Numerical Reasoning Test: What to Expect and How to Prepare
Numerical reasoning tests measure your ability to interpret data, perform calculations, and draw logical conclusions from numbers under time pressure. They are one of the most widely used psychometric assessments in hiring, appearing in graduate schemes, professional roles, and internal promotions across nearly every industry. Whether you are applying to Goldman Sachs, Deloitte, P&G, or the NHS, there is a strong chance you will face a numerical reasoning test as part of the selection process.
This guide covers everything you need to know: what the tests measure, the question types you will encounter, how scoring works, which employers use them, and a practical preparation plan you can start today.
What Does a Numerical Reasoning Test Measure?
A numerical reasoning test measures how accurately and efficiently you can work with quantitative information in a professional context. Unlike a school math exam, these tests focus on applied reasoning rather than pure computation. You are given data in the form of tables, charts, or graphs and asked to extract the right numbers, choose the correct operation, and arrive at an answer within a strict time limit.
The core skills being assessed include:
- Data interpretation -- reading and understanding tables, bar charts, line graphs, and pie charts
- Arithmetic fluency -- performing calculations with percentages, ratios, fractions, and averages
- Logical deduction -- identifying which data is relevant and which is a distractor
- Speed and accuracy -- balancing quick decision-making with precision under time pressure
Employers value these skills because they mirror real workplace tasks: analyzing sales figures, interpreting financial reports, comparing performance metrics, and making data-driven decisions. That is why companies like Amazon, Unilever, and Goldman Sachs make numerical reasoning a mandatory part of their hiring process.
💡Numerical reasoning tests are not about advanced math. They test whether you can quickly read data, pick the right approach, and calculate accurately under pressure.
Common Question Types With Examples
Numerical reasoning questions fall into several recurring categories. Knowing what to expect removes surprises on test day and lets you practice each type systematically.
Percentage change -- You are given a value at two points in time and asked to calculate the percentage increase or decrease. This is one of the most frequently tested concepts.
Ratios and proportions -- Questions ask you to divide quantities in a given ratio, find a missing part, or compare ratios across groups.
Averages and weighted averages -- You may need to calculate a mean, median, or weighted average from a data set, then use it to answer a follow-up question.
Data interpretation from tables -- A table presents figures across rows and columns (for example, quarterly revenue for five product lines), and you must locate the right cell, combine values, or compare trends.
Chart and graph reading -- Bar charts, line graphs, and pie charts each require slightly different reading techniques. Pie chart questions often combine visual estimation with percentage calculations.
Currency conversion and unit conversion -- Some tests include exchange rates or unit conversion factors as part of the data set, adding an extra calculation step.
Profit, loss, and margin calculations -- Common in finance-oriented tests. You calculate gross margin, net profit, or break-even points from provided financial data.
If you want to explore free practice questions across all these categories, visit our free tests page for sample questions with full explanations.
Worked Examples With Step-by-Step Solutions
Seeing how to approach problems systematically is one of the best ways to prepare. Here are three representative questions.
Example 1: Percentage Change
Scenario: A company's revenue was $4.2 million in Q1 and $4.83 million in Q2. What was the percentage increase in revenue from Q1 to Q2?
Step 1: Find the difference. $4.83M - $4.2M = $0.63M
Step 2: Divide the difference by the original value. $0.63M / $4.2M = 0.15
Step 3: Multiply by 100 to get the percentage. 0.15 x 100 = 15%
Answer: Revenue increased by 15% from Q1 to Q2.
Common mistake: Dividing by the new value ($4.83M) instead of the original value ($4.2M). Always divide by the starting figure when calculating percentage change.
Example 2: Ratio Problem
Scenario: A project budget of $180,000 is split between marketing, development, and operations in the ratio 2:5:3. How much is allocated to development?
Step 1: Add the ratio parts. 2 + 5 + 3 = 10
Step 2: Find the value of one part. $180,000 / 10 = $18,000
Step 3: Multiply by development's share. $18,000 x 5 = $90,000
Answer: Development receives $90,000.
Example 3: Data Interpretation From a Table
Scenario: The table below shows unit sales for three products across four quarters.
| Product | Q1 | Q2 | Q3 | Q4 |
|---|---|---|---|---|
| Alpha | 1,200 | 1,350 | 1,500 | 1,400 |
| Beta | 980 | 1,100 | 1,050 | 1,200 |
| Gamma | 1,500 | 1,450 | 1,600 | 1,550 |
Question: Which product had the highest average quarterly sales?
Step 1: Calculate the average for each product.
- Alpha: (1,200 + 1,350 + 1,500 + 1,400) / 4 = 5,450 / 4 = 1,362.5
- Beta: (980 + 1,100 + 1,050 + 1,200) / 4 = 4,330 / 4 = 1,082.5
- Gamma: (1,500 + 1,450 + 1,600 + 1,550) / 4 = 6,100 / 4 = 1,525
Answer: Gamma had the highest average quarterly sales at 1,525 units.
