Nonverbal Reasoning
What the Nonverbal Reasoning battery measures
The Nonverbal Reasoning battery assesses reasoning with shapes and figures rather than words or numbers. The upper-level subtests are Figure Matrices, Figure Classification, and Paper Folding; primary levels use picture analogies and classifications. Because the content is visual, the exam is testing whether you can isolate a transformation or shared attribute and apply it precisely.
Core rules
- Figure matrix. Compare figures across rows and down columns. Track several features separately: shape, size, fill or shading, orientation, number of sides, number of dots, and position. The rule must hold for the whole row or column, not just one pair.
- Figure classification. Name a property shared by all three stem figures, then test each option. A good rule fits every stem and exactly one answer. Prefer a rule that uses all the visible features rather than a single accidental match.
- Paper folding. Follow the folds in order, punch the holes, then unfold in reverse order. Each fold line acts as a mirror: a hole produces a matching hole on the opposite side of that fold. Count holes created by every layer, and watch for holes that fall on a fold line or overlap.
Worked example
A 2×2 matrix has a small filled circle at top-left, a large filled circle at top-right, and a small unfilled square at bottom-left. What belongs at bottom-right?
Across the top row, size changes from small to large while shape and fill stay the same. Down the left column, shape changes from circle to square and fill changes from filled to unfilled. Applying both rules to the missing cell gives a large unfilled square. A large filled square would copy the size rule but ignore the fill change; a small unfilled square would copy the column rule but ignore the row rule.
Traps that cost marks
- Fixating on one feature (size) and missing a second feature (fill or orientation).
- Confusing rotation with reflection; a mirrored figure cannot be produced by turning it.
- In paper folding, counting each punch once even though it passes through several layers.
- In classification, choosing an option that shares one property with the stems but not the defining rule.
- Assuming the row rule must also be the column rule.