The useful way to practise GMAT questions is not to collect a large number of answers. It is to identify the decision each question requires, set up the information correctly, eliminate tempting wrong options, and finish with a quick sense-check.
The questions below cover quantitative reasoning, verbal reasoning and data interpretation. Answer each one from the screen, then read the explanation immediately afterwards. Record not only whether you were right, but why you chose an option and how long the reasoning took.
A MySummaries question screen for this set would look like this:
A shop increases the price of an item by 20% and then applies a 20% discount to the increased price. The final price is what percentage of the original price?
The correct answer is 96%. If the original price is 100, the increase gives 120 and the discount removes 24, leaving 96. The strongest distractor is 100%, which incorrectly treats a 20% increase and a 20% decrease as cancelling; the second percentage is applied to 120, not to 100.
The first question tests whether you apply percentages in sequence rather than combine them as if they used the same base. The efficient set-up is to use 100 as the original value.
If x and y are positive numbers, x/y = 3/5 and x + y = 64, what is the value of y - x?
The ratio has 8 equal parts because 3 + 5 = 8. Each part is 8, so x = 24 and y = 40; therefore y - x = 16. The strongest distractor is 24, which is y - x only if the ratio parts are incorrectly read as the values themselves or if x is omitted from the subtraction.
This is a ratio question, so avoid introducing unnecessary algebra. Translating the ratio into equal parts is faster and reduces the chance of reversing x and y.
A committee of 3 people is selected from 5 engineers and 4 designers. How many committees contain at least one designer?
Count all committees and subtract committees made entirely of engineers: C(9,3) - C(5,3) = 84 - 10 = 74. The strongest distractor is 80, which subtracts the wrong value by treating the five engineers as five possible three-person committees instead of using combinations.
“At least one” often signals a complement. Counting the excluded case is shorter than separately counting one, two and three designers.
A project has four tasks. Task A must finish before Task B can start. Task C must finish before Task D can start. The two chains can be completed independently. If A takes 3 days, B takes 5 days, C takes 4 days and D takes 2 days, what is the minimum completion time?
The A-B chain takes 3 + 5 = 8 days, while the C-D chain takes 4 + 2 = 6 days. Because the chains run independently, the project finishes when the longer chain finishes: 8 days. The strongest distractor is 9 days, which usually comes from adding an extra day between dependent tasks without any stated delay. The correct answer is actually option 1, 8 days.
This question contains a useful checking discipline: calculate each dependency chain, then compare them. Notice that the answer index should match the option text; an explanation that contradicts its own selected option is a marking error to catch before trusting the result.
A company surveyed 200 customers. Of these, 120 used product X, 90 used product Y, and 50 used both products. How many customers used neither product?
Customers using at least one product = 120 + 90 - 50 = 160, because the 50 who used both were counted twice. Therefore 200 - 160 = 40 used neither. The strongest distractor is 20, which results from adding 120 and 90 and subtracting both-product users twice.
The key operation is interpreting overlap. In data questions, write down what each number represents before calculating. “Both” is not a separate group to remove twice; it is the overlap already included in each product total.
A researcher wants to determine whether a new training programme increases employee productivity. Which finding would most weaken the conclusion that the programme caused the increase?
The correct answer is that participants were already the highest-performing employees and volunteered themselves. This creates selection bias: the difference may reflect pre-existing performance rather than the programme. The strongest distractor is the greater increase among participants, but without comparable starting points or random assignment that difference does not establish causation; it is evidence to investigate, not a control for bias.
This tests causal reasoning rather than vocabulary. Separate an observed association from a reason that rules out an alternative explanation. A familiar trap is choosing an option that sounds supportive of the conclusion instead of one that weakens its causal basis.
All analysts in a firm have completed statistics training. Some people who have completed statistics training work remotely. Which statement must be true?
The statement says that some statistics-trained people work remotely, so it must be true that some remote workers have completed statistics training. The strongest distractor is that some analysts work remotely. The overlap described is between trained people and remote workers; it does not say that any of those remote workers are analysts.
