The GRE's "Select One or More" format is uniquely punishing: you must identify every correct answer from the available choices, and there is no partial credit. Miss a single correct choice or select one extra wrong choice and you receive zero points. When this format is combined with Data Analysis content — normal distributions, conditional probability, standard deviation properties, and successive percent changes — precision becomes everything. Below you will learn the five question patterns that appear, work through two interactive walkthroughs, and then practice with questions drawn from real GRE-style problem sets.
What Is the Select One or More Format?
On the GRE Quantitative Reasoning section, most multiple-choice questions ask you to pick exactly one answer from five choices. "Select One or More" questions are different. They present square checkboxes instead of round radio buttons, and the directions explicitly state: "Indicate all such [statements/values/years]." You may need to select one, two, three, or even all of the choices.
The critical rule is that there is no partial credit. If a question has three correct answers and you identify only two, you receive zero points — exactly the same as selecting zero correct answers. This makes it essential to evaluate every single choice independently, rather than stopping once you have found "enough" correct answers.
No partial credit: You must select every correct answer and only correct answers. One missed selection or one extra wrong selection means zero points on the entire question. Always check every choice before submitting.
Five Question Patterns You Will See
Nearly every Select One or More Data Analysis question on the GRE follows one of five patterns. Recognizing the pattern immediately tells you what analytical approach to use.
1
Weighted Average Range Testing
The question introduces a weighted average with a constrained variable (such as group size within a range) and asks which values could be the overall average. You must find the range of possible averages and check which choices fall inside it.
2
Normal Distribution Probability Comparisons
Two normally distributed variables with different means and standard deviations are given. You must determine which probability statements are true by converting to z-scores and using the symmetry property P(Z>a)=P(Z<−a).
3
Statistical Sufficiency
You are given partial information about a data set (such as the range of two subgroups) and asked which additional facts would individually suffice to determine a statistic like the overall range. Each choice must be evaluated independently.
4
Conditional Probability from Tables
A two-way frequency table is provided. You must compute conditional probabilities and percentages for specific subgroups, then compare each to a threshold. Careful fraction comparison is essential.
5
Successive Percent Changes
Annual percent changes are given for consecutive years. You must determine which statements about cumulative changes are true, remembering that successive percent changes compound multiplicatively, not additively.
How to Solve MCM Data Analysis Questions
These strategies apply across all five patterns. The overarching principle is to evaluate each choice independently — do not let your assessment of one choice influence your judgment of another.
When comparing probabilities across two different normal distributions, convert each value to its z-score using z=σvalue−μ. Once in z-score form, you can directly compare probabilities using the symmetry property: P(Z>a)=P(Z<−a) for any value a. Equal z-score magnitudes mean equal tail probabilities.
For questions that ask "which of the following exceed X?" or "which satisfy condition Y?", first compute the threshold value in the same units as the data. Then check each choice against that threshold independently. Do not try to compute and compare simultaneously — this leads to errors, especially when values are close to the boundary.
A +5% change followed by a -3% change is NOT +2%. The correct calculation is 1.05×0.97=1.0185, a net increase of 1.85%. Always multiply the successive factors rather than adding the percentages. This is the single most tested concept in percent change questions.
The range requires knowing the maximum and minimum. The mean requires knowing the sum and the count. Information about one statistic does not necessarily determine another — the mean tells you nothing about the range, and the median tells you nothing about the range. For each choice, ask: does this information, combined with what is already given, uniquely pin down the requested quantity?
When a conditional probability is close to a threshold (such as 40/93 compared to 9/20), do not rely on decimal approximations. Cross-multiply: 40 x 20 = 800 versus 93 x 9 = 837. Since 800 < 837, the fraction 40/93 < 9/20. This eliminates rounding errors that can flip your answer.
Golden rule: Evaluate every choice. The most common mistake on Select One or More questions is stopping after finding two correct answers. There may be three, four, or even more. Check all of them.
Worked Example: Possible Values of the Standard Deviation
This walkthrough teaches you how to reason about the possible range of standard deviations for a bounded data set. Work through each step — you must answer correctly to advance.
