GMAT Data Sufficiency: Necessary vs Sufficient

In GMAT Data Sufficiency, a statement is sufficient if it lets you reach exactly one answer to the question, and necessary if the question cannot be answered without it. The five answer choices only ask about sufficiency, but the two ideas explain every choice: a statement can be sufficient without being necessary (the other statement works too), or necessary without being sufficient (it only works when combined with the other).

Getting that distinction straight is most of the battle in Data Sufficiency. This guide explains it with original worked examples and shows how to use it under time pressure.

Where Data Sufficiency sits on the GMAT

Data Sufficiency is one of the question types in Data Insights, one of the GMAT's three sections alongside Quantitative Reasoning and Verbal Reasoning. According to GMAC, Data Insights has 20 questions in 45 minutes, its questions may need maths, data analysis, verbal reasoning or all three, and, unlike Quantitative Reasoning, it gives you an on-screen calculator. The GMAT is computer adaptive: GMAC describes it as selecting each question for the test taker based on their ability level. You can bookmark and review as many questions as you want and change up to three answers per section.

Each Data Sufficiency question gives you a question and two statements, labelled (1) and (2). Your job is not to answer the question but to decide which information would be enough to answer it.

The five answer choices

The answer choices are always the same, so learn them before test day. In GMAC's description, they amount to:

Read the C and D rows closely, because they are the necessary-versus-sufficient question in miniature. In C, each statement is needed, yet neither does the job on its own. In D, each statement does the job on its own, so neither is needed.

Necessary vs sufficient: the logic

A condition is sufficient for an outcome if the outcome must follow whenever the condition holds. A condition is necessary if the outcome cannot happen without it. Being a square is sufficient for being a rectangle but not necessary, since plenty of rectangles are not squares. Having four sides is necessary for being a rectangle but not sufficient, since many four-sided shapes are not rectangles.

In Data Sufficiency, the outcome is "the question has exactly one answer". What that means depends on the question:

• A value question ("What is the value of x?") needs a single value. If the information allows two values, it is not sufficient, however close you got.
• A yes/no question ("Is x > 0?") needs a definite answer. "Always yes" is sufficient and so is "always no". Only "sometimes yes, sometimes no" is insufficient.

That second point is easy to miss: a statement that proves the answer is "no" is just as sufficient as one that proves it is "yes".

Four worked examples

Example 1: necessary but not sufficient (answer C)

Question: What is the value of x?

• (1) x² = 9
• (2) x > 0

Statement (1) allows x = 3 or x = −3: two values, not sufficient. Statement (2) allows any positive number: not sufficient. Together, x must be 3, so the answer is C. Each statement is necessary, since without either one the value is not fixed, but neither is sufficient alone.

Example 2: sufficient but not necessary (answer D)

Question: What is the value of n?

• (1) 2n + 3 = 11
• (2) n² − 8n + 16 = 0

Statement (1) gives n = 4: sufficient. Statement (2) is (n − 4)² = 0, which also gives only n = 4: sufficient. Each statement works alone, so the answer is D, and neither statement is necessary. Notice that (2) is a quadratic with only one solution; seeing a squared term is not a reason to assume two answers.

Example 3: a definite answer from one statement (answer A)

Question: Is the integer k even?

• (1) 3k is even.
• (2) k² + k is even.

Statement (1): if k were odd, 3k would be odd, so k must be even. That is a definite "yes": sufficient. Statement (2): k² + k = k(k + 1) is the product of two consecutive integers, which is even for every integer k. It tells you nothing about k: not sufficient. The answer is A.

Example 4: two equations that are really one (answer E)

Question: What is the value of x?

• (1) x + y = 5
• (2) 2x + 2y = 10

Neither statement alone fixes x. Together they look like two equations in two unknowns, which usually means a unique solution. But (2) is just (1) multiplied by 2, so together they are still one equation, with infinitely many solutions. The answer is E. "Two equations, two unknowns" is a test you have to check, not a rule you can assume.

Six habits that make Data Sufficiency faster

1. Decide what sufficient means before reading the statements. Is the question asking for one value or a definite yes/no? Rephrase it if you can: if the question tells you that x = 2y, then "Is x > 0?" becomes "Is y > 0?".
2. Judge each statement alone first. When you move to statement (2), forget what (1) told you. Carrying information across is a classic route to a wrong C.
3. Prove insufficiency with two cases. If you can find two values that satisfy the statement and give different answers, the statement is not sufficient. Picking numbers is often the quickest way to do this.
4. Do not solve to the end. As GMAC puts it, the goal is not to calculate an answer but to judge whether the information is enough. Once you know there is exactly one value, stop.
5. Look for hidden constraints. Integers, positive quantities such as lengths or numbers of items, and squares that cannot be negative all narrow the possibilities and can turn an apparently insufficient statement into a sufficient one.
6. Weigh certainty against time. A guide published on GMAC's site makes the point that newcomers tend to overrate how sure they are, so early in your preparation it pays to prove your hunches. Later, judge whether the extra certainty is worth the minutes.

The same words in Critical Reasoning

The words necessary and sufficient also appear when you analyse an argument's assumptions in Critical Reasoning, part of the Verbal Reasoning section. The logic is identical. An assumption is necessary if the argument falls apart without it: deny it, and the conclusion no longer follows. An assumption is sufficient if adding it to the premises makes the conclusion follow. Practice with the distinction in one question type carries over to the other.

How to practise

Data Sufficiency rewards volume and review: the logic is simple, but applying it reliably to unfamiliar questions takes many repetitions. Our GMAT question bank includes Data Sufficiency within its Data Insights questions, filterable by difficulty, with the on-screen calculator available as on the exam and a worked explanation for every answer. When you review a miss, classify it: did you misjudge what sufficient meant, carry information between statements, or miss a constraint?

Then test the habit inside a full section. Our GMAT practice tests adapt question by question and let you change up to three answers at the end of each section, as the real exam does, and the GMAT practice overview sets out how the three sections fit together.

Frequently asked questions