GMAT Data Sufficiency strategy: the AD/BCE method and 7 traps

Updated 24 September 2026 · 8 min read

Data Sufficiency is the one GMAT question type you will not have seen at school. You get a question and two statements, and instead of solving the problem you decide whether the information is enough to solve it. On the GMAT Focus Edition it sits in the Data Insights section, which has 20 questions in 45 minutes and an on-screen calculator.

Most Data Sufficiency mistakes are not maths mistakes. They come from a sloppy process: mixing the statements up, assuming things the question never said, or stopping as soon as an answer looks possible. This guide gives you a fixed routine, the AD/BCE method, and then walks through the seven traps that cost test-takers the most points, each with one of our own worked examples.

How Data Sufficiency works

Every Data Sufficiency question has the same structure: a question stem, statement (1), statement (2), and the same five answer choices.

Because the five choices never change, learn them now so you never have to reread them on test day.

Two ground rules:

• The statements are true. You never question them; you only ask whether they are enough.
• "Sufficient" means exactly one answer. For a question that asks for a value, sufficient means one value. For a yes/no question, sufficient means a definite yes or a definite no.

GMAC does not publish how many of the 20 Data Insights questions are Data Sufficiency, and we could not confirm a figure, so do not plan around a fixed number. What is certain is that it is one of five Data Insights question types, alongside Multi-Source Reasoning, Table Analysis, Graphics Interpretation and Two-Part Analysis. Our Data Insights guide covers all five.

The AD/BCE method, step by step

The AD/BCE method turns the five choices into a two-branch decision so you never pick an impossible answer.

Step 1: Understand the question before you read the statements. Is it a value question ("What is $x$?") or a yes/no question ("Is $x > 3$?")? Rephrase it into its simplest form. "Is $2x + 6$ divisible by 4?" might become "Is $x + 3$ even?", which is "Is $x$ odd?" if $x$ is an integer. Note any constraints: is $x$ an integer? Positive? Different from $y$?

Step 2: Judge statement (1) alone.

• If (1) is sufficient, the answer can only be A or D.
• If (1) is not sufficient, the answer can only be B, C or E.

Step 3: Judge statement (2) alone, forgetting statement (1) completely.

• On the AD branch: (2) sufficient means D; not sufficient means A.
• On the BCE branch: (2) sufficient means B; not sufficient means go to step 4.

Step 4 (BCE branch only): Combine the statements. Together sufficient means C; still not sufficient means E.

You can start with statement (2) if it looks easier; the grid simply flips to BD/ACE. The point is to make a decision about one statement, lock the branch, and move on.

Why this helps with timing. Data Insights gives you about 2 minutes 15 seconds per question on average. If you get stuck on statement (2) after deciding (1), you have already narrowed the answer to two choices, which makes an educated guess far better than a random one.

Value questions vs yes/no questions

Value questions ask for a single number. A statement is sufficient only if it pins down exactly one value. Two possible values means not sufficient, even if both look reasonable.

Yes/no questions ask whether something is true. A statement is sufficient if it gives an answer that is always yes or always no. "Always no" is just as sufficient as "always yes". It is insufficient only when some allowed cases say yes and others say no.

For yes/no questions, the fastest proof of insufficiency is a pair of numbers: one that satisfies the statement and gives yes, and one that satisfies the statement and gives no. When you test numbers, try deliberately different kinds: a positive and a negative, zero, a fraction, a large number.

The 7 Data Sufficiency traps

Trap 1: Letting statement (1) leak into statement (2)

When you judge statement (2) alone, everything you learned from statement (1) is off limits. This sounds obvious, but after two minutes of working with (1) it is easy to assume, for instance, that $x$ is still positive. The fix is mechanical: when you move to (2), cover (1) with your hand on the whiteboard, and rewrite the question with only (2) next to it. You combine the statements only in step 4.

Trap 2: Assuming what the question does not say

Unless the stem says so, a variable can be negative, zero or a fraction, and two variables can be equal. Many "sufficient" statements stop being sufficient the moment you try $x = -2$ or $x = \tfrac{1}{2}$. Before judging any statement, reread the stem and list its actual constraints. If it says "$n$ is an integer", use that. If it says nothing, test non-integers.

Trap 3: Forgetting the second root

Squares, absolute values and even powers hide a second solution. $x^2 = 9$ gives $x = 3$ or $x = -3$. If the question asks for a single value, a squared equation is usually not sufficient on its own unless something else rules out one root. Here is one of our own questions where this trap decides the answer:

Trap 4: Stopping at "an answer exists" instead of "only one answer"

A statement can be consistent with the scenario without determining the answer. This is especially common with inequalities and integer word problems: the statement allows a solution, you find one, and you stop. But sufficiency needs one answer. Always ask, "Could it also be something else?"

In this example, statement (2) allows several different numbers of lattes, all with more lattes than teas. Finding one valid combination is not enough; the second valid combination is what proves insufficiency.

Trap 5: Treating "no" as insufficient

On a yes/no question, a statement that proves the answer is always no is sufficient. Test-takers who expect a yes reject a sufficient statement because the answer came out negative. Train yourself to ask "Is the answer definite?", not "Is the answer yes?".

Trap 6: Confusing a relationship with a value

A ratio, a percentage or a difference tells you how quantities relate; it does not always tell you the quantities. "Blue marbles are 60% of the box" gives the ratio of blue to red but not the count. The reverse trap is just as common: a fixed difference looks as if it should fix a proportion, but it does not. "6 more litres of Q than of P" is a 7-to-1 mix if the total is 8 litres and much closer to 1-to-1 if the total is 100 litres.

Trap 7: Missing an identity that makes a statement say nothing

Some statements look informative but are true for every allowed number, so they add nothing. The product of two consecutive integers is always even, so "$n^2 + n$ is even" is true for every integer $n$ and tells you nothing about whether $n$ is even. When a statement seems too convenient, test it on several numbers of different types before trusting it.

A related trap is the "C trap": a hard-looking question where the two statements obviously work together tempts you to choose C. Before you do, check each statement alone properly. Many questions that look like C are really A, B or D.

A 2-week drill plan for Data Sufficiency

If Data Sufficiency is costing you points, work on it narrowly rather than mixing it into general Quant practice.

• Days 1–3: Process only. Do easy and medium questions and write the AD/BCE branch after every statement, even when the answer seems obvious. Accuracy first, no timer.
• Days 4–7: One topic at a time. Pick the topics that cost you most, for example inequalities, number properties or word problems, and do a short set on each. Log every miss by trap type: leak, assumption, second root, "one answer", "no" answer, relationship vs value, identity.
• Days 8–11: Timed sets. Add a timer at about 2 minutes 15 seconds per question. If you pass the target, lock the branch you are on and guess between its choices.
• Days 12–14: Review. Redo every question in your log without looking at the old work. If a trap type still appears, give it another short set.

Our error log guide explains how to keep the log, and the study plan guide shows how to fit a block like this into a longer plan. For a first set, the free Data Sufficiency practice page has five questions with explanations. Data Sufficiency uses the same arithmetic and algebra as Problem Solving, so a weak topic in one usually shows up in the other. The GMAT Focus Quant topics list covers every topic both draw on.