EPSO AST/SC common mistakes
EPSO AST/SC · Understand the traps and how to avoid them.
In verbal reasoning, answering from personal knowledge or opinion about the topic (e.g. an EU policy you already know) instead of from the passage, so you pick a statement that is true in the world but not supported by the text.
The passages are often drawn from real institutional or policy material, so prior knowledge feels like a shortcut and the brain substitutes recognition for reading. Candidates also confuse 'true in reality' with 'stated in the text'.
What to do instead
Force the question into one of four verdicts before looking at the options: stated, contradicted, not stated, or inferred. Underline the exact sentence that supports your choice. If you cannot point to a sentence, the answer is 'cannot say', not 'true'.
In verbal reasoning, treating 'cannot say' as a last resort and choosing between the two substantive options because one feels more plausible, which systematically loses marks on the not-stated items.
Test-takers are trained by school exams to always pick a 'correct' statement, and 'cannot say' feels like a non-answer. Plausibility is mistaken for evidence.
What to do instead
Before reading the options, decide yourself whether the passage settles the claim. Practise sets where roughly a third of items are 'cannot say', and deliberately count how many you assign to each verdict; if you never choose 'cannot say', you are over-reading.
In verbal reasoning, reading the passage first and the question second, then re-reading the whole passage for every option, which burns time and pulls attention toward irrelevant detail.
It mirrors how people read in daily life and feels thorough, but the item only tests one or two claims, and the stem tells you which ones matter.
What to do instead
Read the question stem first, then scan the passage for the keyword or paraphrase it names. Read only the sentence containing it plus the one before and after. Only fall back to full reading if the claim is genuinely about the passage's overall thrust.
In numerical reasoning, using the calculator for every operation and losing time on trivial arithmetic, or conversely refusing the on-screen calculator and making slips in multi-step percentage and ratio work.
Candidates either distrust mental arithmetic under time pressure or over-rely on the tool out of anxiety. Neither reflects the real constraint, which is deciding what to calculate.
What to do instead
Practise a fixed split: estimate first (round to convenient numbers) to identify the right answer's rough size, then calculate exactly only when two options are close. Learn the standard shortcuts for percentage change, ratios and weighted averages so you can sanity-check the calculator.
In numerical reasoning, answering the question that was asked literally but not the one intended, e.g. giving the absolute difference when the stem asks for percentage change, or giving a per-year figure when it asks for the total.
Under time pressure candidates latch onto the numbers and start computing before finishing the sentence; units and 'per' words are skimmed.
What to do instead
Circle the unit and the comparison word in the stem ('increase', 'ratio', 'average per month', 'total over the period') before touching the data. After computing, state the answer in words with its unit and check it matches the stem.
In numerical reasoning, misreading tables and charts: taking the wrong row or column, ignoring a footnote or a different base year, or reading a stacked bar as if segments started at zero.
Layouts are dense and deliberately similar; the eye jumps to the visually salient number rather than the labelled one. Footnotes are set in small type and skipped.
What to do instead
Before calculating, say aloud what each axis, row header and unit means, and read every footnote. Trace the row and column with a finger or cursor. For stacked charts, subtract the lower segment to get the upper segment's value.
In abstract reasoning, hunting for a single rule across the whole series instead of testing candidate rules one at a time and eliminating options, so you freeze or guess when the pattern is compound (e.g. rotation plus alternating shading).
The items are designed with two or more simultaneous transformations, but candidates expect one clean rule and abandon a partially correct hypothesis too early.
What to do instead
Work through a checklist in fixed order: number of elements, position/rotation, size, shading/fill, shape type, and addition/subtraction of elements. Test each against consecutive figures and keep options that survive. Expect two rules and combine them.
In abstract reasoning, spending several minutes on one stubborn item and then rushing the rest, so a single hard question costs three or four later ones.
Sunk-cost thinking: having invested time, candidates feel the answer is nearly within reach, and the perceived unfairness of skipping makes them persist.
What to do instead
Set a personal cap per abstract item (about 45–60 seconds). If no option has survived your checklist by then, mark one plausible answer, flag it if the format allows, and move on. Practise with a visible timer until the cap becomes automatic.
In situational judgement, answering as the idealised manager or as your own personality rather than as the grade being recruited, e.g. choosing to escalate everything to a head of unit or to work heroically alone instead of consulting peers at your level.
Candidates either import private-sector habits or answer to look conscientious, and 'escalate' and 'do it yourself' both feel safe. The test scores against a defined competency framework for the specific grade.
What to do instead
Anchor on the grade's expected autonomy: resolve within your remit, consult the people affected, inform the line manager rather than asking permission for routine matters, and escalate only genuine risks of harm, legality, or deadline failure. For each option, ask 'is this proportionate to my grade and does it respect procedure and colleagues?'
In situational judgement, picking the most decisive-sounding option in every item and ignoring the 'least effective' or ranking format, so the same behaviour is chosen as best in one item and worst in another.
Candidates read the scenario and react to the behaviour, forgetting that the instruction (most effective, least effective, rank all) changes the task. Decisiveness is also over-rewarded by intuition.
What to do instead
Read the instruction line twice and paraphrase it before reading the options. For 'least effective' items, deliberately identify the option that is plausible but inappropriate (e.g. ignores consultation, or over-escalates) rather than the obviously bad one. For ranking items, sort all options before assigning any rank.