NNAT3 Level D common mistakes
NNAT3 Level D · Understand the traps and how to avoid them.
On Pattern Completion items, the candidate picks the option that matches only the most prominent feature of the missing patch (a colour block or a single line), then moves on without checking the remaining edges.
The missing region is large and the eye is drawn to the visually dominant element, so a single match produces a false sense of closure. Under time pressure the candidate stops scanning once one feature lines up.
What to do instead
Adopt a fixed checklist: trace each boundary segment of the hole in the main figure (top, bottom, left, right, and any interior lines that cross the hole) and confirm the option continues every one of them. Only after all boundary lines are verified should colour and shading be checked. Practise by covering the answer choices, sketching the expected patch from the surrounding edges, then comparing.
On Pattern Completion, the candidate ignores the background grid or pattern that runs behind the missing area and selects a patch whose internal lines are the right shape but offset by one cell.
The pattern across the whole figure is regular but not always aligned to the hole's edges, so a patch can look locally plausible while breaking the global rhythm. Candidates focus on the hole instead of stepping back.
What to do instead
Before looking at options, extend the repeating pattern mentally across the hole and note where lines should fall relative to the hole's corners. Then reject any option whose spacing is shifted. A quick check: count the number of repetitions from the left edge to the hole and predict the phase of the pattern at the hole.
On Reasoning by Analogy, the candidate identifies a relationship between the first pair that is too specific, such as 'the arrow points left', and applies that literal feature to the second pair instead of the underlying transformation.
Concrete surface features are easier to notice than relational ones, and the first pair often shares several incidental similarities that tempt a literal rule. This is especially likely when the correct rule is a rotation or a count change rather than a simple addition.
What to do instead
After spotting a candidate rule, state it in words as a general operation ('the shape rotates 90 degrees clockwise and gains one dot') and test it against the first pair in reverse. Then apply it to the third figure and check that exactly one option results. If two options fit, the rule is too loose and must be tightened.
On Reasoning by Analogy, the candidate assumes the transformation goes from left to right only and never considers that the second pair might require the inverse operation, producing an option that is the mirror of the correct answer.
Reading order biases candidates toward forward application. When the third figure is the result rather than the input, the direction reverses and the same rule produces a different answer.
What to do instead
Explicitly label the third figure as input or output before choosing. If it is the output, invert the rule (rotate counter-clockwise instead of clockwise, subtract instead of add) and verify against the first pair. Train with pairs where the blank sits in different positions.
On Serial Reasoning, the candidate tracks one changing attribute (for example, the number of sides) across the series but ignores a second attribute that changes on a different cycle, such as colour alternating every step while shape size grows every two steps.
Series items commonly combine two or three independent progressions, and the most obvious one captures attention. The candidate confirms the first rule against the options and stops, satisfied that one option matches.
What to do instead
For every series, list each attribute separately (shape, count, orientation, fill, position) and write its sequence across the given frames. Determine the step for each, then predict the missing frame on all attributes at once. Reject any option that satisfies only one list.
On Serial Reasoning, the candidate extrapolates a linear step when the series actually accelerates, decelerates, or repeats cyclically, so the predicted next frame is one step off.
Most practice series use constant steps, creating an expectation of arithmetic progression. Series with doubling counts or a repeating cycle of period three break that expectation and the error is invisible unless the differences are computed.
What to do instead
Compute the successive differences or ratios between frames rather than eyeballing the trend. If differences are not constant, check whether they themselves form a pattern (increasing by one, doubling) or whether the series cycles. Confirm the cycle length by checking that frame four equals frame one when a period of three is suspected.
On Spatial Visualization, the candidate mentally rotates a figure but keeps the original orientation of an internal detail, such as a dot that should move from the top corner to the right corner, and selects an option that is a reflection rather than a rotation.
Mental rotation is approximate; the overall outline is easy to rotate but asymmetric internal features are lost. Reflections are especially seductive because they preserve shape and size and differ only in chirality.
What to do instead
Pick one asymmetric landmark on the figure (a notch, a dot, an unequal arm) and track only that landmark through the rotation, then verify a second landmark. Distinguish rotation from reflection by checking that the clockwise order of features is preserved; a reflection reverses it. Physically turning the page or using a finger to mark the landmark helps.
On Spatial Visualization, the candidate folds or unfolds a net by imagining the whole sheet at once and loses track of which face ends up adjacent to which, producing an answer where two faces that should be opposite are shown touching.
Folding involves several simultaneous spatial transforms, and working memory drops intermediate states. Without an anchor face, the candidate cannot reconstruct the final arrangement reliably.
What to do instead
Assign one face as the base and fold the net one panel at a time, naming the position of each panel as it lands (front, top, right). Note that faces separated by exactly one panel in a straight line become opposite each other. Verify the final answer by unfolding it back to the net.
The candidate spends disproportionate time on one hard item, leaving too little time for later items, and rushes the final questions where accuracy drops sharply.
The test is timed and items are not ordered strictly by difficulty, so an early hard item can consume the budget. Candidates also feel committed to finishing an item they have already invested in.
What to do instead
Set a per-item ceiling (roughly 45 to 60 seconds at this level), mark and move on when it is exceeded, and return only if time remains. Practise with a visible timer in blocks of ten items to build the habit of abandoning and returning.
The candidate answers from the first impression and does not eliminate wrong options, so on four-option items with two close distractors the choice is effectively a coin flip.
Recognition feels faster than analysis, and the correct answer is often the one that looks most familiar. Close distractors are designed to differ by a single attribute that only elimination reveals.
What to do instead
Force a two-pass routine: first strike out every option that violates a stated rule, then compare the survivors attribute by attribute against the prediction made before looking at the options. If two remain, the rule was underspecified; return to the figure and find the attribute that separates them.