5 decision rules that govern every GMAT Data Sufficiency item

GMAT Data Sufficiency is the section of the exam that punishes haste and rewards discipline. The format is deceptively simple: a short problem stem, two labelled statements, and a fixed five-option answer choice asking whether the data, alone or together, is enough. The cognitive trap is that test-takers treat it as a regular quant problem, plunge into arithmetic, and forget the question they are actually being asked. The GMAT (formerly the GMAT Focus Edition) edition preserves the original DS architecture: 18 items inside the Quant module, each scored on a unique logic grid where the correct response is one of five canonical answer choices (A, B, C, D, E). Reading that grid correctly, before any calculation, is the single skill that separates a high-30s DS subscore from a low-40s one.

The anatomy of a Data Sufficiency stem and why it must be rephrased first

A DS question is a small engine with three moving parts. The stem poses a question whose target value (a single number, a yes/no, a ratio, an expression) is hidden inside ordinary prose. Statement 1 and Statement 2 are independent fact packets that may or may not lock the answer down. The five answer choices are identical across the entire section, and they ask one meta-question: 'Is the information you have enough to answer the stem uniquely?' Rephrasing the stem into a precise target is the first non-negotiable step; without it, candidates waste time calculating the wrong thing.

Consider a stem such as: 'What is the value of x?' The target is a single real number. The statements can be read with one question in mind: 'Does this packet of information yield a single, unique value of x, no matter which scenario it allows?' If yes, the statement is sufficient. If the packet produces two or more legal values, the statement is insufficient, even if those values are numerically close. The trap is that test-tellers design stems where two values differ by a sign, a factor, or a hidden integer, and candidates stop the moment they find one answer.

Yes/no stems invert the logic. A stem such as: 'Is x greater than y?' is satisfied by Statement 1 only if, in every legal scenario the statement allows, the answer to the question is unambiguously 'yes.' If the statement permits even one scenario where x is less than y, Statement 1 is insufficient. The grammatical switch from 'What is…' to 'Is…' flips the sufficiency criterion from uniqueness to universal truth. Candidates who fail to notice this switch will mark a 'Yes' answer to the meta-question and a 'Yes' to the embedded question, conflating two different layers of the test.

Three rephrasing moves to make before reading the statements

• Convert the prose into a one-line target. 'What is the value of x?' becomes find a unique real number. 'Is the product of a and b positive?' becomes is sign(a × b) = +1 in every legal scenario.
• Strip social-context words. Phrases such as 'on average', 'at least', 'no more than', and 'an integer' are load-bearing; words such as 'in total' and 'together' often are not.
• Translate constraints into algebra. 'x is a positive integer' becomes x ∈ ℤ⁺, which is a domain restriction that can collapse two mathematical solutions into one DS answer.

This rephrasing habit is what allows a candidate to glance at Statement 1 and immediately judge sufficiency without re-reading the stem. In practice, candidates who skip rephrasing spend 30-45 extra seconds per item rewriting the question inside their head, which is more than the 2-minute DS budget can absorb.

The five canonical answer choices and the only logical relationships they encode

Every GMAT Data Sufficiency item resolves to one of five answer choices. Memorising them is not a study tip; it is the only way to avoid a common error pattern in which the test-taker produces the correct sufficiency conclusion but selects the wrong letter. The five choices, in the order they always appear, are:

• A: Statement 1 alone is sufficient, but Statement 2 alone is not.
• B: Statement 2 alone is sufficient, but Statement 1 alone is not.
• C: Both statements together are sufficient, but neither alone is sufficient.
• D: Each statement alone is sufficient.
• E: Even both statements together are not sufficient.

The middle three options describe the most important logical relationships. C requires the test-taker to verify that the two statements are complementary — each blocks a different route to ambiguity, and the union of the two blocks all ambiguity. D is the strongest possible data state and tends to appear in items where the stem is a value-target and both statements are independent algebraic locks. E is the most underappreciated option; candidates who reflexively try to combine statements sometimes miss cases where the union still allows two scenarios.