Tip: In a timed test, you could estimate by comparing totals rather than computing exact averages, since the product with the highest total also has the highest average when the number of periods is the same.
Example 4: Currency Conversion
Scenario: A UK-based company reports expenses of 45,000 GBP. The exchange rate is 1 GBP = 1.27 USD. What is the expense in USD?
Step 1: Multiply the amount by the exchange rate. 45,000 x 1.27 = 57,150 USD
Answer: The expense is $57,150. Watch for questions that reverse the conversion direction -- always check which currency is the base rate.
Test Providers: Format Differences and What to Expect
Different employers use different test providers, and each provider has its own format. Understanding these differences helps you prepare for the specific test you will face.
| Provider | Questions | Time Limit | Calculator | Format Notes |
|---|---|---|---|---|
| SHL | 18 questions | 17-25 min | Usually allowed | Data presented in tables and charts; multiple-choice with 4-5 options |
| Cubiks (now Talogy) | 20 questions | 25 min | Allowed | Mix of tables, graphs, and text-based data; some questions share a data set |
| Kenexa (IBM) | 24 questions | 20 min | Not always | Faster pace; simpler data but tighter time pressure |
| Revelian (now Criteria) | 20 questions | 20 min | Allowed | Adaptive difficulty in some versions; mix of business scenarios |
| Aon (cut-e) | 37 items | 12 min | Not typically | Very short time per item; uses a true/false/cannot say format with numerical data |
SHL tests are the most common globally and are used by employers such as Deloitte, Unilever, and many NHS trusts. Aon (cut-e) tests are used by several European and Australian employers and stand out for their unusual speed requirements. Cubiks (Talogy) tests appear frequently in P&G hiring processes.
💡Find out which test provider your employer uses before you start practicing. Provider-specific practice gives you a significant advantage because you will already be familiar with the format, pacing, and question style.
Time Management Strategies That Work
Time pressure is the defining challenge of numerical reasoning tests. Most candidates report that they could answer the questions correctly if they had unlimited time -- the difficulty is doing so within roughly 60 to 90 seconds per question.
Here are strategies that consistently help candidates improve their scores:
Read the question before the data. Know what you are looking for before you scan a table or chart. This prevents you from processing irrelevant information.
Estimate before you calculate. Round numbers to the nearest whole figure to check whether your answer is in the right ballpark. If the options are 12%, 15%, 22%, and 31%, a rough estimate will often eliminate two or three choices immediately.
Skip and return. If a question is taking more than 90 seconds, mark it and move on. Spending three minutes on one hard question means losing time on two easier ones.
Use elimination. Even if you cannot solve a question precisely, you can often rule out obviously wrong answers and make an informed guess from the remaining options.
Practice with a timer. Untimed practice builds understanding, but timed practice builds speed. Alternate between the two during your preparation.
Ready to put these strategies into action? Start practicing with full-length timed tests that simulate the real assessment experience.
Calculator vs. No-Calculator Tests
Whether a calculator is allowed depends entirely on the test provider and the employer. Some assessments explicitly provide an on-screen calculator; others prohibit any calculator use.
When a calculator is allowed:
- Use it for multiplication and division but try to estimate first so you catch keying errors
- Do not rely on it for every step -- reading and interpreting the data still takes most of the time
- Practice with the same type of calculator you will have on test day (basic, scientific, or on-screen)
When no calculator is available:
- Strengthen your mental arithmetic, especially multiplication tables up to 12, common fraction-to-percentage conversions, and dividing by 5, 10, 25, and 50
- Round aggressively and use estimation to narrow down answer choices
- Memorize key benchmarks: 10% is dividing by 10, 5% is half of 10%, 1% is dividing by 100
For either format:
- The bottleneck is usually data interpretation, not arithmetic. Even with a calculator, you need to know which numbers to use and which operation to apply.
💡Calculator access does not make the test easy. The real challenge is reading the data correctly and choosing the right calculation. Practice both mental math and data interpretation regardless of the format.
Common Mistakes and How to Avoid Them
Knowing the most frequent errors helps you build habits that prevent them.
Misreading the question. The question asks for the difference between two values, but you calculate only one of them. Always re-read the question after finding your answer to confirm you answered what was actually asked.
Using the wrong base for percentage calculations. Percentage increase uses the original value as the denominator. Percentage of a total uses the total as the denominator. Mixing these up is the single most common error in numerical reasoning tests.
Confusing units or scales. A table might show figures in thousands while the answer options are in millions, or a chart axis might start at 50 rather than zero. Check units and scale labels before you start calculating.
Spending too long on one question. Every question is usually worth the same number of points. A question you find hard may be easy for others, but the two questions you skipped because you ran out of time are guaranteed lost points.
Not checking whether the answer is reasonable. If a company's revenue grew from $10 million to $11 million and your calculated growth rate is 110%, something went wrong. A quick sanity check catches these errors.
