Civil Service Clerical common mistakes
Civil Service Clerical · Understand the traps and how to avoid them.
Answering verbal reasoning questions from your own knowledge of the subject rather than from what the passage actually says - e.g. marking 'True' on a health-and-safety statement because it sounds like standard practice, even though the passage says nothing about it.
Familiar topics (office procedures, public services, well-known events) trigger prior belief, and the brain fills the gap with what 'should' be true. Under time pressure the passage gets skimmed once and the answer is supplied from memory.
What to do instead
Force a source line before selecting: put a finger on the exact sentence that supports your answer. If you cannot point to one, the answer is 'Cannot say', not 'True'.
Treating a likely or sensible inference as 'True' - choosing True when the passage only makes the outcome probable, and choosing False whenever the passage merely fails to mention something.
Candidates collapse the three-way True/False/Cannot Say scale into a two-way true/false scale, so 'probably' silently becomes 'definitely' and 'not mentioned' silently becomes 'false'.
What to do instead
Apply the strict test for each option: (a) is it stated or logically forced by the passage? -> True; (b) does the passage directly contradict it? -> False; (c) neither? -> Cannot say. Also note whether the stem says 'stated', 'suggested' or 'implied' - they are not interchangeable.
Launching into full long multiplication or division on every numerical item, using the whole time allowance on one table question and then having to guess the last few.
School habit: the numbers look awkward, so candidates reach for the exact method. Test designers choose options far apart so that exact arithmetic is rarely needed.
What to do instead
Round first. Convert the figures to easy numbers (e.g. 4,850 -> 5,000; 19.6% -> 20%), compute the estimate, and pick the only option of that magnitude. Only do exact arithmetic if two options sit within a few per cent of your estimate, and leave anything longer than about 45 seconds.
Using the wrong base in percentage questions - taking a discount off the already-reduced price, dividing by the new figure instead of the original, or assuming a 20% rise followed by a 20% fall returns to the starting value.
The numbers in the question are the ones in front of you, so the most recent figure gets used as the base by default. The 'rise then fall' trap exploits the intuition that opposite changes cancel out.
What to do instead
Write 'base = ___' on your rough paper before dividing, and always name the original quantity or starting year. For chained percentage changes, sanity-check with round numbers (100 -> 120 -> 96, not 100).
Ignoring units and column headings - comparing a weekly figure with an annual one, treating '£000s' as pounds, or missing that a total is 'per 1,000 staff' rather than a headcount.
Under time pressure the eye jumps straight to the digits and skips the small print above them; the labels look like decoration rather than data.
What to do instead
Before any calculation, say the unit out loud or jot it down (e.g. 'col 2 = £ thousands, per year'), and after calculating check the magnitude of your answer against the options: if your figure is a thousand times an option, you have dropped a unit conversion.
Comparing two strings by eye in one fast left-to-right sweep, so that differences appearing after the fourth or fifth character - or on the line below - are missed entirely.
Attention is strongest at the start of a string and decays along it, and without a physical guide the eye drifts onto adjacent rows in a dense list.
What to do instead
Break each pair into blocks of three or four characters and compare block against block. Use a straightedge, ruler or folded paper to isolate one row at a time, and move it down only after the row is finished.
Stopping as soon as one error is found in a line or record, and moving on - so a second error later in the same line (a wrong digit in the reference, a misspelt surname) is never recorded.
Finding an error feels like completing the task, and the relief produces a mental 'done' signal. Many items are deliberately built with two or three differences.
What to do instead
Treat every field in the record as an independent check. After finding a difference, tick that field and continue to the end of the line, then re-scan the whole line once before moving on.
Missing transposition and look-alike errors - 'ie' for 'ei', 'ht' for 'th', '45' for '54', '10' for '01' - because the two strings 'look the same' at a glance.
Skilled readers recognise words by shape rather than letter by letter, so swapped letters inside a familiar word are invisible. Digit pairs have no shape to recognise and get read as a single chunk.
What to do instead
For any pair that looks identical, deliberately re-read the second item backwards, or aloud in your head, and check numbered fields digit by digit. Treat 'they look the same' as the trigger to slow down, not to move on.
Choosing the warmest, most helpful-sounding response in a situational judgement item - promising a caller that their case will be approved, or personally fixing something outside the role's authority - instead of the option that is helpful within the rules.
Candidates optimise for 'being nice' and read the question as a personality test. The distractors are written to be emotionally attractive; the correct answer is usually courteous but procedural.
What to do instead
Filter in a fixed order: (1) is it within my authority and does it avoid promising an outcome? (2) does it protect accuracy, confidentiality and safety? (3) does it escalate or refer rather than guess? Then pick the option that passes all three and still treats the customer with respect.
Misreading whether the item asks for the most or the least appropriate response, and treating a prioritisation scenario as a 'what would you like to do' question rather than ranking by urgency and consequence.
Most-to-least wording is a single word buried in a long stem, and candidates read the four options first and the stem second, anchoring on the options. Prioritisation items feel subjective, so gut preference takes over.
What to do instead
Underline 'most' or 'least' before reading the options. For prioritisation, list the four tasks and rank them by what happens if delayed (safety or a missed statutory deadline first, routine filing last) before looking at the answer choices.