Scio NSZ OSP common mistakes
Scio NSZ OSP · Understand the traps and how to avoid them.
Reading a verbal passage once at normal speed and then answering from memory of the gist, without returning to the exact lines each option depends on.
Under time pressure, candidates treat verbal items as general-knowledge questions and trust the impression the text left. Options in this exam are usually built by changing one word (a quantifier, a modal verb, a direction of causation), so a gist-level memory cannot distinguish the correct option from a near-miss.
What to do instead
Adopt a fixed two-pass routine: first pass for the main claim and the author's stance (about 60–90 seconds), then answer each option by locating the sentence that would have to be true for it to hold. If you cannot point to that sentence, the option is unsupported and should be eliminated even if it sounds sensible.
Treating 'true according to the text' and 'must be true in reality' as the same question, and picking the option that matches real-world facts rather than the passage.
Strong prior knowledge about the topic (history, biology, economics) feels like an advantage, so candidates import outside information and override what the author actually wrote. The exam deliberately uses plausible topics where real facts and text facts diverge.
What to do instead
Before answering, label the task: inference-from-text versus evaluation-of-claim. For inference items, physically cover the options, restate the passage's claim in your own words, then uncover and choose the option that follows from that restatement alone. Any option requiring knowledge not in the passage is out of scope.
Ignoring quantifier and scope words (all, some, most, only, at least, no more than, unless) and answering as if the statement were absolute.
The eye skips short function words and lands on content words, so 'most experts agree' is encoded as 'experts agree'. These words are exactly where distractors are built, and errors here are invisible on review because the candidate still believes the statement was absolute.
What to do instead
During practice, circle every quantifier and negation in the passage and in each option before choosing. Build a short personal list of scope words and drill statements of the form 'Only A are B' versus 'All B are A', checking which conversions are valid. On the exam, do this circling on the question booklet for every verbal item until it becomes automatic.
Solving numerical word problems by grabbing the two most visible numbers and applying the operation that 'looks right', without writing the relationship between quantities first.
Test anxiety and time pressure push candidates toward arithmetic before modelling. Percentages, ratios, and rates are usually presented with an extra irrelevant number, and the most visible pair is often the wrong pair.
What to do instead
Force a modelling step: write one line naming the unknown and one equation or proportion before any calculation. Underline units (per hour, per person, total, change) and convert everything to a single unit first. If your equation has no unknown, you have not modelled the problem yet.
Making percentage and ratio errors of direction - computing a 20% increase as multiply by 0.2, or reversing the base in 'A is 25% more than B' problems.
Percentage language is asymmetric and candidates memorise 'multiply by the percent' without tracking the base. The reverse-base error is especially common because both directions produce plausible-looking answers that appear among the options.
What to do instead
Always write the base explicitly: 'increase of p% on base X means X × (1 + p/100)'. For comparison items, set the smaller quantity to 100 and express the other relative to it, then check by substitution with simple numbers (e.g. B = 100, A = 125, so A is 25% larger and B is 20% smaller). Do ten such checks in a row until the direction is reliable.
Spending more than about two minutes on a single numerical item and then rushing the remaining items, producing a cluster of avoidable errors at the end of the section.
Sunk-cost thinking: having invested in a calculation, candidates refuse to abandon it. Because numerical items are scored equally, one hard item can cost three easy ones, and the resulting panic degrades arithmetic accuracy on those.
What to do instead
Set a hard per-item ceiling (roughly 90 seconds) and a visible checkpoint at the halfway mark of the section. If the ceiling is hit, mark the item, enter a plausible estimate if there is no penalty for wrong answers, and move on. Practise full timed sections so the abandon-and-return habit is rehearsed, not improvised.
Drawing no diagram for ordering, grouping, or scheduling puzzles and trying to hold the constraints in working memory.
These puzzles feel easy at first because the first two constraints are memorable, but the third and fourth silently drop out of memory, and the candidate then answers from an incomplete constraint set with high confidence.
What to do instead
Use a standard compact notation: a row of slots for sequences, a small grid for grouping, and a separate list of constraints with a tick for each one applied. After choosing an answer, re-verify it against every listed constraint, not just the one you were thinking about.
Assuming a single arrangement is the answer when several arrangements satisfy the constraints, and rejecting a valid option because it does not match the first arrangement found.
Once a candidate constructs one consistent arrangement, it feels like 'the' solution. Questions asking which option 'could be true' or 'must be true' require reasoning over all arrangements, and the first one found is only one witness.
What to do instead
Classify the question before solving: 'must be true' needs a deduction that holds in every arrangement; 'could be true' needs one valid witness; 'cannot be true' needs a contradiction with a stated constraint. For must/cannot items, test the option by trying to build a counter-arrangement rather than by recalling your own.
Evaluating an argument by whether the conclusion is agreeable rather than by whether the stated premises support it, and choosing the option that attacks a position the candidate dislikes.
Everyday argumentation trains us to judge conclusions, not links. On critical-reasoning items the conclusion is often deliberately bland or deliberately provocative, and personal agreement masquerades as logical strength.
What to do instead
Rewrite each stimulus in two lines: 'Premises: ... Therefore conclusion: ...'. Then ask only whether the conclusion follows, and label the gap between them. The correct strengthen or weaken option must touch that labelled gap; options that merely restate, agree with, or emotionally reject the conclusion are eliminated on principle.
Confusing necessary and sufficient conditions, and treating a correlation in the stimulus as proof of causation.
Conditional language ('only if', 'unless', 'requires') is compressed in natural prose, and causal claims are the default way people explain observed patterns. Both errors produce distractors that read smoothly and feel like common sense.
What to do instead
Translate every conditional into 'If P then Q' and immediately test the converse: does Q require P, or merely accompany it? For causal stimuli, ask what alternative explanation would produce the same correlation (reverse causation, a common cause, selection). Practise by writing the negation of each conditional you meet, since 'P only if Q' is falsified by P without Q.