CogAT Grades 9-12 common mistakes
CogAT Grades 9-12 · Understand the traps and how to avoid them.
Spending the same amount of time on the first verbal analogy as on the last, instead of rationing time across the whole section.
Analogy items feel solvable, so candidates keep re-reading until a pair 'clicks', burning minutes on one item and leaving the final items rushed or blank. CogAT sections are tightly timed, and unanswered items cost the same as wrong ones.
What to do instead
Before the section starts, divide the item count by the minutes allowed to get a per-item budget (for example, 20 items in 10 minutes means 30 seconds each). Practice with a visible timer so you learn what 30 seconds feels like, then move on the moment you exceed it and mark the item to revisit only if time remains.
Choosing an analogy answer because the two words are related in real life rather than because the relationship matches the stem's relationship in the same direction.
Strong real-world associations hijack judgment. A candidate sees 'doctor' in an answer and picks it because it belongs with 'hospital' in the stem, without checking that the stem is a worker-to-workplace relation in that order and the answer reverses it.
What to do instead
Write the stem relationship as a short sentence with a direction arrow, e.g. 'A scalpel is a tool used by a surgeon.' Then test each option by substituting it into that same sentence. Eliminate any pair that fits the words but not the sentence, and any pair that only works if you reverse the arrow.
Treating sentence-completion or verbal-classification items as a search for a familiar-sounding word instead of predicting the needed meaning before reading the options.
Answer choices anchor thinking. Once a plausible word is visible, the brain builds a justification for it, so a candidate talks themselves into a near-fit and stops evaluating the rest.
What to do instead
Cover or ignore the options and generate your own word or category first. Then read all options and pick the one closest to your prediction. If two options seem to fit, look for the precise shade of meaning (degree, connotation, part of speech) that separates them.
Ignoring the qualifier words in verbal items, such as 'not', 'except', 'most nearly', 'always', and 'best'.
Under time pressure, candidates skim for content words and skip small function words. These qualifiers flip the task, so a correct-sounding answer becomes wrong.
What to do instead
Adopt a fixed scan habit: before answering, circle or mentally flag any negative or superlative word in the prompt and restate the task to yourself in one phrase ('find the one that does NOT belong'). Build this into a checklist you rehearse on practice sets until it is automatic.
Setting up number-series or quantitative-comparison problems by guessing the next term before checking the differences between consecutive terms.
Patterns like +3, +5, +7 are easy to spot, so candidates jump to arithmetic and miss alternating, multiplicative, or two-interleaved-sequence rules. When the first guess fails, they assume the item is unfair rather than trying another structure.
What to do instead
Use a fixed ladder of tests in order: first differences, second differences, ratios, then check whether odd and even positions form separate sequences, then check for squares, primes, or doubling-plus-one rules. Write the differences under the series so the structure is visible rather than held in memory.
Estimating quantitative comparison items by rough feel instead of reasoning about the quantities, especially when variables or fractions are involved.
Items are designed so the two columns look close, and mental arithmetic on fractions or percentages is error-prone. Candidates also forget that a variable can be zero, negative, or a fraction, so they test only one convenient value.
What to do instead
For comparisons, test at least three values: a positive integer, a proper fraction between 0 and 1, and a negative or zero value where allowed. If the columns' relationship changes across those tests, the answer is 'cannot be determined'. For pure arithmetic, convert fractions to a common denominator or to decimals before comparing.
Losing track of the question in multi-step word problems and answering an intermediate quantity instead of the quantity requested.
Word problems require building a model, and the last number computed feels like the answer. Candidates also skip writing units, so they cannot tell whether a total, a rate, or a difference is being asked for.
What to do instead
Before calculating, write the target quantity and its units at the top of your scratch space (for example, 'cost per ticket, in dollars'). After computing, check the units of your result against that target and reread the final sentence of the prompt to confirm you are answering what was asked.
Rotating or flipping a figure mentally in a paper-folding or figure-transformation item and choosing an answer that is a mirror image rather than a true rotation.
Mental rotation is effortful and the visual system takes shortcuts, so a mirrored version of the correct shape feels identical to the rotated original. Candidates rarely verify by tracking one asymmetric feature.
What to do instead
Track a single distinctive landmark, such as a notch, a shaded corner, or a longer arm, through each fold or turn, and note its final position relative to the whole shape. Then check that the chosen option places that landmark in exactly that spot, which rules out mirror images.
Solving figure-matrix and classification items by overall visual similarity rather than by the specific rule that changes across rows or columns.
The eye groups shapes by how alike they look, so candidates pick the option that 'matches the style' of the other figures while ignoring the attribute that actually varies, such as number of sides, shading, or position.
What to do instead
Name the attributes aloud or in writing before choosing: shape type, count, size, shading, orientation, and position. Determine which attribute changes along each direction and which stays constant, then select the option that satisfies both rules. Confirm the answer against every figure in the matrix, not just one.
Leaving non-verbal items blank when the rule is not immediately visible, rather than eliminating options or making a reasoned choice.
Nonverbal items can feel like all-or-nothing insight, so candidates freeze or skip. Because there is no penalty for guessing on this exam, a blank is strictly worse than an eliminated-choice guess.
What to do instead
Use a fallback routine: eliminate any option that violates a confirmed rule, then choose among the rest by the attribute most likely to matter (count, then shading, then orientation). Answer every item before time expires, and flag uncertain ones only for a quick second look if time allows.