is automatic. Mixed practice teaches you nothing if your foundation is shaky.
3. Time individual questions from day one.Untimed DS practice creates a false sense of competence. The 1:45-2:00 budget is part of the skill.
4. Log every miss with a trap tag. Was it a C trap? A forgotten case? A hidden constraint? A sign error? The same four traps repeat across hundreds of questions. Tag them.
5. Re-do missed questions cold a week later.If you can't reproduce the right reasoning without looking, you didn't internalise the lesson.

The short version

Five answer choices, every time. AD/BCE process to evaluate each statement in isolation. Don't compute when you don't have to. Watch for the four trap families. Time in the 1:45 to 2:00 range. Once the workflow is automatic, DS becomes one of the highest-accuracy question types on the test.

The platform

Zakarian GMAT's DS practice is tagged by sub-type and by trap family, so you can drill the C trap specifically or the inequality sign-error specifically. The error log uses the four trap tags this guide describes, which means the patterns surface within two weeks of consistent logging. Free diagnostic if you want to see where your DS sits today.