Interactive Walkthrough0/5 steps
Possible Values of the Standard Deviation
A data set contains 10 values, all of which lie between 20 and 40, inclusive.
Which of the following could be the standard deviation of the data set? Indicate all that apply.
0
3
7
10
15
1
Step 1: Can the standard deviation be 0?
A standard deviation of 0 means all values are identical. Is it possible for all 10 values to equal the same number between 20 and 40?
2
Step 2: Find the maximum possible standard deviation
3
Step 3: Is SD = 10 achievable?
4
Step 4: Test the middle values
5
Step 5: Compile the answer
Worked Example: True Statements About a Data Set
This walkthrough teaches you how to compute and compare basic descriptive statistics — mean, median, mode, and range — and evaluate which statements about a data set are true.
Interactive Walkthrough0/5 steps
True Statements About a Data Set
A data set consists of the values: 3, 5, 5, 7, 8, 8, 8, 10, 12.
Which of the following statements are true? Indicate all that apply.
The mean is greater than 7
The median equals the mode
The range is 10
The mean is an integer
1
Step 1: Compute the mean
Sum =3+5+5+7+8+8+8+10+12=66. The mean is 966. Is the mean greater than 7?
2
Step 2: Find the median
3
Step 3: Find the mode
4
Step 4: Compute the range
5
Step 5: Is the mean an integer?
Practice Questions
Apply what you learned. Each question uses the "Select all that apply" format — check every choice you believe is correct, then submit. After submitting, click through the step-by-step solution to compare against your reasoning.
Question 1 — Weighted Average Range
A university has two campuses, East and West. The average GPA of students at the East campus is 3.2, and the average GPA of students at the West campus is 3.8. The East campus has between 200 and 1,000 students, inclusive, and the West campus has exactly 400 students. Which of the following could be the average GPA of all students at the university? Indicate all such values.
Question 2 — Successive Percent Changes
The annual percent change in the consumer price index (CPI) for a certain city over five consecutive years is: Year 1: +5%, Year 2: -3%, Year 3: +4%, Year 4: +2%, Year 5: -4%. If the CPI at the beginning of Year 1 was C, which of the following statements must be true? Indicate all such statements.
Transportation Survey Data
Results of a survey of 200 adults classified by age group and preferred mode of transportation
Car
Public Transit
Bicycle
Total
Under 30
18
32
20
70
30-50
35
22
13
70
Over 50
40
15
5
60
Total
93
69
38
200
Question 3 — Conditional Probability from a Table
Based on the transportation survey data above, which of the following statements must be true? Indicate all such statements.
Question 4 — Quartile and IQR Properties
A data set consisting of 100 test scores has a first quartile Q1 = 62 and a third quartile Q3 = 78. Which of the following statements must be true? Indicate all such statements.
Question 5 — Dice Sum Distribution
A fair six-sided die is rolled three times, and the sum S of the three results is recorded. Which of the following statements must be true? Indicate all such statements.
Question 6 — Standard Deviation Under Transformations
A data set S has n values (n≥2) with standard deviation d, where d>0. For each of the following transformations applied to every value in S, a new data set is created. Which of the following statements about the standard deviation of the new data set must be true? Indicate all such statements.
Question 7 — Combinatorial Probability
A jar contains 5 red marbles, 4 blue marbles, and 3 green marbles. Three marbles are drawn at random without replacement. Which of the following statements must be true? Indicate all such statements.
Discrete Probability Distribution
Probability distribution of random variable X
x
0
1
2
3
4
P(X=x)
0.10
0.25
0.30
0.20
0.15
Question 8 — Expected Value, Variance, Median, and Mode
Based on the probability distribution table above, which of the following statements are true? Indicate all such statements.
Question 9 — Regression and Correlation
A researcher collected data on study hours and exam scores for 30 students. The correlation coefficient is r = 0.85, and the regression equation is y-hat = 40 + 5.2x, where x is hours studied and y-hat is the predicted score. Which of the following statements must be true? Indicate all such statements.
