EPSO Numerical Reasoning Test

What the test is

The EPSO numerical reasoning test measures whether you can extract the right number from a small data set, do a short calculation and pick the correct answer in roughly 60 to 90 seconds. The format is multiple choice. Each item gives you a table, bar chart, line chart or short stem of statistics, asks a quantitative question, and offers four or five candidate answers. You read, decide which figures matter, do the maths, and choose.

The test is intentionally not a maths exam. The arithmetic is at upper-secondary level: percentages, ratios, weighted averages, growth rates and basic algebra. The difficulty comes from time pressure and from data sets engineered to punish careless reading. EU administrators routinely interpret budget, statistical and indicator data; the test exists to predict that.

This page sits inside our broader EPSO competitions guide; pair it with the abstract and verbal reasoning sub-articles to cover the full reasoning battery.

What is being measured

The competency targeted is analysis and problem-solving applied to quantitative data. EPSO is checking three things at once: can you locate the relevant figures in the data set, can you choose the correct calculation, and can you finish before the timer expires. A candidate who knows the maths but reads slowly will score badly. A candidate who reads quickly but rushes the calculation will also score badly. The test selects for the intersection.

Question types you will see

• Percentage change. "By what percentage did X grow between year A and year B?" The trap is identifying the correct base.
• Ratios and shares. "X as a percentage of total Y." Often combined with a chart that shows components versus total.
• Weighted averages. "Average of metric M across countries weighted by population." Multi-step.
• Conversion. Currency, units (tonnes versus kilotonnes), or time scaling (annual versus monthly).
• Reverse calculation. "If the share of X rose from 12% to 15% and total Y is constant at 240, by how much did X grow?"
• Two-step problems. Two of the above chained together. Time pressure matters here most.

Recent format changes (2024–2026)

Numerical reasoning has remained part of EPSO's reasoning battery through the 2024–2025 modernisation. Some recent cycles split the reasoning components into a single integrated test rather than three separate timed sections. Item difficulty appears unchanged. Format may evolve further; check the official EU Careers portal on the day of the notice.

Common mistakes

1. Wrong base. Computing percentage change relative to the new value instead of the old. Always anchor on the year/value asked.
2. Mixing units. Totalling figures in millions with figures in thousands. Annotate units before computing.
3. Averaging averages. The mean of three per-capita ratios is not the per-capita of the combined population. Build a single weighted ratio.
4. Reading the wrong column. Tables often list both "value" and "share". Confirm which the question wants.
5. Trusting the chart over the data. If a question gives both a chart and a table, the table is exact; the chart is for orientation.
6. Skipping rounding. Many items expect the answer rounded to a specific decimal. The closest unrounded option is often a distractor.
7. Spending two minutes on one item. The opportunity cost is at least two further items. Budget 60–90 seconds and move on.
8. Forgetting the on-screen calculator. Mental arithmetic is slower and more error-prone than the calculator for any multi-digit operation.
9. Practising offline only. The EU Careers sample tests use a specific click-and-drag interface; build muscle memory there.

Preparation resources

Official

• EU Careers sample tests — current item types, difficulty calibration, on-screen interface.
• Eurostat — real EU statistics. Export a table, design your own question, time yourself answering. Free, unlimited, the closest you can get to the data style EPSO uses.

Public study material

Any GMAT or GRE quantitative reasoning workbook covers the same arithmetic at similar difficulty. Use them for volume; calibrate timing against EPSO samples, not against the GMAT timer (which is more generous).

Related guides on this site

• EPSO competitions overview
• Abstract reasoning
• Verbal reasoning
• Situational judgement
• EPSO AD5 graduates competition

Test-day strategy

Three rules. First, read the question before the data set. Knowing what you are looking for changes how you scan the table by an order of magnitude — you head straight to the relevant cells instead of absorbing every figure. Second, do the unit and base check before computing. Half of all wrong answers come from arithmetic on the wrong figures, not from arithmetic errors. Third, eliminate options before computing where you can — three of the five options are usually too far from any plausible answer once you have a rough order of magnitude.

Practice over six to eight weeks if you can. The skills compound: you will find that by week four you are reading tables in a different way, eyes drawn straight to the columns that matter. That is the skill EPSO is selecting for, and it transfers directly to working life as an EU administrator. The job is, in large part, exactly this — under exactly this kind of time pressure.

A second-order tactic is order-of-magnitude estimation. Before reaching for the on-screen calculator, glance at the answer options. If the spread between options is wide — say €30,000, €36,000, €37,500, €40,000 — you can rule out half of them with a back-of-envelope estimate (1,440 ÷ 40 is "in the mid-thirty thousands") and only do the precise division to discriminate between the two finalists. Estimation saves about ten seconds per item; over a 25-item test that is four extra minutes — enough time to crack two more items you would otherwise have rushed.

On test day, manage your physical state. Numerical reasoning is the most cognitively expensive of the three reasoning components and tends to land in the middle of the test slot when fatigue is starting to bite. Hydration, a light meal beforehand, and a 30-second eye break between items are not folklore; they cost almost nothing and they do measurably help sustained accuracy. EPSO administers these tests in a window of two to three hours; pacing your physical energy matters as much as pacing your time.

A six-week preparation plan

If you have six weeks before the test, the highest-leverage allocation we have seen is roughly: weeks one and two on the official EU Careers samples to calibrate item style and difficulty; weeks three and four on volume practice using a GMAT/GRE quantitative workbook with a strict 75-second timer; weeks five and six on mixed mocks at full test length and on diagnosing your specific weak channel — most candidates discover, by mocking, that a single category of question (often two-step weighted ratio problems, or unit-conversion problems) accounts for a disproportionate share of their misses. The last two days should be light: one short timed mock to keep the muscle warm, then rest. Fatigue is a bigger threat to numerical reasoning than under-preparation in the final 48 hours.

Two small habits compound. First, after every practice block, write down each item you got wrong and the single mistake you made — wrong base, wrong unit, wrong row, calculation slip. After ten blocks the pattern is unmistakable. Second, keep a one-page formula sheet for percentage change, ratio, weighted average and compound growth. You will not be allowed to take it into the test, but rewriting it weekly bakes the formulae into reflex. By the test day you should not be reaching for "what's the formula?" any more than for "what's two plus two?". That mental space is then available for reading the question carefully — which is where the marks actually are.

A common question is whether to invest in paid preparation courses. The honest answer is: only after you have exhausted the official EU Careers samples and a free GMAT-style workbook, and only if your weakest channel is one a course teacher can address better than a textbook. Most candidates do not need a paid course. Those who benefit most are usually candidates whose first language is not the language of the test, who get value from working through items aloud with a tutor.

Frequently asked questions