NNAT3 Level E common mistakes
NNAT3 Level E · Understand the traps and how to avoid them.
In Pattern Completion, choosing an answer that matches the overall colour scheme of the larger design but contradicts the internal line pattern at the missing patch's edges, instead of tracing each line and shape that crosses the patch boundary and matching its continuation.
The eye is drawn to global colour and texture, so a plausible-looking option feels right; candidates do not consciously project the lines that terminate at the gap's border, which is the only reliable way to reconstruct the missing piece.
What to do instead
Before looking at the options, trace every line and shape that touches the edge of the missing patch with a finger or pencil and note its direction, thickness, and colour. Then eliminate any option that breaks even one of those continuations, and only compare surviving options on fill and shading.
In Pattern Completion, treating a symmetric design as if the missing patch must be a mirror image of the opposite side, and selecting the mirrored option even when the design is rotationally symmetric rather than bilaterally symmetric.
Many practice designs are left-right symmetric, so candidates over-apply a mirror rule; they do not check whether the whole figure maps onto itself by rotation instead, which produces a different patch.
What to do instead
Test the design's symmetry explicitly: ask whether the figure looks identical when flipped left-right or when rotated 180 degrees. Use the symmetry type you actually observe to predict the patch, and treat mirror-image options as traps when rotation is the governing rule.
In Reasoning by Analogy, comparing the two top figures by overall appearance or by the number of elements rather than by the single transformation that relates them, so the chosen bottom pair applies a different relationship than the top pair demonstrates.
Analogy items often include a distractor pair that shares surface features with the prompt but uses a different rule; candidates who read the top pair holistically rather than naming the rule are easily captured by it.
What to do instead
State the top relationship in one sentence before viewing the options, for example 'the outer shape loses one side and the inner dot moves from left to right'. Then apply that exact sentence to the third figure and reject any pair that satisfies a different rule, however attractive.
In Reasoning by Analogy, assuming the transformation runs in the same direction as reading order and applying it from the wrong figure, which reverses the operation (for example shrinking instead of enlarging, or rotating clockwise instead of counter-clockwise).
The direction of the analogy is not always left-to-right in the way candidates assume, and some transformations are not their own inverse, so a reversed application yields a plausible but wrong answer.
What to do instead
Fix the direction explicitly: label the first figure A, the second B, the third C, and the answer D, and write the rule as 'A to B is X, so C to D is X'. Check by asking whether applying the inverse of X to B returns A; if not, the rule is stated backwards.
In Serial Reasoning, identifying the rule that governs one attribute (such as shape type cycling through three forms) and ignoring a second attribute that changes on a different cycle (such as shading or position), then picking an option that is correct on the first attribute only.
Sequences in this domain usually vary two or more attributes simultaneously, and the most visually salient attribute dominates attention; the candidate stops checking once one pattern fits.
What to do instead
For every sequence, list each attribute separately (shape, count, colour or shading, orientation, position) and predict its next value independently before looking at the options. Accept an option only if it matches every prediction; if two attributes conflict, re-derive the cycle length for the one that seems off.
In Serial Reasoning, extrapolating a progression linearly when it actually advances by an increasing step (for example counts rising by 1, then 2, then 3), or assuming a cycle restarts when it merely continues, producing an answer one step off.
Candidates fit the first visible difference and project it forward without checking it against the later terms of the sequence, and short sequences make a constant-step rule look adequate.
What to do instead
Compute the differences between consecutive terms and verify the rule against every term, not just the first pair. If the differences themselves change, look for a second-order pattern or a repeating cycle whose length divides the number of visible terms, and confirm the predicted term is consistent with all given terms.
In Spatial Visualization, mentally rotating a figure but losing track of chirality, so a mirror image is accepted as a valid rotation of the original when no rotation can produce it.
Mental rotation is easier for familiar or near-symmetric shapes, and the brain often 'snaps' a mirrored shape onto the target because it matches in outline; handedness is the attribute most easily lost in imagination.
What to do instead
Pick one asymmetric landmark on the figure, such as a notch or a single offset dot, and track its position relative to a second landmark through the rotation. If the clockwise ordering of three landmarks reverses, the option is a reflection, not a rotation, and must be eliminated.
In Spatial Visualization, folding or unfolding a net by matching edges that look adjacent on the page rather than edges that actually meet when folded, so faces are assigned to the wrong side of the assembled solid.
Paper nets are drawn flat, and the candidate does not simulate the fold sequence; shared vertices in the drawing are confused with shared edges after folding, and opposite faces are mistaken for adjacent ones.
What to do instead
Fold mentally one panel at a time from a fixed base face, naming which net edge joins which, and remember that panels separated by exactly one panel in a straight strip become opposite faces. Check each candidate solid against the adjacency list you built, not against the flat drawing.
In Spatial Visualization, counting cubes or blocks by tallying only the visible ones and ignoring hidden support cubes, so a stack is undercounted and the option matching the visible total is chosen.
Hidden cubes leave no visual trace, and the candidate does not apply the physical constraint that every cube above the base must rest on another cube; the visible count feels complete.
What to do instead
Count column by column from a fixed corner, and for each column record the height implied by the topmost visible cube, assuming a solid stack unless the item states otherwise. Sum the column heights rather than counting visible faces, and sanity-check the total against the outline of the structure.
Spending a fixed amount of time on every item regardless of difficulty, so a single hard spatial item consumes the time needed for several later items, and easy items at the end are left unanswered.
The test is timed and items are not ordered by difficulty within a domain, so a candidate committed to finishing in sequence can be stalled by one unfamiliar figure and never recover the pace.
What to do instead
Set a per-item time ceiling in practice (roughly 45 to 60 seconds), and when it is hit, mark the best current answer, move on, and return only if time remains. Practise with a visible timer so the ceiling becomes automatic, and always leave enough time to confirm that no item is blank, since there is no penalty for a reasoned guess.