EPSO AD5 common mistakes
EPSO AD5 · Understand the traps and how to avoid them.
Answering verbal items from your own knowledge of EU affairs instead of from the passage - e.g. marking an option correct because it is factually true about the EU, even though the text never says it.
AD5 candidates usually follow EU policy closely, so a familiar topic triggers prior beliefs and fast 'recognition' feels like comprehension; under time pressure the brain substitutes memory for reading.
What to do instead
For every option, force yourself to point to the exact sentence (or the absence of one) that decides it. Adopt the rule: 'true but not in the text = wrong'. Practise on passages about topics you know well, precisely because they are the trap.
Skimming over quantifiers, negations and scope words ('all', 'only', 'some', 'unless', 'except', 'not necessarily') and then choosing an option that overstates or reverses the passage's claim.
Test items are built by inserting one small logical operator into an otherwise identical sentence; fast reading compresses such words, so two options look the same.
What to do instead
Circle or mentally flag every quantifier/negation before evaluating options, then test the swap: if replacing 'some' with 'all' (or deleting 'not') changes the truth, the option is not supported. Do a final re-read of the stem to confirm exactly what is being asked (true / false / cannot say).
Computing a percentage change with the wrong base - dividing by the new value instead of the original, or treating a rise from 20% to 25% as '5%'.
Formulas blur under a per-question countdown, and the distractor options are deliberately built from the wrong-base result, so a plausible number appears and confirms the error.
What to do instead
Write the denominator down before touching the calculator ('change ÷ ORIGINAL'). For percentage-point versus percent questions, state in words what the base is. Sanity-check: a change from 20 to 25 must give an answer above 20%.
Working with mixed units or scales (thousands vs millions, per-inhabitant vs absolute, index vs euro) and then selecting the option closest to a half-finished calculation.
Tables and charts mix units and the answer options are designed around intermediate results; the nearest-looking number is often a deliberate distractor.
What to do instead
Before calculating, write the unit of the requested answer and convert all inputs to it. Estimate the order of magnitude first (e.g. 'about 3 million'), then compute exactly; if the exact result and the estimate disagree by a factor of 10 or 100, find the unit error before answering.
Locking onto the first visible transformation in a figure series (e.g. 'it rotates 90°') and choosing the answer on that rule alone, without checking the other elements (count, shading, position, size, inner shape).
Human pattern detection latches onto the most salient change and then confirms it, ignoring the second rule that the item actually requires; the near-miss options are built exactly for this.
What to do instead
Run a fixed checklist on every series: number of elements, shape type, position/movement, rotation, shading/fill, size, intersections. Track each attribute separately across all given frames in a small table before looking at the options.
Deriving a rule from the first two or three figures and applying it to the answer without verifying it against the last given figure - so a sequence that accumulates or cycles fails at frame four or five.
Test sequences often look regular at the start and change regime later; candidates run out of time and skip the validation step.
What to do instead
Always validate the candidate rule against every provided frame; if it breaks at frame 4, the rule is wrong. Then eliminate options rather than only confirming one: a correct option must satisfy all rules simultaneously.
Choosing the most forceful, decisive-sounding option (escalate to the Director, impose a decision) over the option that respects your grade, the chain of command and the applicable procedure.
Candidates import private-sector or national-administration reflexes, and equate 'leadership' with visible decisiveness; the EU institutional logic rewards informing the hierarchy, consulting those affected and acting within your remit.
What to do instead
Rank options on two axes: effectiveness for the unit's output, and appropriateness given your grade and the rules. Prefer: gather facts, inform the line manager, consult colleagues, act within remit, then escalate if the issue exceeds your authority. Reject any option that bypasses a level or commits someone else's resources.
Answering 'what I personally would do' and ignoring whether the question asks for the MOST or the LEAST effective response, or picking the conflict-avoiding option that defers the problem.
Situational judgement feels like a personality questionnaire, so candidates express values instead of judging effectiveness; the instruction word is read once and forgotten.
What to do instead
Underline MOST/LEAST (or 'first thing to do') before reading the options. Judge each option by impact on the team and the deliverable - being agreeable but passive scores low. Practise ranking all options from best to worst, not just selecting one, so the relative judgement is explicit.
Writing the case study as an unstructured list of personal opinions, with no statement of the problem, no legal/procedural basis, no actor responsibilities, no timeline and no conclusion.
Time pressure pushes candidates to start writing immediately on the topic, and many have never practised the EU drafting conventions (note for file, briefing, options paper) that the markers expect.
What to do instead
Use a fixed skeleton every time: context and problem → relevant legal/policy basis → options assessed against stated criteria → reasoned recommendation → who does what by when, plus monitoring. Add headings, keep sentences short, and reserve five minutes to check that each section names the responsible actor and the deadline.
Making unverifiable or invented factual claims in the written case (precise article numbers, figures, dates) or recycling one prepared story as the answer for several competency questions.
Candidates want to sound authoritative and have memorised fragmentary facts; separately, they prepare only two or three examples and stretch them, which produces vague, generic answers.
What to do instead
Describe procedures accurately in words rather than guessing citations - an approximate but correct description scores, a wrong article number damages credibility. For the competency assessment, prepare six to eight distinct verified examples mapped to the competences (analysis, communication, teamwork, resilience, leadership), each in Situation–Task–Action–Result form with a quantified result and a one-line reflection on what you would do differently.