CogAT Level 12 common mistakes
CogAT Level 12 · Understand the traps and how to avoid them.
On verbal analogies, the candidate matches the answer choice to only the first word of the stem pair (e.g., picking 'kitten' for 'dog : puppy :: cat : ___' because it's a young animal) without checking that the same relationship direction and type holds for the whole pair.
Young test-takers read left-to-right and often lock onto the first familiar word, treating the analogy as a vocabulary association task rather than a relational one. The relationship (whole-to-part, category-to-member, degree, function) is what is being tested, but the surface meaning of individual words is more salient to a 12-year-old reader.
What to do instead
Teach a fixed sentence-frame routine: read the stem as a full sentence ('A dog is a grown-up puppy, so a cat is a grown-up ___'), then plug each answer choice into the same sentence and reject any that break it. Practice with pairs where the tempting distractor shares a topic but not the relationship.
The candidate treats verbal classification items as 'which word is most different in meaning' and eliminates the odd word based on personal familiarity or emotional salience (e.g., picking 'volcano' from a list of landforms because it seems scary or unusual) instead of finding the category that exactly three words share.
Classification requires constructing a rule that fits exactly three items, which is a hypothesis-testing skill that is still developing at this age. Familiarity bias makes a well-known or vivid word feel 'special' and therefore 'the odd one out,' even when it belongs to the same category as the others.
What to do instead
Have the candidate name the shared category for three words first and only then check the fourth. If no clean three-word category emerges, the rule is wrong and they should try a narrower or different category (e.g., 'things that erupt' vs. 'landforms'). Drill with lists that contain a vivid member of the shared category.
On sentence-completion items, the candidate selects the answer that sounds most sophisticated or that echoes a word already in the sentence, rather than the one that makes the sentence logically complete and grammatically consistent.
Test-wise students overcorrect toward 'impressive' vocabulary, and younger readers are drawn to lexical overlap (picking 'cold' for a sentence containing 'cold') as a false cue. They also under-read the full sentence, stopping at the blank instead of using the second clause as a constraint.
What to do instead
Instruct the candidate to cover the choices, predict their own word for the blank, then match. Require them to reread the completed sentence aloud (subvocally) and check both meaning and grammar. Flag any choice that merely repeats a word from the sentence as a likely trap.
The candidate solves quantitative comparison or number-series items by continuing the visible pattern for one or two steps without verifying the rule against every given term, so a series like 2, 6, 12, 20, 30 is answered as '+4, +6, +8, +10' but the next term is computed from the wrong increment.
Working memory load: holding the sequence and the inferred rule simultaneously is demanding, and children tend to extrapolate from the last visible difference rather than the rule that generated all differences. They also stop checking once a rule 'seems to work' for the first two terms.
What to do instead
Enforce a two-pass method: first write the differences (or ratios) under every adjacent pair, then state the rule in words ('add consecutive even numbers'), then apply it once more to the final term. Practice series where the first two differences suggest one rule but a later term falsifies it.
On quantitative items involving units, time, or money, the candidate computes correctly but answers in the wrong unit or ignores the unit asked for (e.g., giving minutes when hours are requested, or cents when dollars are requested).
The arithmetic captures attention and the final unit check is skipped under time pressure. At Level 12, multi-step word problems are new enough that the candidate treats the computation as the whole task rather than the last of several steps.
What to do instead
Make 'circle the unit in the question, write the unit next to your answer' a standing habit. Include a 5-second final check on every quantitative item: does my number have the right size and the right unit for what was asked?
The candidate guesses on quantitative items instead of using estimation and answer-choice structure, spending equal time on hard items and easy ones and running out of time before reaching items they could solve.
Young candidates often lack a sense of pacing and treat every item as equally weighted. They also do not yet use the answer choices as information (e.g., eliminating choices that are wildly the wrong magnitude) and are reluctant to mark and skip.
What to do instead
Practice with a timer in short sets and rehearse a three-step routine: if the method is not visible within about 30 seconds, estimate to eliminate, mark the item, and move on. Review marked items only after finishing the section. Build estimation drills (round to nearest ten, check magnitude) so guessing becomes informed.
On figure analogies and figure classification, the candidate compares whole figures by overall visual similarity ('these two look alike') rather than isolating the transformation (rotation, reflection, size change, shading, number of elements) that relates the first pair.
Holistic visual matching is the default perceptual strategy and is fast, so it wins over analytic feature comparison unless explicitly trained. Distractors are deliberately built to be visually similar to the stem while applying the wrong transformation, which punishes global matching.
What to do instead
Train a feature checklist applied in fixed order: count elements, note shading/fill, note orientation, note size, note position. Say the transformation out loud ('the shape rotated 90 degrees clockwise and the shading inverted') before looking at choices, then test each choice against the stated rule.
The candidate ignores paper-folding and figure-transformation items involving reflection versus rotation, assuming a folded-and-punched hole will appear in the same orientation on the unfolded paper, or confusing a mirror image with a rotated figure.
Mental rotation and mental folding are distinct spatial operations, and reflection reverses chirality in a way that is easy to miss without physical experience. Children who have not folded paper and punched holes tend to default to translation ('the hole just moves over').
What to do instead
Have the candidate physically fold and punch paper, then unfold and compare, several times. Teach the fold-count rule (each fold doubles the number of holes) and the reflection rule (mirror images cannot be rotated into each other). Use tracing paper to check orientation on practice items.
The candidate answers only the item types they have practiced and treats unfamiliar formats (e.g., a matrix with three changing attributes, or a series of figures with two independent patterns) as unanswerable, leaving them blank or guessing randomly.
Novel formats increase cognitive load and anxiety, and without a general strategy the candidate has no entry point. Blank answers also cost more than educated guesses on a scored test with no wrong-answer penalty in most administrations.
What to do instead
Teach a universal entry routine for any nonverbal item: find the attribute that changes across the row or series, track each attribute separately, and eliminate choices that fail on any single attribute. Practice mixed-format sets so no format is novel on test day, and rehearse 'always answer, eliminate first' as a rule.
The candidate rushes through the first items to 'save time' and makes avoidable errors on easy questions, then loses the time gained to hard items later, producing a net loss because early items are the ones most likely to be answered correctly.
Time anxiety and the belief that speed equals competence lead to skipping verification on items the candidate actually knows. The scoring reality - easy items are worth the same as hard ones - is not intuitive to a 12-year-old.
What to do instead
Run timed practice where the candidate must finish the first half of a section with at least 80% accuracy before touching the second half. Debrief accuracy by item position to make the point concrete: points lost on easy items cannot be recovered by extra time on hard ones.