GMAT Focus Data Sufficiency Questions – The Ultimate Guide
GMAT Focus Data Sufficiency questions test whether given statements can answer a question — not the answer itself. Master the AD/BCE elimination method and the SCAN framework (Sort, Convert, Arrange,...
GMAT Focus Data Sufficiency questions test whether given statements can answer a question — not the answer itself. Master the AD/BCE elimination method and the SCAN framework (Sort, Convert, Arrange, Navigate) and you will cut careless mistakes by half. This guide covers everything, with an interactive decision tree and practice questions built in.
Most test-takers approach GMAT Data Sufficiency questions the wrong way. They try to solve the problem. The question never asks you to. It asks whether you can.
This shift in mindset is what separates a 655 from a 705. GMAT Data Sufficiency questions now live in the Data Insights section of the GMAT Focus Edition, and they remain one of the highest-leverage question types on the exam. Once you understand the structure, a lot of the difficulty dissolves.
This guide walks you through every layer: the question format, the two types, the AD/BCE method, the SCAN framework, worked examples including the traps that catch strong students, common pitfalls, and an interactive practice question you can try right here.
Prepathon is a free Data Insights practice tool built for the GMAT Focus Edition. Adaptive DS questions, instant step-by-step explanations, no sign-up needed.
What Are GMAT Focus Data Sufficiency Questions?
A question format unique to the GMAT — one that tests analytical thinking more than arithmetic.
Data Sufficiency (DS) gives you a problem and two statements. Your job is not to solve the problem. It is to decide whether the information in those statements is enough to solve it.
Every DS question has the same five answer choices. These never change, which is an advantage once you internalize them:
- (A) Statement (1) ALONE is sufficient, but statement (2) alone is not sufficient.
- (B) Statement (2) ALONE is sufficient, but statement (1) alone is not sufficient.
- (C) BOTH statements TOGETHER are sufficient, but NEITHER statement ALONE is sufficient.
- (D) EACH statement ALONE is sufficient.
- (E) Statements (1) and (2) TOGETHER are NOT sufficient.
Internalise the five DS answer choices with practice and recall. Eliminate choices on finding statements sufficient or insufficient.When you stop having to think about what each letter means, you free up working memory for the actual problem.
Every DS question has three parts in this order: (1) the question stem (setup + question), (2) two statements labeled (1) and (2), and (3) the five fixed answer choices. The question stem may include background information, then asks either for a specific value or a yes/no answer.
Two Types of GMAT Data Sufficiency Questions
Understanding which type you are facing changes exactly how you evaluate sufficiency.
Type I: Value DS Questions
You need to find one unique numerical value for the unknown. Sufficiency means a single definitive number. Any range or multiple possibilities makes the statement insufficient.
- Trigger phrases: “What is the value of x?”, “How many…”, “How much…”
- If a statement gives you two possible values for x: it is insufficient
- If a statement constrains x to exactly one value: it is sufficient
Type II: Yes/No DS Questions
The question has a yes/no answer. Sufficiency means the statement always leads to the same answer. Always “yes” or always “no.” A statement that sometimes gives yes and sometimes gives no is insufficient.
- Trigger phrases: “Is x greater than y?”, “Is n even?”, “Does…”
- A definitive “no” is sufficient. It consistently answers the question
- A “sometimes yes, sometimes no” is insufficient
What Concepts Are Tested in GMAT Data Sufficiency?
DS questions can draw from any math concept in the GMAT quant universe — and also appear in non-math, logic-based form.
Math-based DS questions draw from:
Non-math DS questions test logical reasoning through logic puzzles — think seating arrangements, sequencing, or comparison-based puzzles. They look similar to Critical Reasoning questions. No formulas needed — just careful reading and logic.
Data Sufficiency is one question type inside the broader GMAT Data Insights section, which also includes Table Analysis, Two-Part Analysis, and Graphics Interpretation. Getting strong at DS is the fastest path to a high Data Insights score.
