Last updated July 07, 2026 · Reviewed by the AnvayaPrep team
Introduction
The Statistics and Probability unit covers the quantitative tools used to summarize data sets and measure uncertainty. Topics include mean, median, mode, range, standard deviation, graphs, tables, data interpretation, scatterplots, probability, independent events, conditional probability, expected value, permutations, combinations, the counting principle, sets, Venn diagrams, statistical reasoning, and weighted averages. This unit accounts for approximately 15-20% of GMAT Quantitative Reasoning questions.
The GMAT tests statistics conceptually more than computationally. Questions rarely ask for a manual standard deviation calculation; they test whether the student understands what standard deviation means when data is transformed. Questions about mean and median rarely involve more than four or five data points; they test the relationship between these measures in skewed distributions. Probability questions test logical structure (independent vs. dependent events, "at least one" vs. "exactly one") rather than complex numerical computation.
Combinatorics (permutations and combinations) is the most computationally intensive subunit and also one of the most precisely worded. The difference between "choose 3 people from 10" and "arrange 3 people from 10 in order" changes the calculation entirely. Developing sensitivity to problem wording is as important as knowing the formulas.
This unit connects directly to Data Insights: the statistics and graphical analysis skills here overlap with Table Analysis and Graphics Interpretation question types. Strong performance in this unit improves performance across two GMAT sections simultaneously.
Learning Objectives
- Calculate mean, median, mode, range, and weighted average for data sets and determine how adding or removing values affects each measure
- Explain what standard deviation measures, compare standard deviations across data sets without computation, and determine how transformations (adding a constant, multiplying by a constant) affect standard deviation
- Compute basic probability as favorable outcomes divided by total possible outcomes, and distinguish between independent and dependent events
- Calculate compound probability for multiple events using multiplication (AND) and addition (OR) rules, accounting for dependence and mutual exclusivity
- Apply the complement rule (P(at least one) = 1 - P(none)) to simplify probability problems
- Use the counting principle, permutation formula, and combination formula appropriately based on whether order matters
- Solve set problems using Venn diagrams and the inclusion-exclusion principle
- Interpret data from tables, graphs, and scatterplots, including identifying trends, outliers, and correlation direction
- Apply conditional probability P(A|B) = P(A and B) / P(B) to problems involving restricted sample spaces
High-Yield Concepts
| Topic | Core Rule or Formula | Common GMAT Trap | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Mean | Sum of values / number of values; Total = Mean x Count is more useful algebraically | Averaging group averages when group sizes differ (requires weighted average) | ||||||||
| Median | Middle value when sorted; for even count, average the two middle values | Assuming median equals mean; they are equal only in symmetric distributions | ||||||||
| Standard Deviation | Measures spread around the mean; adding a constant to all values does NOT change SD; multiplying all values by c multiplies SD by | c | Assuming adding a constant to all values changes standard deviation | |||||||
| Probability | P(Event) = favorable outcomes / total outcomes; 0 <= P <= 1 | Not accounting for "without replacement" in dependent event problems | ||||||||
| Independent Events | P(A and B) = P(A) x P(B); the occurrence of A does not affect P(B) | Applying independent event multiplication when events are actually dependent | ||||||||
| Complement Rule | P(at least one) = 1 - P(none); P(A) = 1 - P(not A) | Computing "at least one" by adding individual probabilities instead of using the complement | ||||||||
| Combinations | C(n,r) = n! / (r! x (n-r)!); order does not matter | Using permutation formula when order does not matter | ||||||||
| Permutations | P(n,r) = n! / (n-r)!; order matters | Using combination formula when order does matter | ||||||||
| Sets / Venn Diagrams | A union B | = | A | + | B | - | A intersect B | Forgetting to subtract the overlap when computing union | ||
| Expected Value | E = sum of (value x probability); used for long-run average outcomes | Confusing expected value with the most likely single outcome |
For any "at least one" probability problem, the complement approach is almost always faster than direct calculation. Instead of summing P(exactly 1) + P(exactly 2) + ... + P(exactly n), compute 1 - P(none). This reduces a multi-term sum to a single multiplication chain.
