Numerical reasoning tests explained

The skill

What is a numerical reasoning test?

A numerical reasoning test presents figures in a workplace context — a table of regional sales, a chart of monthly costs, a short financial extract, or a dashboard spread across several tabs — and asks you to decide what those figures support. You are given the data. Your job is to locate the part that matters, choose one calculation, and answer before the clock moves you on.

It is not a mathematics exam. The operations are the ones you met at school: percentages, ratios, rates, averages and simple currency conversion. What makes the test difficult is performing them accurately, from unfamiliar data, at a pace someone else set.

The response format is not fixed either. Some assessments want a number. Aon's numerical assessment instead gives you statements to mark true, false or cannot say, and Pearson's NDIT includes free-response items rather than only multiple choice.

Try it before anything else

Free numerical reasoning practice questions

Six original questions covering the formats above, hardest habits first. Commit to an answer before opening the solution — the solution shows the method and names why every wrong option is on the list.

What employers are looking at

What do numerical reasoning tests measure?

Employers use numerical reasoning to predict how well someone will work with the figures a role actually produces: budget variance, conversion rates, headcount ratios, pricing and forecast accuracy. Pearson describes its own numerical test as assessing the interpretation and manipulation of the numerical data routinely encountered in the workplace.

Three abilities are being separated. Data selection: can you find the one row that matters in a dense table. Method choice: can you turn a question into a single calculation instead of three. Judgement: can you tell that an answer of 940% is wrong before you commit to it.

Reading your result

How numerical reasoning tests are scored

Your raw score is rarely what the employer sees. Providers convert it into a comparison against other people, which changes what a good result means and what is worth optimising for.

• Results are percentiles, not marks out of ten. Criteria sends employers a raw score alongside a percentile, and Korn Ferry compares your result with a reference group of similar candidates. A 70th-percentile result means you beat 70% of the comparison group, whatever the raw count was.
• No provider publishes a pass mark. Korn Ferry states directly that it does not set one; any benchmark belongs to the hiring organisation. If a recruiter has disclosed a target, that is the only target worth working to.
• The norm group shapes the number. The same raw performance scores differently against graduate applicants than against the general population. This is why comparing scores across different assessments, or across friends' tests, says very little.
• Unanswered questions are part of the design. Aon tells candidates they are not expected to answer every task, and Criteria reports that fewer than 1% of candidates complete all 50 CCAT questions. Accuracy at a sustainable pace usually beats rushed completion.
• On an adaptive test, difficulty is not feedback. SHL's Verify Interactive and Korn Ferry's Elements select the next question from your last answer, so a harder question often follows a correct one. You cannot read your score from how hard the test feels.

Question formats

What do the questions look like?

The material and the response format both vary by provider. Learn the mechanics first and add the interface last.

• Table and chart interpretation. A data set followed by a question needing one or two values from it. The work is finding the right two numbers before any arithmetic starts.
• Percentages and percentage change. Increase, decrease, reverse percentages, and the difference between a percentage change and a percentage-point change.
• Ratios, proportions and shares. Splitting a total, comparing two shares, or scaling a known ratio up to a new total.
• Rates, averages and per-unit figures. Cost per unit, output per hour, weighted averages, and the mean of a subset rather than the whole table.
• Currency and simple financial arithmetic. Converting before comparing, applying a margin, or working with a figure that is already net of something.
• Statement evaluation. Rather than a number, you judge whether a statement about the data is true, false or cannot be determined. Aon's scales numerical uses exactly this response set.
• Sufficiency questions. You decide whether the information supplied is enough to answer at all. Pearson's RANRA includes a dedicated sufficiency section.

Launch coverage

Which providers use numerical reasoning tests?

The underlying maths is portable. The interface, the response format and the clock are not. Match the name in your invitation to the route below.