The AD/BCE Method: Step by Step
The core decision framework — works because the five answer choices split cleanly into two groups based on how Statement (1) performs alone.
| If Statement (1) alone is… | Remaining options | Next step |
|---|---|---|
| Sufficient | A or D | Check Statement (2) alone. Sufficient → D. Not sufficient → A. |
| Not Sufficient | B, C, or E | Check Statement (2) alone. Sufficient → B. Not sufficient → go to C/E. |
| If neither alone works: Check both together. Sufficient → C. Not sufficient → E. | ||
Evaluate each statement in isolation first. Never let what you know from Statement (1) bleed into your evaluation of Statement (2). This is the single most common way test-takers accidentally choose C when the answer is B.
Try the AD/BCE Decision Tree INTERACTIVE
You have read the question stem and both statements. Start here.
Is Statement (1) alone sufficient to answer the question?
Statement (1) is sufficient. Eliminate B, C, E. Now check Statement (2) alone.
Statement (1) is not sufficient. Eliminate A and D. Now check Statement (2) alone.
Neither statement alone is sufficient. Now combine (1) and (2) together.
EACH statement ALONE is sufficient. Both (1) and (2) independently answer the question.
Statement (1) ALONE is sufficient, but statement (2) alone is not sufficient.
Statement (2) ALONE is sufficient, but statement (1) alone is not sufficient.
BOTH statements TOGETHER are sufficient, but NEITHER statement ALONE is sufficient.
Statements (1) and (2) TOGETHER are NOT sufficient. You cannot determine the answer even with both statements combined.
Prepathon’s adaptive DS questions let you practice statement isolation and the AD/BCE method with step-by-step feedback on every attempt.
Run SCAN Before You Touch a Statement
AD/BCE tells you which answer to pick. SCAN is what happens before that — twenty to thirty seconds, four decisions, in order.
Sort the question type, then commit to its pathway
This is not a label. It switches on a different mode of thinking, and you stay in that mode until you have an answer.
Value → convergence mode. You are hunting for the single number the statement pins you down to. Nothing else matters until that number either appears or doesn’t.
Yes/No → corner-case mode. You are hunting for the exception. Assume the statement gives you “yes,” then go looking for one scenario where it flips to “no.” If you find even one, the statement is insufficient — you don’t need to resolve the question, just break its consistency.
These are two different searches, and hence two different mental pathways. Convergence mode looks for one answer. Corner-case mode looks for a counter-example. Decide which search you’re running before you touch Statement (1), and don’t drift into the other one mid-question.
Convert the stem to match the statements
Units first: if the stem is in kilograms and the statements are in grams, convert the stem once. Do not convert each statement back as you go.
Complexity next: if the stem gives multiple variables and equations, simplify it before reading either statement.
Arrange the data
- More than three data points, or mutually exclusive categories — create a table (make sure you have a column for total and a row for total).
- “Either / both / neither,” three or fewer overlapping groups — a Venn diagram is the best way to structure the data.
- Everything else — jot the values down. Simple numbers or specific conditions in your working memory are what slips under time pressure.
Navigate the question
What do you need? State precisely what “sufficient” means here and jot what you need to answer the actual question. This is usually going to lead you to the answer to the DS question itself.
Which statement first? Scan both statements first and work on whichever one lets you eliminate answer choices immediately. A statement with no real constraint — for example, “x is a non-negative integer,” with nothing pinning it down further — can often be ruled insufficient on sight. If that’s Statement (2), you eliminate B and D in the first thirty seconds, before you’ve done any real evaluation. You don’t have to follow the order of Statement (1) then Statement (2) strictly.
Bookmark or push through? If the stem allows multiple cases (positive/negative, even/odd, unconstrained) on a question early in the section, flag it and move on rather than spending three minutes. Use the bookmark if, after the first 30 seconds, you are still struggling to make sense of the information given and the ask of the question.
