Data Sufficiency is one of the most unusual question types on any standardized test. It doesn't ask you to solve a problem — it asks you whether a problem CAN be solved with given information. This distinction trips up even strong math students. Here's the complete framework for mastering it.
The Format
Every DS question has the same structure:
Question stem: A question about a mathematical relationship (e.g., "What is the value of x?" or "Is y > 0?")
Statement 1: A piece of information
Statement 2: Another piece of information
- (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) Statements 1 and 2 together are not sufficient
These five answer choices never change. Memorize them until they're automatic — you should never waste time re-reading them during the test.
The AD/BCE Decision Tree
This is the systematic approach that eliminates guesswork:
Step 1: Evaluate Statement 1 ALONE (ignore Statement 2 completely)
Pretend Statement 2 doesn't exist. Can you answer the question using only Statement 1?
- If YES → You're in the AD branch (Answer is A or D)
- If NO → You're in the BCE branch (Answer is B, C, or E)
Step 2A: If in AD branch → Evaluate Statement 2 ALONE (ignore Statement 1)
Now pretend Statement 1 doesn't exist. Can you answer the question using only Statement 2?
- If YES → Answer is D (each statement alone is sufficient)
- If NO → Answer is A (only Statement 1 is sufficient)
Step 2B: If in BCE branch → Evaluate Statement 2 ALONE (ignore Statement 1)
Can you answer the question using only Statement 2?
- If YES → Answer is B (only Statement 2 is sufficient)
- If NO → Go to Step 3
Step 3: Evaluate both statements TOGETHER
Combine the information from both statements. Can you answer the question now?
- If YES → Answer is C (together but not alone)
- If NO → Answer is E (not sufficient even together)
The Two Types of DS Questions
"Value" Questions
These ask: "What is the value of X?" Sufficiency requires that you can determine exactly one value for X. If a statement gives you two possible values (e.g., x = 3 or x = -3), it's NOT sufficient.
Example:
> What is the value of x?
> (1) x² = 9
> (2) x > 0
- Statement 1 alone: x = 3 or x = -3. Two values → NOT sufficient (BCE branch)
- Statement 2 alone: x could be anything positive → NOT sufficient
- Together: x² = 9 AND x > 0 → x = 3. Exactly one value → Answer: C
"Yes/No" Questions
These ask: "Is [something] true?" Sufficiency requires a definitive answer — always yes OR always no. A statement is sufficient even if the answer is "no" — as long as it's definitively "no."
A statement is NOT sufficient if sometimes the answer is yes and sometimes no.
Example:
> Is x > 5?
> (1) x > 3
> (2) x > 7
- Statement 1 alone: If x = 4, answer is no. If x = 6, answer is yes. Not definitive → NOT sufficient (BCE branch)
- Statement 2 alone: x > 7, so x is always > 5. Definitive yes → sufficient → Answer: B
Critical insight: For yes/no questions, "always no" IS sufficient. If a statement proves that x is NEVER greater than 5, that's a definitive answer — the statement is sufficient.
The Five Most Common DS Traps
Trap 1: Solving When You Should Only Evaluate
You don't need to find the value of x. You need to determine WHETHER you CAN find it. This is the most fundamental DS mistake.
- You're doing extensive calculations
- You're trying to get a numerical answer
- You're spending more than 2.5 minutes on one DS problem
The fix: As soon as you know that a unique answer CAN be determined, stop. You don't need to calculate what that answer actually is.
Example:
> What is the value of 3x + 5y?
> (1) x + y = 7
> (2) 2x + 4y = 22
You DON'T need to solve this system. You need to determine if 3x + 5y can be uniquely determined. Two independent linear equations with two unknowns? Yes, you can solve for x and y, therefore you can find 3x + 5y. Answer: C. Move on.
Trap 2: Forgetting Hidden Constraints
The question stem often contains implicit information that affects sufficiency.
- "n is a positive integer" → n ≥ 1, and n is whole
- "the number of students" → positive integer
- "the price of the item" → positive real number
- Geometry context → lengths are positive, angles are between 0° and 180°
Example:
> If n is a positive integer, what is the value of n?
> (1) n² < 5
Without the "positive integer" constraint, this gives infinitely many values. WITH it: n = 1 or n = 2. That's two values → not sufficient. But notice how close it is — this is the kind of constraint awareness that separates right from wrong answers.
Trap 3: Not Testing Multiple Cases
When evaluating sufficiency, test at least 2-3 cases to confirm your conclusion. Pick values that are as different as possible:
- Test a case you think gives "yes"
- Test a case you think gives "no"
- If both give the same answer, the statement is likely sufficient
- If they give different answers, the statement is not sufficient
- If you get one value from a test case, try to find a different value that also satisfies the statement
- If you can't find a different value, the statement is likely sufficient
Trap 4: Combining Statements Prematurely
When in the BCE branch, evaluate Statement 2 ALONE before combining. Many students jump from "Statement 1 is insufficient" directly to combining both, skipping the check of Statement 2 alone. This leads to choosing C when the answer is B.
Trap 5: Redundant Statements
Sometimes the two statements provide identical information in different forms.
Example:
> What is the value of x?
> (1) 2x = 10
> (2) x + 3 = 8
Both statements tell you x = 5. Each alone is sufficient → Answer: D.
This seems easy, but the GMAT disguises redundancy:
> (1) x - y = 3
> (2) 2x - 2y = 6
Statement 2 is just Statement 1 multiplied by 2. They provide the same information. If Statement 1 alone isn't sufficient, combining them won't help — the answer jumps from the BCE branch to E, not C.
Number Properties in DS: The Most Tested Topic
A disproportionate number of DS questions test number properties. Know these cold:
- Odd + Odd = Even
- Even + Even = Even
- Odd + Even = Odd
- Odd × Odd = Odd
- Even × anything = Even
- Positive × Positive = Positive
- Negative × Negative = Positive
- Positive × Negative = Negative
- 0 is even
- 0 is neither positive nor negative
- 0 is a multiple of every integer
- Division by 0 is undefined
- If a is divisible by b, and b is divisible by c, then a is divisible by c
- If a and b are both divisible by c, then (a + b) and (a - b) are also divisible by c
DS Practice Protocol
- The question type (value or yes/no)
- Your evaluation of Statement 1 (with test cases)
- Which branch you're in (AD or BCE)
- Your evaluation of Statement 2 (with test cases)
- If needed, your evaluation of both together
### Week 2: Building speed
Do 10 DS problems timed (2.5 minutes each). Still write out the framework, but work faster. Identify where your time is going.
### Week 3+: Mixed with other DI types
Integrate DS into full DI practice sets. By now the framework should be automatic — you shouldn't need to consciously think through AD/BCE; it should be reflex.
- Which step in the framework did you go wrong?
- Did you test enough cases?
- Did you miss a hidden constraint?
- Did you solve instead of evaluate?
- Would drawing a number line or simple diagram have helped?
The Speed Secret
- What type it is (value vs. yes/no)
- What mathematical concept is being tested
- What kinds of statements would be sufficient
This pattern recognition comes from practice. There are really only 15-20 core DS "scenarios" (number properties, inequalities, systems of equations, divisibility, etc.). Once you've seen 50+ problems of each type, you start recognizing them instantly.
DS is the most learnable question type on the GMAT. The framework works every time. The traps repeat. The math is usually not hard. What's hard is the logic — and logic is a skill you can build.