How Employers Use Your Results: Scoring and Percentiles
Your raw score on a numerical reasoning test is converted into a percentile, which shows how you performed relative to a comparison group (called a norm group). A percentile of 65 means you scored higher than 65% of people in that norm group.
How scoring typically works:
- Each correct answer earns one point. There is usually no penalty for wrong answers, so you should never leave a question blank.
- Your raw score is compared against the norm group, which may be recent graduates, experienced professionals, or applicants at a similar level.
- Employers set a cut-off percentile. For competitive graduate schemes at firms like Goldman Sachs or Deloitte, the threshold might be the 70th or 80th percentile. For other roles, the 50th percentile may be sufficient.
What employers do with the results:
- Screening stage: Many companies use numerical reasoning as a pass/fail filter early in the hiring process. If you do not meet the threshold, your application does not advance.
- Ranking candidates: When many applicants pass the threshold, employers may rank candidates by score to decide who moves forward.
- Combined assessment: Some employers combine numerical reasoning scores with abstract reasoning, verbal reasoning, and situational judgment to create a composite score.
- Verification testing: A growing number of employers administer a shorter verification test during an assessment center to confirm that you achieved a similar score without assistance.
💡Most tests have no penalty for guessing. Never leave a question unanswered. Even a random guess gives you a 20-25% chance on a multiple-choice question.
Your 2-Week Preparation Plan
Consistent, focused practice is the most reliable way to improve your numerical reasoning score. Here is a practical plan you can follow in the two weeks before your test.
Week 1: Build foundations
- Days 1-2: Review core concepts -- percentages, ratios, fractions, averages, and percentage change. Work through untimed practice questions to build accuracy.
- Days 3-4: Practice data interpretation. Work with tables, bar charts, line graphs, and pie charts. Focus on identifying relevant data quickly.
- Days 5-7: Take your first timed practice test. Review every question you got wrong or guessed on. Identify your weakest question type and spend extra time on it.
Week 2: Build speed and confidence
- Days 8-9: Practice estimation techniques. Try solving questions by rounding first, then checking with exact calculations.
- Days 10-11: Take two more timed practice tests under realistic conditions (quiet room, no interruptions, strict time limit).
- Days 12-13: Review all mistakes from both weeks. Look for patterns in your errors.
- Day 14: Take a final timed test. Aim for improvement over your Week 1 score, not perfection.
For a comprehensive set of practice tests that track your progress and adapt to your skill level, explore the All Test Package, which includes numerical reasoning alongside verbal, abstract, and other aptitude tests.
Frequently Asked Questions
Are calculators allowed in numerical reasoning tests?
It depends on the test provider and the specific employer. SHL and Cubiks (Talogy) tests generally allow a basic on-screen calculator. Aon (cut-e) tests often do not. Kenexa tests vary by employer. Always read the test instructions carefully before you begin, and if you are unsure, ask the recruiter. Even when a calculator is allowed, mental estimation remains a critical skill because it saves time and helps you catch errors.
How many questions can I expect?
Most numerical reasoning tests contain 15 to 25 questions with a total time limit of 17 to 35 minutes. The notable exception is Aon (cut-e), which can include 37 items in just 12 minutes. On average, you will have roughly 60 to 90 seconds per question. The key is recognizing that you may not finish every question, so accuracy on the questions you do attempt matters more than attempting every single one.
What if I am not strong at math?
Numerical reasoning tests require competence with basic arithmetic, not advanced mathematics. You will not see calculus, trigonometry, or complex algebra. The math involved is percentages, ratios, fractions, averages, and basic multiplication and division. If these areas feel rusty, two weeks of focused practice is usually enough to reach a competitive level. Start with untimed practice to build understanding, then gradually add time pressure.
Can I retake a numerical reasoning test if I fail?
Most employers allow you to reapply after a waiting period, typically 6 to 12 months. Some test providers lock your results for a set period. If you did not pass, use the waiting period productively: identify your weak areas, practice consistently, and retake with a stronger foundation. Your second attempt will almost always be better than your first if you prepare properly.
Do employers see how long I spent on each question?
Some test platforms do record time-per-question data, but most employers focus on your overall score and percentile rather than individual question timings. That said, extremely uneven timing patterns (spending five minutes on one question while rushing through others in ten seconds) could theoretically flag for review, especially in proctored assessments. Aim for a steady, consistent pace.
How much can practice actually improve my score?
Research on psychometric testing consistently shows that practice improves scores. Most candidates see meaningful improvement within two to four weeks of regular practice. The gains come from three sources: familiarity with question formats (so you waste less time figuring out what is being asked), improved calculation speed (so you work faster), and better test strategy (so you manage time and avoid common mistakes). Structured practice with feedback is more effective than simply repeating random questions.
Start Preparing Today
Numerical reasoning tests are predictable. The question types, the data formats, and the calculation methods follow established patterns. That predictability is your advantage -- the more you practice, the faster and more accurate you become.
Begin your preparation with realistic numerical reasoning practice tests and track your improvement over time. Every practice session brings you closer to the score you need.