SCAN does not replace AD/BCE. It runs before it. Skipping SCAN is why capable students still take four minutes on a question built to take ninety seconds.
Worked Example: Applying SCAN and Solving the DS Question
Let us walk through a complete example so you can see exactly how SCAN changes the way you attack a question that looks harder than it is.
(1) A total of $1,000 was withdrawn from Emily’s savings account last month.
(2) At the beginning of last month, Emily’s account balance was $500.
They read the stem, skip straight to the statements, and start inventing numbers. Statement (1) says $1,000 was withdrawn, so they imagine a starting balance, subtract $1,000, check if it dips below $200, then try a different starting balance and check again. Statement (2) says the opening balance was $500, so they imagine a withdrawal amount and check again. Multiple scenarios, multiple guesses, no structure. This is trial and error dressed up as reasoning, and it eats the full two minutes before landing anywhere solid.
S — Sort: “Did the bank charge a fee” is a yes/no question. Corner-case mode: assume “no fee” and go looking for one scenario that forces “yes,” or vice versa.
C — Convert: No units to align. The stem itself needs decoding — “falls below $200 at any time during the month” is the real condition, and it is stricter than it first sounds. It does not ask about the balance at month-end. It asks whether the balance dipped below $200 at any single point, even briefly.
A — Arrange: Jot the keywords as you read, don’t hold them in your head — condition: balance below $200, at any time; Statement (1): $1,000 withdrawn (total, no timing given); Statement (2): opening balance $500.
N — Navigate, what do you need: to answer this with certainty, you need three things — the opening balance, every withdrawal, every deposit, and the timing of each. Miss any one of these and you cannot rule out a dip below $200 at some point in the month.
Which statement first: doesn’t matter much here, both fail fast. Statement (1) alone gives a withdrawal total but no opening balance and no timing — insufficient. Eliminate A and D. Statement (2) alone gives an opening balance but no withdrawal detail — insufficient. Eliminate B. Down to C or E, inside thirty seconds.
Combine: opening balance $500, minus $1,000 withdrawn, means the account went negative at some point — well below $200. That part looks locked. But this is where the trap sits. The question never says $1,000 was the only movement on the account last month. A savings account with withdrawals almost always has deposits too, and neither statement nor the combination rules that out. If a deposit landed before the withdrawal, the balance may never have dipped below $200 at all.
Here is where a lot of test-takers talk themselves into C. The thought goes: “why would I assume there were deposits? Nothing in the question mentions any.” That sounds careful. It is actually the opposite. There is a difference between a fact and an assumption. It is a fact that a savings account will have deposits, just as it will have withdrawals — that is simply how such an account operates. We cannot rule that out unless the stem explicitly states that withdrawals were the only transactions last month. It does not. Assuming a quiet account with no deposits is not caution. It is adding a constraint the question never gave you.
Final Answer: E.
The giveaway that C is wrong is not spotting the deposit — it is asking “what do I need” before you touch the statements. If you had listed opening balance, withdrawals, deposits, and timing as the four things required, you’d have seen deposits missing from both statements combined, and E would have been obvious the moment you finished reading. The people who fall for C are not bad at maths. They are skipping the step where you define sufficiency before you go hunting for it.
Practice: Try a GMAT Data Sufficiency Question Yourself
Apply what you have just learned. Read the question, run SCAN before touching the statements, route yourself through AD/BCE, then select an answer below.
Data Sufficiency: Medium Difficulty
(1) n is divisible by 3.
(2) n is divisible by 2.
Select the best answer:
Statement (1) alone: n is divisible by 3. This does not guarantee divisibility by 6. For example, n = 9 is divisible by 3 but not by 6. Insufficient. Eliminate A and D.
Statement (2) alone: n is divisible by 2. Not enough. n = 4 is divisible by 2 but not by 6. Insufficient. Eliminate B.
Both together: n is divisible by both 2 and 3, which means n is divisible by LCM(2,3) = 6. Sufficient. Answer is C.