How to test the answer choice before committing

The efficient method is a two-pass judgement. First, evaluate Statement 1 in isolation. If it is sufficient, the answer is either A or D. Then evaluate Statement 2 in isolation. If it is also sufficient, the answer is D; if not, the answer is A. If Statement 1 is insufficient, the answer cannot be A or D, so the candidate moves on. This binary tree, applied consistently, removes the paralysis that afflicts candidates who try to read all five options after each statement. It also forces the test-taker to give every statement a clean sufficient/insufficient verdict, which is the only decision the answer choices are asking for.

The 90-second budget per item is a soft average, not a rule. Some DS items — particularly those with integer constraints, hidden squares, or domain restrictions — demand 120-150 seconds, and the candidate must trade that time against an easier item elsewhere in the section. In my experience, the candidates who plateau at a low DS subscore are the ones who refuse to spend more than 75 seconds on the hard items, then lose 4-5 minutes of cumulative time to re-checking the easy ones. The mental model should be: spend the time the item deserves, then move on without re-verification.

Statement 1 alone versus Statement 2 alone: reading the logical frame of a Data Sufficiency prompt

Statements 1 and 2 are deliberately written to feel symmetric. They often share a variable, a function, or a context (a business, a tank, a sequence), which leads candidates to assume their sufficiency status will match. In reality, the second statement is engineered to break the symmetry: it adds, removes, or transforms the very piece of information that made the first statement ambiguous. The candidate's job is to identify that breaking element, not to assume parallelism.

Consider the stem: 'A rectangle has perimeter 40. What is its area?' Statement 1: 'The length of the rectangle is twice its width.' This locks length and width uniquely: 2L + 2W = 40 and L = 2W give W = 20/3 and L = 40/3, so the area is uniquely 800/9. Statement 1 alone is sufficient. Statement 2: 'The diagonal of the rectangle is 10√2.' A perimeter of 40 and a diagonal of 10√2 also uniquely fix the rectangle, since the diagonal and perimeter together determine the side lengths. Statement 2 alone is sufficient. The answer is D. The symmetry holds here, which is precisely what makes the item a 'fair' test of sufficiency judgement.

Now consider a stem where the symmetry breaks. 'A positive integer n is multiplied by 2. Is the product greater than 50?' Statement 1: 'n > 30.' This guarantees the product exceeds 60, so the answer is always yes — sufficient. Statement 2: 'The product is a multiple of 9.' Multiplying n by 2 yields a multiple of 9 only when n is a multiple of 9/2, but n must be a positive integer, so n ∈ {9, 18, 27, 36, …} and the products are {18, 36, 54, 72, …}. Some of these products are below 50, others above, so Statement 2 is insufficient. The answer is A. The first statement looked modest; the second looked powerful; the sufficiency ranking was inverted by the integer constraint. Candidates who trusted surface impressions chose D and were wrong.

The hidden integer trap

Integer constraints are the single most underweighted feature in DS preparation. A statement that mathematically allows two values (such as x² = 9) becomes sufficient the moment the stem declares x is a positive integer (x = 3, unique). The reverse is also true: a statement that looks sufficient because it yields a unique real solution becomes insufficient the moment x is allowed to be negative, complex, or non-integer. Candidates who fail to map domain restrictions onto the algebra will repeatedly mark A or B when the correct answer is C or E. In a focused preparation block, the highest-yield drill is a 20-item set restricted to integer traps.

Rephrasing, then testing, then committing: the 90-second decision tree

The decision tree for a single DS item, compressed, runs in three phases. Phase one is rephrasing: rewrite the stem as a target, identify the answer type (value, yes/no, expression, ratio), and note any domain restrictions. This phase should take no more than 15 seconds; longer means the stem was misread. Phase two is the two-statement evaluation: judge Statement 1 in isolation, then Statement 2, then both together, using a clean sufficient/insufficient verdict for each. Phase three is the answer mapping: convert the three verdicts into the correct letter using the canonical table.

For most candidates the bottleneck is phase two, where the temptation to compute a numerical answer is strongest. A DS question is never asking for the numerical answer to the stem; it is asking whether the data is rich enough to produce one. Calculating the actual value of x is often a waste of time, except as a tool to disprove sufficiency. If a single computation produces a single value, that alone is enough to declare sufficiency; if the candidate has to enumerate cases, the statement is probably insufficient.

