Skip to main content
Analytical Reasoning 3% exam weight

Classification and Odd One Out

Part of the NAT-I (NTS) study roadmap. Analytical Reasoning topic ar-6 of Analytical Reasoning.

By Last updated 3% exam weight

Classification and Odd One Out

🟢 Lite — Quick Review (1h–1d)

Rapid summary for last-minute revision before your exam.

Classification groups items by a shared attribute (mammals, vowels, quadrilaterals). Odd-one-out isolates the single option that breaks the dominant property. NAT-I items present 4–5 items; the correct answer is the one failing the primary shared trait, not a peripheral one.

  • Definition: Identify the rule uniting 3+ items, then flag the option violating it.
  • Four attribute families tested: semantic (meaning), structural (shape/letter pattern), numerical (digit sum, parity), functional (units, chemical families).
  • Speed tip: Read all five options before choosing; a second pattern can look valid but NAT-I keys use the most syllabus-standard attribute.
Attribute familyExample setOdd option
Semantic (fruits)mango, apple, banana, pearcarrot
Structural (closed shapes)circle, square, triangle, ovalcross
Numerical (even digits)2, 4, 6, 89

🟡 Standard — Regular Study (2d–2mo)

Standard content for students with a few days to months.

Core mechanism

Classification questions resolve into two operations: attribute extraction (find the property shared by ≥3 of 4–5 items) and membership test (decide whether a candidate item satisfies it). Odd-one-out reverses the test — eliminate options that match, isolate the one that does not. The attribute is rarely named in the stem, so the solver must infer it from co-occurrence.

Attribute families in NAT-I practice

NTS item banks draw attributes from four families. Semantic covers categories like professions, vehicles, currencies, and body organs. Structural covers letter patterns (vowels, consonants, symmetrical letters), word patterns (rhymes, prefixes), and shape properties (open vs closed, sides count). Numerical covers digit sums, divisibility by 3 or 5, prime status, and square/cube roots. Functional covers units of measurement, chemical families, and living-versus-non-living divisions.

FamilyDiagnostic questionCommon trap
SemanticDo all items share a category name?Two items belong to overlapping sets
StructuralIs the pattern in letters, shape, or order?Counting syllables instead of letters
NumericalIs digit sum, parity, or factor at play?Confusing sum with product
FunctionalDo they perform the same role or unit?Mixing SI with imperial units

Elimination discipline

  • Read every option before committing; the odd item can sit in the middle of the list.
  • State the attribute aloud in one word (e.g., “fruits”, “even”, “vowels”); if it fails on more than one option, your attribute is wrong.
  • When two patterns both yield an odd item, prefer the attribute requiring the fewest assumptions — in standard papers the literal or numerical one over a metaphorical reading.

🔴 Extended — Deep Study (3mo+)

Comprehensive coverage for students on a longer study timeline.

Edge cases and overlapping categories

The hardest NAT-I items exploit category overlap: “orange” is both a fruit and a colour; “mercury” is both a planet and a metal; “March” is both a month and a verb. The stem in most keys disambiguates (“which is a fruit”), but when it does not, the rule is to choose the classification requiring fewer leaps. Another edge case is second-order attributes — items that share surface category (all animals) but diverge on a deeper property (mammal vs reptile, vertebrate vs invertebrate). Examiner traps target this layer.

Classification feeds directly into analogy (A:B :: C:?) and coding-decoding (letter-number shifts), because all three rest on attribute mapping. Mastering the four attribute families above transfers cleanly into series completion and syllogism premises.

MistakeWhy it loses marksFix
Picking on a trivial differenceWrong attribute layerRecount matches; majority rule wins
Mixing question directionGroup belongs vs does not belongRe-read the verb: “fits” vs “does not”
Chasing a cleverer patternNAT-I keys the obvious oneState attribute in one literal word
Ignoring proper-noun statusMisses technical classCheck for symbols, abbreviations, units

Practice prompts

  1. From {sparrow, bat, penguin, dolphin, ostrich} identify the odd one out on a biological attribute and justify why “penguin” or “bat” is not a stronger answer.
  2. Given {2, 3, 5, 7, 9} state two valid odd-one-out answers using different attribute families, then defend which one NAT-I would mark.

Continue your study

Content adapted based on your selected roadmap duration. Switch tiers using the selector above.

Sources & verification