Worked Example: The Trap Answer
A harder one, for after you’ve got the basic pattern down.
(1) 3n is even.
(2) 5n is even.
They see “3 times something is even” and jump to “so n must be even,” then do the same for Statement (2), and mark D without a second thought. D is the standard trap here, and it looks completely reasonable — which is exactly why it’s dangerous.
S — Sort: “Is n even” is a yes/no question. Corner-case mode: assume yes, then hunt for a scenario that flips it to no.
N — Navigate, what do you need: you need to know whether n is always an integer, and always even, under each statement. This is the detail the trap skips — nobody asked whether n has to be an integer at all.
Evaluating Statement (1) alone: 3n even does not force n to be an integer. Try n = 2: 3n = 6, even, and n = 2 is even — gives “yes.” Now try n = 2/3: 3n = 2, still even, but n = 2/3 is not even — gives “no.” Same statement, two different answers. Insufficient. Eliminate A and D.
Evaluating Statement (2) alone: same trap, same fix. n = 2 gives 5n = 10, even, n even — “yes.” n = 2/5 gives 5n = 2, even, but n = 2/5 is not even — “no.” Insufficient. Eliminate B. Down to C or E.
Combining both statements: reason algebraically rather than hunting for more fractions. 3n is even, and 5n is even. Look at 2×(3n) − 5n. That’s just 6n − 5n = n. And 2×(3n) is even minus even, which is even. So n itself is forced to equal an even number, built entirely from the two given quantities. There’s no fraction that can slip through this, because the combination isn’t testing values anymore — it’s proving n has to be even, directly.
It’s worth sitting with why the single fractions (2/3, 2/5) no longer work here. For n to dodge being even while satisfying both statements at once, it would need a denominator that survives multiplication by both 3 and 5 without ever fully cancelling — and no such non-integer exists once both constraints apply together. The two statements close every gap the other one left open.
Answer: C.
The reason D feels so safe is that “3n even → n even” is true when n is already known to be an integer — and most people import that assumption without noticing they’ve done it. The fix isn’t memorising this example. It’s the habit from N: before you accept a statement as sufficient, ask what n is even allowed to be. The moment you stop assuming “integer” for free, the fractions show up, and the trap disappears.
Deconstructing a Data Sufficiency Question Stem
Some DS questions are hard not because the statements are tricky, but because the stem is carrying a relationship you have to unpack before the statements mean anything. This is exactly the “C — Convert” moment from SCAN, played out in full.
(1) Peter’s annual salary was $30,000 that year.
(2) Anna’s annual salary was $40,000 that year.
Reading the stem literally, first pass: “the difference between Laura’s and Peter’s salaries was twice the difference between Laura’s and Anna’s salaries” translates to L − P = 2(L − A). But stop here, and be honest about what “difference” means — it could just as easily run the other way. This is where the second sentence stops being a throwaway detail. “Laura’s salary was the highest of the three” fixes the direction: if Laura earns the most, then L − P and L − A are both positive by definition, so there is only one way to write this equation correctly: L − P = 2(L − A).
At this point, most people go straight to the average, (L + P + A) / 3, treat it as a three-variable problem, and assume they need to solve for L, P, and A individually before they can find the average. That assumption is the whole trap. It’s what makes this look like it needs both statements, or neither.
Deconstructing further: take the relationship from the stem and isolate L: L − P = 2L − 2A, so L = 2A − P. Now substitute this into the average: (L + P + A) / 3 = (2A − P + P + A) / 3 = 3A / 3 = A. The P terms cancel entirely. The average of all three salaries is just A — Anna’s salary, on its own. You were only ever solving for one variable, and the stem told you which one the moment you deconstructed it properly.
Now the statements are almost trivial. What you need: a value for A, nothing else. Statement (1): Peter’s salary was $30,000. This tells you P, which has already been cancelled out of the equation. It gives you nothing about A. Insufficient. Eliminate A and D. Statement (2): Anna’s salary was $40,000. This is exactly what you need. Average = A = $40,000. Sufficient on its own.