A worked example: applying the decision tree end-to-end

Stem: 'Is x a positive integer?' Statement 1: 'x² − x is a positive integer.' Statement 2: 'x² is a positive integer.'

Phase 1: rephrase. The target is the parity/sign of the integer part of x. Domain note: x can be any real number unless restricted. Phase 2, Statement 1: if x² − x is a positive integer, can x be non-integer? Let x = 0.5: x² − x = −0.25, not a positive integer. Let x = 1.5: x² − x = 0.75, not a positive integer. Let x = 2: x² − x = 2, positive integer. Let x = 0: x² − x = 0, not positive. Let x = 3: x² − x = 6, positive integer. So x can be 2, 3, 4, … or possibly non-integer values that we have not yet ruled out. Try x = (1+√5)/2 ≈ 1.618: x² − x = 1, positive integer. So Statement 1 is insufficient. Phase 2, Statement 2: if x² is a positive integer, x can be √2, √3, 2, 3, 4, etc. The answer to 'Is x a positive integer?' can be yes or no. Insufficient. Phase 2, both together: x² − x is a positive integer and x² is a positive integer. Subtracting, x = (x²) − (x² − x) is the difference of two positive integers, hence an integer. So x is a positive integer. Sufficient. The answer is C.

The work above took roughly 90 seconds in writing, but in live exam conditions a prepared candidate would compress it to 60-75 seconds by recognising the algebraic trick (subtract the two expressions to force an integer) before exploring examples. That is the texture of DS at the high end: a small algebra move, executed cleanly, in a tight time window.

Common pitfalls and how to avoid them in GMAT Data Sufficiency

The most common pitfall is answering the wrong question. Candidates calculate the value of x, see that the statements together give x = 4, and mark C — without checking whether each statement alone is insufficient. In the canonical DS scoring, the answer to a stem that asks 'What is x?' is determined by whether the data forces a unique value. If Statement 1 alone already forces x = 4, the answer is A or D, never C. Discipline here is to read the question literally: 'Is the data sufficient?' not 'What is the answer?'

The second pitfall is forgetting the meta-question. A yes/no stem such as 'Is x > y?' must be evaluated for universal truth, not existence. A statement that produces 'Yes in scenario 1 and No in scenario 2' is insufficient, even though the candidate found a 'Yes' answer with a little effort. This is the most frequent source of the gap between practice-test subscores and real-test subscores in DS.

The third pitfall is trusting the surface of an algebraic statement. 'x² = 4' looks like a unique answer (x = 2), but in DS the candidate must check the domain. If x can be negative, the statement allows two values, and it is insufficient. Always check: is the variable restricted to positive reals, integers, or another domain? If the stem is silent, treat the statement as allowing all reals.

The fourth pitfall is time leakage through re-reading. The DS module budget is roughly 2 minutes per item, including the time to bubble in the answer. Candidates who re-read the stem three times because they did not rephrase it on the first pass are giving away 30-45 seconds per item, which compounds across the section. A 10-item rephrasing discipline at the start of preparation saves 5-7 minutes over a 60-minute module.

The fifth pitfall is option E blindness. Candidates assume the union of the two statements is always sufficient, particularly when the statements share variables. This is a test-teller trap: the union of two insufficient statements can still be insufficient, especially when the variables are constrained by an external domain (such as a count of items, a sum of angles, or a non-negativity rule). Train yourself to construct two concrete counterexamples to sufficiency, not one. If two legal scenarios satisfy all the data and produce different answers to the stem, the data is insufficient.

A self-check routine at the end of every DS item

• Re-read the stem in one breath. Did I answer the right question?
• Re-state the verdict for each statement: sufficient or insufficient?
• Map the verdicts onto the answer choice using the canonical table.
• If the answer is C, confirm that at least one scenario under Statement 1 alone gave an ambiguous result. If not, the answer is A or D, not C.
• If the answer is D, confirm that Statement 1 alone truly forces a unique value, including domain checks. If Statement 1 is only sufficient in tandem with Statement 2, the answer is C, not D.