• Verify Interactive Numerical Reasoning. SHL's product fact sheet lists 18 minutes and a maximum of 10 questions in an interactive, activity-based, adaptive format. SHL Direct describes the classic task as answering from figures presented in statistical tables.
• scales numerical. Multiple data sheets and a series of statements, each marked TRUE, FALSE or CANNOT SAY. Aon lists the module at 12 minutes and 37 tasks, and tells candidates to have a pen, paper and calculator to hand.
• Talent Q Elements Numerical. Adaptive delivery with a 16-minute limit. Korn Ferry tells candidates to have a calculator, paper and pen, and recommends a desktop or laptop because the drop-down answer controls behave poorly on tablets.
• Numerical Data Interpretation Test. 21 items in 30 timed minutes. Pearson permits a calculator and includes free-response items alongside multiple choice.
• Numerical Analysis. Usually met inside a Swift battery. Saville lists Swift Analysis Aptitude at 18 to 24 minutes across verbal, numerical, diagrammatic and abstract reasoning, and Swift Executive Aptitude at 18 minutes.
• Logiks Numerical (Advanced). Talogy publishes the Logiks range at 12 to 25 minutes and 20 to 50 items depending on version, either as a combined general test or as one focused ability.
• CCAT. Numerical items sit inside one mixed test: 50 questions in 15 minutes, no calculator permitted, with verbal and spatial questions interleaved.

Provider names identify the assessment format. Cognivy is independent and is not affiliated with or endorsed by these providers.

What varies, and what does not

Timing and calculator rules

There is no single numerical reasoning test, so there is no single time limit, calculator rule or pass mark. These are the figures the providers themselves publish.

• Time limits and question volumes vary widely. Pearson's NDIT runs 30 minutes for 21 items. Criteria's CCAT gives 15 minutes for 50 mixed questions. Aon lists scales numerical at 12 minutes for 37 tasks.
• Calculator policy is per assessment, never universal. Aon and Korn Ferry both tell candidates to have a calculator ready, and Pearson permits one on the NDIT. Criteria states plainly that no calculators are allowed on the CCAT. Your own instructions are the authority.
• Finishing is not always expected. Aon tells candidates they are not expected to answer every question. Criteria reports that fewer than 1% of candidates complete all 50 CCAT questions. Unanswered items are designed into some of these tests.
• Adaptive delivery changes what a hard question means. SHL's Verify Interactive numerical assessment and Korn Ferry's Elements tests both select the next item from your last answer. An adaptive test may follow a correct answer with a harder item, but you cannot read your score from how hard the questions feel, so treat difficulty as normal behaviour rather than as feedback.

A repeatable approach

A method that survives the clock

The same five moves work whether the data is a table, a chart or a dashboard. Practise them until they are automatic, then add the clock.

1. 1

   Read the question before the data. Know what you are looking for and you will scan the table once instead of three times. Read the data first and you will read all of it.

2. 2

   Name the calculation before you start it. Say what you are about to compute — change over original — before touching a number. Most wrong answers are the right arithmetic on the wrong expression.

3. 3

   Check units, period and label. Thousands or millions, quarter or year, gross or net, this region or the total. A misread label produces correct arithmetic on the wrong quantity, so the answer survives a sanity check on the working.

4. 4

   Estimate, then compute. A rough answer tells you which options are impossible and catches a decimal-place slip immediately. On a no-calculator test it is often the whole method.

5. 5

   Decide your abandon point before you start. Set a fixed ceiling per question. When you reach it, choose the best remaining option and move. One rescued question is never worth three lost ones.

What slows progress

Common mistakes, and the fix for each

• Calculating before checking the label. Fix: read the axis title, the units row and the period column before the first keystroke. A calculation run against the wrong units or the wrong period is wrong however accurate the arithmetic, and nothing later in the question reveals it.
• Confusing percentage change with percentage points. Fix: whenever a figure is already a percentage, decide which of the two the question wants, and write the base you are dividing by next to your working.
• Reading the whole table. Fix: locate only the cells the question names. Dense tables are padded deliberately, and most of the data in a question is not needed for that question.
• Rescuing one hard question. Fix: set an abandon point before you start and honour it. On a test where fewer than 1% of candidates finish, protecting pace matters more than any single item.
• Practising with the wrong calculator rule. Fix: find out whether your assessment permits one before you build a habit. Mental estimation and calculator-first working are genuinely different skills.