Answer: B.
This is a classic case of a question that punishes you for not deconstructing the stem before touching the statements. Solve for the average in its raw form, and it looks like a three-variable problem demanding both statements — which is exactly why C is the trap answer here, not D. The stem was never asking you to find three salaries. It was asking you to find one, disguised inside an average. Deconstruction isn’t an extra step for hard questions. It’s what turns a hard question back into an easy one.
Looking for more questions like this? Work through the official question bank and supplement with GMAT practice tests under timed conditions. That is where DS skill actually sticks.
Three Common Mistakes and the Five-Step Process
Most DS errors come from one of three failure modes. The SCAN process builds the habit that prevents all of them.
Skipping the pathway commitment
Many test-takers don’t decide upfront whether they’re in convergence mode or corner-case mode. They evaluate a value question like a yes/no question, or accept the first result on a yes/no question as final instead of hunting for the exception that breaks it. This is exactly how a trap answer like D gets picked when C is correct — the person never went looking for the counter-example that would have exposed it.
Half-deconstructing the stem
You write down the relationship the stem gives you — an equation, a constraint — and stop there, without pushing it through to what the question is actually asking for. The stem then looks like it needs more variables pinned down than it really does, because the shortcut buried inside it never surfaced.
Mistaking silence for a ruled-out possibility
Nothing in the stem mentions a detail, so the assumption is that it doesn’t exist. This is not caution. Absence of a number is not the same as absence of the thing itself — a savings account has deposits and withdrawals as a matter of how such accounts work, whether or not a figure is given.
The Five-Step Process
For every DS question, run through these in order. Speed comes later. Build the habit first.
Prepathon tracks your DS accuracy across question types and builds an adaptive drill list around your specific weak areas. Free to use.
Still have questions?
Every trap on this page fell apart the same way: someone skipped straight to the statements before deciding what they were even looking for.
That’s the entire lesson. Not a new formula. A pause, spent asking what “sufficient” actually means, before the calculator in your head switches on.
Strip away the algebra and this is a decision-making skill. Knowing when you have enough to decide, and when you don’t, is exactly what a boardroom asks of you. Data Sufficiency is the smallest version of that instinct. The MBA, and the career after it, asks the same question at a larger scale, with no five neat choices to pick from.
Curious how your DS accuracy translates to your overall score? Understanding GMAT scoring helps you prioritise which question types to focus on most in your prep window.
Prepathon is a free adaptive Data Insights practice tool for the GMAT Focus Edition. Take a 10-question DS set, get instant feedback, and see exactly where your accuracy is leaking.
Devmitra Sen is Head of Academics at Crackverbal and has trained over 4,000 students. Her scorers tell the story: GMAT 745, 725, 715, 705 alongside turnarounds like 575→715 and 375→675. She has produced multiple Q90 scores, including a perfect 100th percentile on GMAT Quant — a benchmark very few coaches can claim consistently. On Data Insights, her superpower is changing how students see, observe, and comprehend data: breaking it down, reasoning through it, and zeroing in on exactly what the question asks. The results follow: multiple 90+ percentile DI scores, including a 1st to 99th percentile turnaround in under two and a half months. She carries a quiet interest in the history of mathematical thought — particularly ideas rooted in India long before they were formalised elsewhere — a perspective that gives her an unusually grounded sense of why the subject matters.
TL;DR GMAT Multi-Source Reasoning (MSR) questions give you 2–3 tabs of data and ask you to cross-reference across all of...
TL;DR GMAT Graphics Interpretation (GI) questions appear in the Data Insights section and ask you to read a graph or...
TL;DR The GMAT is accepted by over 7,700 programs at more than 2,400 business schools globally — covering not just...
Your 705+ score starts with
one expert call.
30-minute strategy session. We'll audit your prep, find your gaps, and build a roadmap to your target score.