GMAT Data Sufficiency statement analysis is the single skill that separates a Data Insights candidate stuck in the high-50s from one climbing towards 80+. The question family looks deceptively small: a stem, two statements, and five answer codes. The real work happens in the way a candidate reads each statement as a self-contained claim, tests it with a concrete example, and only then opens the door to the second statement. Most wrong answers on Data Sufficiency come from a candidate who combined the statements before they had a clean ruling on statement 1, or who treated "sufficient" as a feeling rather than a verdict you can defend with one or two numbers.
This article walks through the structural reading that Data Sufficiency actually demands. We will unpack the five answer codes, then build a two-pass protocol: pass 1 isolates statement 1, pass 2 layers statement 2 on top of a clean first verdict. We will translate the most common wordy prompts into working equations, practise the case 1 / case 2 bookkeeping habit that prevents premature certainty, and finish with a daily drill sequence that hardwires the statement-by-statement habit before test day.
What GMAT Data Sufficiency is really asking: a structural reading of the stem
Data Sufficiency items are short, but the stem is doing more work than it appears. Every prompt contains a question, almost always ending in a question mark, and an information gap the statements are meant to fill. The standard prompts are recognisable: "What is the value of x?", "What is the value of y?", "Is x greater than y?", "Is xy positive?", "What is the value of the integer n?", "What is the average of a, b and c?". The shape of the question tells you what the statements have to deliver, and that shape is the first thing to read.
If the question asks for a value, sufficiency means the statements together pin down exactly one number. If the question asks a yes/no, sufficiency means the statements together always give the same yes or the same no. The mistake most candidates make is to start plugging numbers before they have decided which kind of verdict they are collecting. For a value question, finding a single value is not enough; you have to be sure no other value is also possible. For a yes/no question, one consistent yes and one consistent no already means insufficiency, even if the yes was reached first.
On the GMAT Focus edition, Data Sufficiency lives inside the Data Insights section, mixed with tables, graphs, and multi-source reasoning items. The questions still follow the same five-code answer key that has defined Data Sufficiency for years, and the time pressure is real: roughly two minutes per item, sometimes less when the section's adaptive logic routes you to a harder module. A clean reading of the stem buys you the extra ten seconds per question that compounds across 20 items.
The stem also carries signals about which topic the test-maker is leaning on. A prompt with a percent sign, a ratio colon, and the word "of" usually hides a weighted average or a mixture problem. A stem that mentions consecutive integers, distinct prime factors, or the word "integer" is a permission slip to bring parity, sign, and divisibility cases into your analysis. Read the stem twice before you touch the statements: once for the question shape, once for the hidden constraint the topic is signalling.
The five answer codes, decoded in plain language
Every Data Sufficiency item resolves into one of five codes. The codes look like a maze until you translate them into plain rulings. Once the language is plain, the codes start to feel like a checklist rather than a guessing game.
- Statement 1 alone is sufficient, but statement 2 alone is not. Read as: S1 settles the question on its own. S2 cannot.
- Statement 2 alone is sufficient, but statement 1 alone is not. Mirror image of the first code.
- Both statements together are sufficient, but neither statement alone is sufficient. Neither statement is enough by itself; the pair is.
- Each statement alone is sufficient. S1 settles it; S2 also settles it, independently.
- The two statements together are still not sufficient, and additional data is needed. This is the catch-all that punishes any candidate who declared victory too early.
The first two codes force you to compare the statements. The third and fourth codes force you to test the pair. The fifth is your safety net: it is correct more often than nervous candidates believe, especially on the harder items in a high-level adaptive module. A useful habit is to decide, before you look at the answer choices, which of the five codes you expect. If your pre-choice was code 3 and the only viable answer is code 5, do not switch on a hunch. Re-read the stem, re-test the pair, and reissue the verdict.
Many candidates lose points because they let an answer code change the question. The codes are bookkeeping, not hints. The question lives in the stem. The codes only record what the statements did to that question.
