AU common mistakes
AU · Understand the traps and how to avoid them.
Reading only the last line of a long vignette and answering the question that seems most familiar instead of the one asked, then discovering the stem's opening clauses (setting, timeframe, who is asking) would have excluded two options.
Vignette stems are long and time pressure pushes candidates into keyword-matching; the brain locks onto the first recognisable concept and stops encoding the rest of the stem.
What to do instead
Read the final sentence first to know the ask, then read the whole stem once with the ask in mind, and physically underline the three qualifiers (who, when, what outcome) before touching the options.
Choosing an option because it is a true statement, when the question asks for the BEST, FIRST, or MOST LIKELY action, so a factually correct but lower-priority or later-stage action is selected.
Training rewards recognising true statements; candidates are not habitually distinguishing truth from priority, and 'best/first/next' items usually contain several true distractors by design.
What to do instead
Circle the superlative or sequencing word in the stem and rank the surviving options on that single dimension; if two options are both true, the question is asking you to rank, not to recognise.
Mis-handling negative stems such as EXCEPT, NOT, or LEAST by selecting the one correct-sounding option, because the negation is a single small word buried mid-sentence.
Negations are low-salience and working memory drops them while attention shifts to option content; the default response set is 'pick the right answer', which fights the instruction.
What to do instead
Whenever you see a negation, restate the task in your own words as 'three of these belong, one does not' and cross out the three that belong first, so the answer is whatever remains.
Spending disproportionate time on the first difficult item or two, then rushing the final block and leaving several items blank or mis-keyed.
Sunk-cost persistence plus the belief that harder items are worth more marks; candidates do not know their own per-item time budget before the clock starts.
What to do instead
Before starting, compute a per-item budget from total items and minutes, mark any item you cannot resolve within roughly 1.5 times that budget, move on, and reserve a fixed final block for flagged items and a blank-check.
Changing a first-choice answer on review without finding new evidence in the stem, usually converting a correct answer into a distractor.
Anxiety on review mimics doubt, and re-reading a stem a second time feels like new information when it is only familiarity.
What to do instead
Adopt a rule: only change an answer if you can name the specific stem word or fact that the original choice contradicts. If you cannot state the evidence out loud, leave it.
Treating absolute qualifiers (always, never, exclusively, all) and hedge qualifiers (usually, may, in most cases) as interchangeable, and eliminating hedged options as 'too vague'.
Option-elimination heuristics taught for other tests say hedged answers are weak; in applied professional domains the hedged option is often the accurate one, and absolutes are the trap.
What to do instead
Judge each option on whether the absolute claim survives real-world exceptions in the domain; if you can name one legitimate exception, the absolute option is wrong and the hedged version becomes the candidate.
Confusing a pair of closely related domain terms that differ only in scope or sequence, and answering as if the two were synonyms.
The terms co-occur in study material and are learned as a cluster rather than as a contrast; exam writers deliberately place both in the option set.
What to do instead
Build one explicit contrast card per confusable pair: definition, one boundary case where only one applies, and the consequence of swapping them. Test yourself by explaining the difference aloud, not by re-reading definitions.
Applying the framework or value system of one domain to a question from the other domain, producing an answer that is defensible in the wrong context.
Domains are revised in separate blocks, so recent study dominates retrieval; nothing in the item flags which lens is required beyond the stem's own subject matter.
What to do instead
Classify each item to a domain in the first five seconds (based on the object of the question, not the vocabulary), and only then retrieve that domain's decision rules and ordering of priorities.
Guessing at multi-step or calculation items instead of setting them up, because the arithmetic looks long and the answer options are close together.
Time pressure plus numeric options that differ by small margins make estimation feel futile, so candidates abandon the setup and pick the most familiar-looking number.
What to do instead
Write the formula or decision sequence on the scratch sheet first with units and any conversion, estimate the magnitude before computing, and eliminate options outside the estimate; flag rather than guess if the estimate eliminates nothing.
Answering every item in one linear pass and running out of time before reaching a block of items that were easier than the ones already agonised over.
Candidates treat item order as difficulty order; in practice difficult items are often front-loaded and later items are independent and quicker.
What to do instead
Do a first pass answering only items resolved within about 40 seconds, then a second pass on flagged items, then a final pass to confirm no blanks; a blank is a guaranteed loss, a considered guess is not.