CogAT Nonverbal common mistakes
CogAT Nonverbal · Understand the traps and how to avoid them.
On Figure Matrices, checking only one attribute (e.g., shape) and picking the first answer that matches it, without verifying the second attribute (shading, size, orientation, count).
Figure Matrices items are built from two simultaneous transformations across rows and columns; the eye locks onto the most salient change (usually shape) and the brain treats a partial match as a full match, especially under time pressure.
What to do instead
Before looking at the answer choices, state the rule aloud in a full sentence covering both attributes ('down the column the shape stays the same but the fill flips; across the row the count increases by one'). Then eliminate choices one attribute at a time, crossing off any option that violates either rule.
Treating Paper Folding questions as a mental animation of the folding process instead of a static symmetry problem, then losing track of how many holes should appear and where.
Candidates try to simulate each fold in sequence under time pressure, which overloads working memory; the sequence is harder than the end state, and one dropped fold produces a wrong hole count.
What to do instead
Ignore the folding sequence once you have the punch location. Work backwards: each fold line acts as a mirror, so reflect the punch across the most recent fold, then across the previous one, doubling the hole count at each step. Count holes first, eliminate answers with the wrong count, then check mirror positions.
On Figure Classification, grouping the three stem figures by overall visual similarity or theme ('these look like arrows') rather than by a rule that holds for all three and excludes the other two.
Classification items are designed so that surface similarity is a trap; the correct rule is often abstract (rotational symmetry, odd vs. even number of sides, closed vs. open figures), and candidates stop at the first plausible grouping.
What to do instead
Generate at least two candidate rules for the stem set and test each against all five options. The correct rule must include exactly three stem figures and exclude exactly two options. If a rule fits four or five options, it is wrong; keep searching for the discriminating rule.
On Paper Folding and Figure Analysis, forgetting that a fold doubles the number of holes and that a punch on a fold line creates a single merged hole rather than two.
The doubling rule is stated in instruction but not practiced enough; edge cases (punch exactly on the crease, punch on the folded edge) are the ones most often missed, and they change the answer count.
What to do instead
Drill the edge cases explicitly: punch on a crease = one hole shared by both layers; punch inside a folded region = two holes mirrored across the crease. Before answering, write the expected hole count as a number and match it to the options.
On Number Analogies and Number Series, solving by arithmetic trial and error on the answer choices instead of identifying the relationship in the stem pair first.
Analogies present a completed pair (A:B) plus a partial pair (C:?), and candidates jump to testing operations on C without extracting the A-to-B rule; series items tempt guessing the next term by feel rather than by difference pattern.
What to do instead
For analogies, write the A-to-B operation as an equation (e.g., 'B = 2A + 1') before touching C. For series, compute successive differences (and second differences if the first are not constant) on paper; only then predict the next term and check it against the options.
On Sentence Completion and Verbal Analogies, choosing an answer that sounds academic or sophisticated rather than one that fits the sentence's logic or the stem pair's relationship.
Distractors are written to be plausible-sounding and semantically adjacent; candidates who lack a precise definition default to word familiarity as a proxy for correctness, which is exactly the trap the item is built around.
What to do instead
Cover the answer choices, predict your own word from context, then match. For analogies, label the relationship in a short phrase ('part to whole', 'cause to effect', 'degree of intensity') and require the correct option to satisfy the same label, not just the same topic.
On Verbal Classification, rejecting the correct answer because it is a less obvious member of the category, while accepting a familiar word that belongs to a different category.
Classification requires finding the one word that does not share the stem's category; candidates often rely on the most common meaning of a word and miss a secondary sense, or they pick the odd one out on a superficial association.
What to do instead
For each stem word, list its category membership in one or two words. The correct answer is the one that fails every category the other four share. If two options seem to fit, look for a narrower shared property (e.g., all are units of measure, all are metals) that only one option violates.
Spending too long on a single hard item and running out of time, leaving later items unanswered even though they may be easier.
CogAT is timed per battery, and hard items early in a section create a sunk-cost effect; candidates keep reworking one figure matrix while easy items at the end go blank, and unanswered items score the same as wrong ones.
What to do instead
Set a per-item budget (roughly 45-60 seconds on a 20-item section). If no answer is clear after one full pass of elimination, mark your best guess, flag it mentally, and move on. Return only if time remains; never leave a bubble empty.
Reading only the first answer choice that looks right and stopping, rather than evaluating all options to find the best fit.
On quantitative comparison and analogy items, the first plausible option is often a near-miss distractor; candidates who stop early systematically pick the first attractive answer, which correlates with lower scores even when their reasoning is sound.
What to do instead
Adopt a fixed routine: read all options, eliminate at least two, then choose. On number analogies, verify your chosen rule against the stem pair one more time before marking the answer.
Failing to transfer a rule learned in one item type to a structurally identical item, so the same error repeats across a section.
CogAT reuses item formats (matrices, folding, series) with different surface content; candidates treat each item as novel and do not build a reusable checklist, so working memory is consumed re-deriving the same method each time.
What to do instead
After each practice section, write a one-line checklist for that item type (e.g., 'Matrix: check shape, fill, count, orientation, position'). Run the checklist on every item of that type until it becomes automatic, and review only the items where the checklist caught an error.