Practice design: how to train Data Sufficiency judgement over a 4-week block

A focused DS preparation block benefits more from a small, well-chosen set of items than from volume. The structure that consistently produces gains is a 4-week cycle, with three weekly sessions of 90 minutes each, plus one full-length review. The first two weeks are spent on item-type drills: integer-trap items, yes/no items, value-target items, geometry items, and rate/work items. The third week shifts to mixed sets of 15 items, timed at 30 minutes, to simulate the section's pacing pressure. The fourth week is dedicated to error analysis: every missed item is classified by root cause (rephrasing failure, domain omission, statement-trap misread, time exhaustion) and re-attempted 48 hours later.

The diagnostic value of error analysis cannot be overstated. Candidates who score in the mid-30s on a first diagnostic almost always show a clear pattern: they are losing 60-70% of their missed items to one of three root causes, not to a broad skill deficit. Identifying the dominant cause and drilling only that cause for 7-10 days usually moves the subscore by 4-6 points, which is a meaningful jump within the GMAT scoring band.

Sample 4-week preparation timeline

1. Week 1: Rephrasing drill. 30 items, untimed, with the candidate writing the rephrased target above each stem. Goal: build a rephrasing habit that takes <15 seconds per item.
2. Week 2: Two-statement evaluation drill. 20 items per session, with the candidate forced to give a one-word verdict for each statement before any calculation. Goal: build the sufficient/insufficient reflex.
3. Week 3: Timed mixed sets. Three sets of 15 items at 30 minutes each. Goal: lock the 90-second per-item budget and the 2-minute overall pacing.
4. Week 4: Error analysis and re-attempt. Every missed item is logged, classified, and re-attempted 48 hours later. Goal: convert the dominant root cause from a recurring failure into a controlled risk.

The numbers above are not magic; they are the cadence that fits a 4-week block. Candidates with longer timelines can spread each phase over 10-14 days, and the gains compound. Candidates with shorter timelines should prioritise the first two weeks, because rephrasing and the two-statement reflex are the load-bearing skills; the timed mixed sets and error analysis reinforce what is already there.

Beyond the section: how Data Sufficiency judgement transfers to other GMAT tasks

The cognitive habits built in DS — rephrasing, two-statement evaluation, domain awareness, sufficiency judgement — transfer cleanly to the rest of the Quant module and to the Data Insights module. The CR-style 'is the evidence enough?' question is a structural cousin of the DS prompt, and the discipline of testing sufficiency before trusting a conclusion is the same discipline that prevents over-claiming in Reading Comprehension inference questions. Candidates who master DS often find that their Data Insights subscore improves in parallel, partly because the section's two-part (DS and Multi-Source Reasoning) item types share the same answer-choice logic.

The transfer is not symmetrical, however. Reading Comprehension rewards synthesis across paragraphs; DS rewards isolation of a single piece of evidence. Candidates who over-apply DS logic to RC inference questions may under-claim, marking a passage as insufficient when the cumulative weight of three sentences is enough. The right mental model is: DS teaches you to ask 'is this enough?' RC teaches you to ask 'is this enough, taken together?' The two questions look similar; the threshold is different.

Score-translation notes

GMAT Data Sufficiency is scored as part of the Quant band, which spans roughly the 60-90 range in the new score scale. A strong DS subscore correlates with a Quant band in the upper half of that range, but the correlation is not perfect: candidates with strong DS skills but weaker Problem Solving geometry can still post a mid-range Quant band. The lesson is that DS is a multiplier on existing quant skills, not a substitute. The candidates who post the highest Quant bands are the ones who pair DS discipline with fluent Problem Solving — particularly in algebra, number properties, and arithmetic word problems.

TestPrep Europe's Data Sufficiency diagnostic set is a natural starting point for candidates building a sharper preparation plan around statement-trap logic and 90-second decision making.

Conclusion and next steps

GMAT Data Sufficiency rewards a small number of habits executed consistently. Rephrase the stem into a target before reading the statements. Evaluate Statement 1 in isolation, then Statement 2, then both together. Map the three verdicts onto the canonical answer table. Watch for integer traps, domain restrictions, and yes/no inversions. Budget 90 seconds per item on average, and spend more on hard items without re-checking easy ones. Train those habits over a 4-week block, with a deliberate mix of drills, timed sets, and error analysis. The DS subscore will follow the quality of the judgements, not the volume of calculations.

Frequently asked questions