Pass 1: isolating statement 1 before statement 2 has any voice
The single most important habit in GMAT Data Sufficiency statement analysis is the discipline of pass 1. Read statement 1. Cover statement 2 with your hand, your notes, or a piece of paper. Decide, on statement 1 alone, whether the question is settled. Only then move to statement 2.
The reason this habit is non-negotiable is that statement 2 is psychologically loud. It carries new symbols, new numbers, new vocabulary, and the brain wants to weave it into statement 1 immediately. The instant you weave, you lose the ability to rule on either statement cleanly, and the answer codes 1 and 2 stop being reachable. Roughly four out of ten Data Sufficiency items on a typical GMAT Focus section resolve to codes 1 or 2. You cannot afford to forfeit them by reading too fast.
Concretely, pass 1 means three steps. First, translate statement 1 into a single mathematical claim. "x is a positive integer" is one claim. "x is a positive integer less than 10" is one claim. "x is a positive integer less than 10 and greater than 3" is also one claim, just tighter. Second, test the claim with two cases that differ enough to expose any remaining freedom. For a value question, pick two admissible values; if both lead to the same answer, statement 1 is sufficient; if they lead to different answers, statement 1 is not sufficient. Third, write down your ruling before reading statement 2: S1 = sufficient, or S1 = not sufficient. A one-line ruling written on the scratch pad is a stronger memory anchor than a thought in the head.
A common counter-argument from experienced candidates is that the algebra often spans both statements, so isolating feels artificial. The counter-counter-argument is that on every item where the algebra spans both, statement 1 alone almost always leaves at least one free variable, and statement 2 closes the gap. Pass 1 still works; you just record that the free variable is, say, the sign of y, and let pass 2 decide whether statement 2 pins it down.
Time budget for pass 1: aim for 30 to 45 seconds on a typical item. Less, and you have not stress-tested the statement with two cases. More, and you are likely constructing a full model that belongs in pass 2.
Pass 2: layering statement 2 onto a clean statement 1 verdict
Once you have a pass 1 ruling, pass 2 begins. The discipline here is symmetric: do not re-read statement 1 in detail; trust your pass 1 ruling. Read statement 2, translate it into a single claim, and ask the two question types in sequence.
Question A: does statement 2 alone settle the question? If yes, you are in code 1 or code 4 territory; finish the item by comparing to your pass 1 ruling. If no, move to question B. Question B: do the two statements together settle the question? If yes, you are in code 2 or code 3 territory. If no, the answer is code 5.
Pass 2 also demands the same case-testing habit as pass 1. The two cases you test now should specifically target the gap that pass 1 left open. If pass 1 pinned down the magnitude of x but not its sign, pass 2 should test the positive and negative branches of statement 2 against the pinned magnitude. If pass 1 gave you the sum of two numbers but not the individual values, pass 2 should test a swap that keeps the sum constant. Two cases is the floor; three is the ceiling. The aim is not to enumerate every case but to expose any remaining freedom that would let a different answer slip in.
For yes/no questions, the same pass 2 logic applies with one adjustment. The two cases you test must end on opposite verdicts. A single yes is not enough. If you can find even one admissible no, statement 2 (or the pair) is not sufficient. In my experience, this is the most consistent source of silent point loss on value questions rephrased as yes/no. Candidates find a clean yes, declare sufficiency, and miss the fact that a different combination of values would yield a clean no.
Sufficiency versus necessity: the distinction that decides statement-only items
Sufficiency and necessity are not synonyms, and Data Sufficiency tests the difference relentlessly. A statement is sufficient when, on its own, it settles the question. A statement is necessary when the question cannot be true without it. The item type you are answering cares only about sufficiency, even when the wording sounds like it is asking for necessity.
A standard trap reads: "If x is a positive integer, is x divisible by 6?" Statement 1 says x is divisible by 3. Statement 2 says x is even. The trap answer is "both statements together are sufficient" because divisibility by 6 requires both conditions. But the question is whether divisibility by 6 is true, not whether 3 and even are both required. Each statement on its own is not enough, but the pair does settle the question: if x is divisible by 3 and x is even, then x is divisible by 6. The pair is sufficient. That is the correct ruling, and it requires separating the necessary conditions from the sufficient combination.