Adding a constant to every value in a data set (e.g., adding 5 to every score) shifts the mean by that amount but does NOT change the standard deviation or range. Multiplying every value by a constant multiplies the mean, the standard deviation, and the range all by that same constant. Many students incorrectly assume that adding a constant also increases the spread.
Study Strategy
Start with mean, median, mode, and range, because these are the most frequently tested individual statistics and they also appear within Data Insights questions. Develop intuition about when mean and median diverge: in a right-skewed distribution (long tail to the right), the mean exceeds the median; in a left-skewed distribution, the median exceeds the mean. The GMAT tests this relationship in Data Sufficiency questions.
Probability is best studied in order of complexity: basic probability first, then independent events, then dependent events, then the complement rule, then conditional probability. At each step, drill the distinction from the previous level. The most common GMAT probability error is applying the independent event multiplication rule to events that are actually dependent (sampling without replacement).
For combinatorics, master the conceptual distinction between combinations and permutations before learning the formulas. Ask "does order matter?" for every problem. Language clues: "arrange," "sequence," "order," "rank" point to permutations; "select," "choose," "committee," "group," "team" point to combinations.
Data interpretation (tables, graphs, scatterplots) requires no new formulas; it requires careful reading and percent/ratio arithmetic. Practice reading axes carefully, checking units, and computing percent changes from graph values.
Common Mistakes
Mean vs. median: Assuming that knowing the mean provides information about the median or individual values. In Data Sufficiency, a statement giving the mean of a set of five numbers is insufficient to determine the median.
Weighted average: Averaging the averages of two groups without accounting for different group sizes. If Group A (10 people) has a mean of 80 and Group B (5 people) has a mean of 70, the combined mean is (800 + 350) / 15 = 76.67, not (80 + 70) / 2 = 75.
Standard deviation: Confusing range with standard deviation. A set with range 10 could have high or low standard deviation depending on how the values are distributed within that range.
Probability: Forgetting that drawing cards "without replacement" makes the events dependent. After drawing one card from a 52-card deck, only 51 cards remain, changing all subsequent probabilities.
Combinations vs. permutations: Using C(n,r) when the problem involves ordered selections, or using P(n,r) when the problem involves unordered groups. The ratio between the two answers is r!, which is a significant numerical difference for any r greater than 2.
Venn diagrams: Not subtracting the intersection when computing the total in a two-set problem. If 30 people like A, 20 like B, and 10 like both, the total who like A or B is 30 + 20 - 10 = 40, not 50.
Exam Tips
For Data Sufficiency statistics questions, establish exactly what each statement tells you. A statement giving the mean does not determine the individual values or the median. A statement giving the range does not determine the mean or standard deviation. A statement giving the number of elements and the sum uniquely determines the mean but not the distribution.
Expected value problems typically involve a game or investment scenario. Multiply each possible outcome by its probability and sum the products. If the expected value is positive, the situation is favorable in the long run; if negative, it is not. The GMAT sometimes asks whether it is "worth it" to take an action, which is an expected value comparison.
Scatterplot questions test whether you can identify the direction and strength of a correlation, identify outliers, and estimate the line of best fit. A positive correlation means both variables tend to increase together; a negative correlation means one increases as the other decreases. The GMAT does not require computation of correlation coefficients.
For probability questions involving "exactly k successes" in repeated trials, the binomial probability formula applies: P = C(n,k) x p^k x (1-p)^(n-k). This formula is rarely needed explicitly, but recognizing its structure helps set up problems where you need to count favorable outcomes among repeated independent trials.
For combinations and permutations: C is for "Choosing" (order does not matter), P is for "Placing in order" (order does matter). The combination formula divides the permutation formula by r! (the number of ways to arrange r items). If you can never remember which is larger: permutations are always greater than or equal to combinations for the same n and r.
Sign up free to keep reading
Create a free AnvayaPrep account to finish this GMAT guide on Statistics and Probability — plus flashcards and practice questions.