For logic questions, keep the direction of each statement intact. “All analysts are trained” does not mean “all trained people are analysts”. “Some trained people are remote” does not identify their occupation.
A data set has five values with mean 18. If one value, 26, is removed, what is the mean of the four remaining values?
The original total is 5 × 18 = 90. Removing 26 leaves 64, and 64 ÷ 4 = 16. The correct answer is option 1, not option 3: the option list places 16 at index 1. This exposes an answer-key consistency check that should be performed whenever a calculation and selected option disagree.
Do not calculate the original values individually. Mean multiplied by the number of values gives the total, so removing one value is direct. This is also a reminder that careless checking can turn sound reasoning into a wrong submission.
A reading passage argues that a city should replace some car-parking spaces with bicycle storage. Which fact, if true, most strengthens the argument?
The strongest support is that secure storage is the main barrier for residents who might otherwise cycle. It connects the proposed change directly to increased cycling. The strongest distractor is that the city has more cars than bicycles: that describes the current situation but does not show that bicycle storage would change behaviour or meet a significant need.
Argument questions reward a close match between the conclusion and the evidence. The best strengthening fact usually addresses the mechanism: why would the proposed action produce the claimed result?
A machine produces 240 units in 8 hours. After a setting is changed, its production rate increases by 25%. At the new rate, how many units will it produce in 6 hours?
The original rate is 240 ÷ 8 = 30 units per hour. A 25% increase gives 37.5 units per hour, and 6 hours gives 225 units. The correct answer is option 2, not option 3: 225 is the third option when counting from one, or index 2 when counting from zero. The strongest distractor is 240, which applies the rate increase but fails to adjust the number of hours.
This combines rate, percentage and time. Keep the units visible: units per hour multiplied by hours must produce units. If the numerical explanation and answer index disagree, treat that as a quality-control issue rather than accepting the screen uncritically.
How to review the set
Score the questions by the reason for the error, not only by right or wrong:
- Set-up error: you misunderstood the given information or the question demand.
- Method error: you chose a longer or unsuitable route.
- Elimination error: you accepted a tempting option without testing it.
- Calculation error: the approach was sound but the arithmetic failed.
- Pacing error: you spent too long before committing to a defensible answer.
A short written response can also expose whether your reasoning is complete. For a GMAT-style data question, a useful response states the given values, identifies the operation, shows the calculation and checks whether the result fits the units.
A marked response might look like this:
A survey of 300 customers finds that 180 use service A, 140 use service B, and 60 use both. Explain how many use neither service.
7/9I add 180 and 140 to get 320, then subtract the 60 who use both, giving 260 who use at least one. Therefore 40 use neither service.
The set-up and conclusion are correct, and the overlap is subtracted once. The response gives a defensible calculation under time pressure.
Missed
−State explicitly that 300 is the total surveyed and that 260 is the number using at least one service.
Model answerAdd the service totals and subtract the overlap once: 180 + 140 - 60 = 260 customers using at least one service. Subtract from the total: 300 - 260 = 40 customers using neither service. The overlap correction prevents the 60 customers using both services from being counted twice.
After reviewing, rank the topics by marks lost and repeat the relevant question type under a tighter time limit. Keep a small error log with the original choice, the correct choice, the trap, and the shortest reliable method.
The most useful final step is to turn repeated errors into short retrieval prompts rather than rereading explanations. For example: “When does an overlap count get subtracted?” or “What does a percentage change use as its base?”
You lost this mark twice: when two groups overlap, which quantity is subtracted once from the sum, and why?
How MySummaries helps
MySummaries can turn your own GMAT notes and completed question reviews into a revision board, then generate flashcards, timed written practice and audio explanations. Its question review is most useful when you label the error precisely: wrong set-up, inefficient method, weak elimination, careless calculation or poor pacing. Start at portal.mysummaries.app.