Question 10 — Overlapping Sets and Conditional Probability
In a company of 120 employees, 75 speak English, 55 speak Spanish, and 15 speak neither English nor Spanish. Which of the following statements must be true? Indicate all such statements.
Question 11 — Constraints from Median and Mean
A data set S consists of 11 values with a median of 40 and a mean of 38. Which of the following statements must be true? Indicate all such statements.
Question 12 — Survey Methodology and Bias
A researcher wants to estimate the average number of hours per week that adults in a city exercise. She sends an online survey to 1,000 randomly selected email addresses from the city's public directory. Of these, 200 people complete and return the survey. Which of the following are valid concerns about the methodology? Indicate all such concerns.
Common Traps to Avoid
Trap 1 — Adding successive percent changes. A +5% change followed by a -3% change is not +2%. It is1.05×0.97=1.0185, a net increase of only 1.85%. Each percent change applies to a different base. Always multiply the factors.
Trap 2 — Confusing r with r-squared. The correlation coefficient r=0.85 does not mean 85% of variability is explained. The coefficient of determination r-squared = 0.7225 does. This distinction appears frequently in MCM format questions because it is easy to create a plausible-sounding wrong choice.
Trap 3 — Strict vs. non-strict inequalities. When a choice says "greater than 1/3" and the computed value is exactly 1/3, the statement is false. The GRE uses this distinction deliberately. Read the inequality direction carefully every time.
Trap 4 — Stopping after finding two correct answers. The most common meta-error on Select One or More questions is not evaluating all choices. A question might have one, three, or even five correct answers. Never assume a number — always check every single choice.
How to Decide How Many Choices Are Correct
There is no shortcut to knowing how many choices are correct. The GRE does not follow a pattern such as "always two" or "always three." Use this table to calibrate your expectations.
Number Correct
How Common
Example Pattern
1 of 4-5 choices
Common
Only IQR is guaranteed; median, count, and outlier fence are not
2 of 3-4 choices
Very common
Two z-score equalities hold; one does not
3 of 5 choices
Common
Three conditional probabilities exceed threshold; two do not
4+ of 5-8 choices
Rare but possible
Six of eight years satisfy a graph-reading condition
The key discipline is to evaluate each choice as a standalone true/false question. Compute the relevant quantity, compare it to the stated threshold or value, and record your verdict. Only after checking every choice should you count how many you have selected.
Study Checklist
MCM Data Analysis Mastery Checklist0/8 complete
Frequently Asked Questions
How does the Select One or More format differ from standard multiple choice on the GRE?
In Select One or More questions, you must identify every correct answer from among the choices. There is no partial credit: if a question has three correct answers and you select only two, you receive zero points. The choices use square checkboxes instead of round radio buttons to signal this format. The directions will explicitly say "Indicate all such [values/statements/years]."
What data analysis topics appear most often in Select One or More questions?
The most common topics are normal distribution probability comparisons (using z-scores and symmetry), conditional probability from two-way tables, properties of standard deviation under linear transformations, successive percent changes (compounding), and statistical sufficiency for determining measures like range, mean, or median.
How many correct answers should I expect in a Select One or More question?
There is no fixed number. A question might have one correct answer, two, three, or even all choices correct. In official GRE materials, the number of correct answers ranges from one to six out of three to eight total choices. Never assume a specific count — evaluate each choice independently against the conditions stated in the question.
What is the z-score method for normal distribution questions?
Convert each value to a z-score using z=σvalue−μ. This standardizes values from any normal distribution onto the same scale. Then use the symmetry property: P(Z>a)=P(Z<−a). Two probabilities with equal z-score magnitudes are equal, regardless of the original means and standard deviations. This method reduces complex probability comparisons to simple arithmetic.
What is the most common mistake on Select One or More Data Analysis questions?
The most common mistake is selecting too few answers. Students often find two correct choices and stop looking, missing a third or fourth correct choice. Because there is no partial credit, this means zero points. The second most common mistake is confusing r with r-squared in regression questions, or adding successive percent changes instead of compounding them. Always evaluate every single choice before submitting your answer.