In practice

How do you solve percentage questions under time pressure?

Write the base before the arithmetic. Every percentage question turns on one decision — what is this a percentage OF — and most wrong options on real tests are the right arithmetic on the wrong base. Change over original for percentage change; the reduced figure over one-minus-the-rate for reverse percentages; never add a discount back onto a sale price.

Keep percentage points and percent apart. When the data is already in percentages, a rise from 62% to 68% is 6 percentage points but a 9.7% increase — sample question 1 above turns entirely on that distinction, and so do real items in SHL-style and Aon-style sets.

Then make the fraction equivalents automatic: 12.5% is an eighth, 33⅓% a third, 16⅔% a sixth. On calculator-free tests like the CCAT they replace the calculator; on the rest they are the estimate that catches a slipped decimal.

In practice

How do you read a data table quickly without missing the trap?

Read the question first, and let it name the cells you need. A dense table is padded deliberately — sample question 1 carries ticket volumes that no correct calculation touches — and candidates who read data before question read most of it twice.

Then spend three seconds on the frame before any value: units row, period columns, and whether figures are gross or net, thousands or millions. A misread label produces clean arithmetic on the wrong quantity, which is the one error a checking pass will not catch, because the working looks right.

Finally, compute only what the comparison needs. Asked which region improved most, a glance at direction often eliminates half the rows before any division, and on a 15-minute test those seconds are the score.

In practice

How do you avoid unit and currency conversion errors?

Convert before you compare, never after. Figures in different currencies, or different magnitudes, cannot be ranked as they stand — sample question 2's largest local-currency number converts to the smallest sterling profit, and the exchange-rate row exists precisely to punish skipping it.

Anchor the direction of every rate before dividing. Write it as a sentence — one pound buys 13 kronor, so kronor become pounds by dividing by 13 — and the upside-down conversion, which multiplies instead, becomes impossible to commit silently.

Sanity-check magnitude at the end: a Swedish quarterly profit of £63 million from a kr4.9 million base fails the glance test. Order-of-magnitude checks catch conversion errors faster than redoing the arithmetic.

A clear route

How to prepare in the days you have

Most candidates get a few days of notice, not a study term. This order gets the most out of a short run-up.

1. 1

   First, fix the arithmetic rather than the data. Drill percentages, reverse percentages, ratios and averages untimed until the method is automatic. If this is where you lose marks, no amount of chart practice will help.

2. 2

   Next, one calculation per question. Work untimed on tables and charts, writing the expression before computing it. Review each miss by cause: data selection, method, arithmetic or pace.

3. 3

   Then add the provider's format. Switch to the response format named in your invitation — numeric entry, drop-downs, or true, false and cannot say — and to the calculator rule that comes with it.

4. 4

   Time it. Short timed sets first, then a full-length rehearsal at the real pace with the real interaction.

5. 5

   On the last day, repair rather than cram. Re-solve only the questions you got wrong, and stop. New material the night before adds noise, not skill.

Cognivy uses your assessment date to choose the route rather than asking you to predict a study schedule. When you sit down to practise, you choose the session length that fits that day.

Pace

Getting faster without losing accuracy

Speed on a numerical test comes from removing work, not from working faster. Four things remove the most.

• Estimate first, always. An approximation is enough to reject options that are an order of magnitude out, and it tells you whether a full calculation is needed at all.
• Learn the fraction equivalents. 12.5% is an eighth, 33⅓% is a third, 16⅔% is a sixth. Recognising them turns a calculator step into a mental one, which matters most where no calculator is allowed.
• Keep one working layout. Write the same way every time: expression, numbers, answer, units. A consistent layout is what lets you catch your own slip without redoing the question.
• Rehearse the interaction, not only the maths. Drop-downs, tabbed data sheets and free-response boxes each cost seconds until they are familiar. Korn Ferry specifically warns that Elements Numerical drop-downs render badly on tablets.

Keep exploring

Related test types and how they differ

Numerical reasoning sits next to several assessments that look similar on screen and are scored differently.

Direct answers

Numerical reasoning test FAQs