OLSAT Level G common mistakes
OLSAT Level G · Understand the traps and how to avoid them.
On antonym items, the student selects a word that is merely unrelated to or less similar than the prompt word, rather than a true opposite, and moves on without checking the pairing.
Under time pressure the brain treats 'different from' and 'opposite of' as the same relation, and a distractor placed early in the list feels like a satisfying contrast.
What to do instead
Force the pair into a sentence frame ('Hot is the opposite of ___') and reject any option that only yields 'X is not really Y.' Then check that the two words match on intensity and part of speech, since near-synonyms and mild contrasts are the standard traps.
On sentence-completion items, the student reads only for topic and drops in the option that sounds smooth, ignoring the logical signal word (although, because, therefore, unless) that sets the direction of the blank.
Fluent reading is predictive: the eye completes familiar-sounding sentences from context and skips small function words that carry the actual logic.
What to do instead
Before reading the options, underline the connector and say the direction aloud ('contrast' or 'continuation'), then predict a word of your own for the blank. Only options matching both the direction and your prediction are eligible.
On verbal analogies, the student matches on shared subject matter (both pairs are about medicine, both about weather) instead of testing the specific relationship, and does not re-check the order of the answer pair.
Topical overlap is highly salient and gives an instant feeling of 'these go together,' which masks an imprecise or reversed bridge.
What to do instead
Write a one-sentence bridge with placeholders ('A is a ___ that is used to ___ B') and confirm it is true of the stem pair. Then impose the same bridge, in the same direction, on each option; discard any option where the bridge reverses actor and object.
On 'which one does not belong' items, the student picks the odd word on a superficial trait (unusual spelling, longer word, a single letter pattern) and never verifies that the rule covers all remaining words.
Serial processing stops at the first rule that seems to work, so one counterexample among the other four words goes unnoticed.
What to do instead
Invert the task: treat each option in turn as the excluded word and state the category the other four share. Only accept an answer when the category is stated explicitly and no leftover word violates it.
On pictorial series items, the student follows only the most eye-catching change (the shape rotating) and misses a second simultaneous rule such as a count increase or an alternating colour flip.
Visual attention latches onto motion or orientation, and working memory holds roughly one transformation at a time unless the changes are externalised.
What to do instead
Build a quick feature table with columns for position/orientation, shading, size and count, fill it in for every panel, and check each column separately for its own rule before choosing the next panel.
On picture analogies and pictorial classifications, the student chooses the option whose picture looks most like the stem picture, rather than the one produced by the same transformation.
Visual similarity is processed faster than relational reasoning, so a matching-looking distractor wins before the relation is ever articulated.
What to do instead
Say the transformation in words out loud before looking at the choices ('the figure loses its outline and doubles'), then evaluate each option only against that stated rule, not against visual resemblance.
On paper-folding and hole-punching items, the student unfolds mentally in a single step, so holes are not mirrored across each fold line and the total count is understated.
Folds are physically stacked, but most people reason about them as a flat picture, ignoring that each fold doubles the hole and reflects its position
What to do instead
Number the folds and undo them in reverse order, drawing the mirror image of every existing hole across each crease and doubling the count at each step; sanity-check that the final total is a power-of-two multiple of the punches.
On figural matrices, the student finds a rule that explains one row or one column and applies it without confirming the same rule holds in the perpendicular direction or across the third row.
The first coherent pattern found feels like the answer, and confirming it costs the time the student is anxious to save.
What to do instead
Adopt the rule only after it holds for two full rows and two full columns; if a feature varies freely as a distractor, note that explicitly so it is not forced into the rule.
On number series, the student assumes a constant difference and extends it forward, or extends a pattern from the first three terms without checking the later terms that break it.
Arithmetic sequences are the most over-learned schema, and early terms often fit several rules at once.
What to do instead
Compute the successive differences and the successive ratios in two rows before choosing a rule, then check for interleaved sequences by examining alternating terms and re-verify the candidate rule against every given term, including the last one.
On quantitative word problems, the student solves for an intermediate quantity, then selects that number because it appears in the options, or drops the units and mis-scales the answer.
The value just calculated is fresh and feels like the goal; question stems put the actual request at the end, where it is read last and least carefully.
What to do instead
Before computing, write the target quantity and its unit on paper ('asked: cost per person, in dollars'), label every intermediate value, and finish by asking whether the answer's size and unit are plausible for the situation described.