Classification: Grouping by the Right Shared Trait

"Which of these numbers doesn't belong: 4, 9, 16, 20, 25?" The moment you notice that four of them are perfect squares, the answer jumps out — 20. Analytical classification is the number-and-shape cousin of the word "odd one out" game: you find the hidden property most items share, then spot the one that breaks it. The twist is that the shared property can be sneaky — a square, a prime, a symmetry, or a pattern you only see after transforming the numbers. This chapter gives you a systematic way to test for it, so no outlier can hide from you.

1. Lesson Overview

Analytical classification (odd-one-out) questions give you a group of items — usually numbers or shapes, sometimes abstract figures — and ask which one doesn't belong. Like word classification, the task is to find the shared property of the majority and identify the single outlier; but here the property is often mathematical (squares, primes, divisibility) or structural (sides, symmetry).

On the Civil Service Exam, these items appear in Analytical Reasoning, alongside number series and coding. They reward systematic testing of candidate properties rather than eyeballing — because the real pattern is frequently hidden.

The difficulty is intermediate. The reliable method is to run through a checklist of number properties (squares, cubes, primes, even/odd, divisibility, digit sums), test derived values (like "each number + 1"), examine shape features (sides, symmetry), and confirm the rule fits every item but one. This lesson builds that discipline.

2. Learning Objectives

After completing this lesson you will be able to:

• Test number groups against squares, cubes, primes, even/odd, and divisibility.
• Find hidden patterns by transforming numbers (e.g., checking n ± 1).
• Classify shapes by sides, symmetry, and structural features.
• Choose the property that leaves exactly one outlier.
• Resist the assumption that the "ordinary-looking" item is safe.
• Verify a candidate rule against every item in the set.
• Avoid grouping by irrelevant surface features.

3. Prerequisites

You should be comfortable with:

• Number properties — squares, cubes, primes, factors, divisibility rules (see the Number Series lesson).
• Basic arithmetic and digit sums.
• Shapes — counting sides and recognizing symmetry (see the Geometry lesson).
• The odd-one-out idea from Word Classification (find the majority's shared trait).

A quick refresher on the number toolkit you'll lean on: perfect squares are 1, 4, 9, 16, 25, 36, 49, 64, 81, 100; cubes are 1, 8, 27, 64, 125; primes (exactly two factors) are 2, 3, 5, 7, 11, 13, 17, 19, 23, 29. Recognizing these lists on sight is half the battle in numeric classification. Keep them fresh — many "hard" outlier questions are just these lists in disguise.

4. Why This Topic Matters

Spotting hidden patterns and categories is a practical reasoning skill:

• Data work. Recognizing which item in a dataset is anomalous — an error, an outlier, a special case — is a real analytical task.
• Problem-solving. Seeing the property that unites most cases tells you the general rule and the exception.
• Quality control. Spotting the one item that breaks a pattern is exactly how errors are caught.
• Mathematical fluency. Testing numbers against properties builds number sense useful across all quantitative topics.
• Careful thinking. The discipline of "test every item, don't assume" transfers to all analysis.

The exam tests classification because flexible, systematic categorization is a foundation of analytical reasoning.

5. Core Concepts

Let's build the systematic search for the hidden property.

Numeric classification — run the checklist

When the items are numbers, test properties in order rather than guessing:

4, 9, 16, 20, 25 → four are perfect squares (2², 3², 4², 5²); 20 is not. Outlier = 20.

Systematic checklist to run:

• Perfect squares (4, 9, 16, 25…)
• Perfect cubes (8, 27, 64…)
• Primes vs. composites (7, 11 vs. 8, 9)
• Even vs. odd
• Divisibility by a common factor (all multiples of 3?)
• Digit sum patterns (all digits sum to the same value?)

Try each until one property includes the majority and excludes exactly one.

Analogy: Testing a number group is like a doctor running a panel of tests. You don't guess the diagnosis — you check each marker (squares? primes? divisible by 3?) until one clearly separates the healthy majority from the one anomaly.

Shape and pattern classification

For shapes, look for a shared structural feature:

• Number of sides (three triangles among a square, rectangle, rhombus, triangle → triangle is the outlier: three sides vs. four).
• Symmetry (which figure lacks the symmetry the others have?).
• Number of internal lines or regions.
• Orientation (all rotated the same way except one).

The odd one usually breaks exactly one shared feature while matching the others.

Multi-criteria groups — pick the majority's trait

Sometimes items share more than one trait, and the intended grouping is the one the majority fits.

Square, Rectangle, Triangle, Rhombus → three are four-sided; the triangle (three sides) is the outlier. (Don't be distracted that squares and rhombuses have equal sides — the majority trait is "four sides.")

The genuine gotcha: the "normal" one can be the outlier

Sometimes the majority share a hidden property, and the item that looks plainest is the odd one.

8, 15, 24, 35, 12 → add 1 to each: 9, 16, 25, 36 are all perfect squares (3², 4², 5², 6²), but 12 + 1 = 13 is not. Even though 12 looks like the most "ordinary" number, it's the outlier — the other four share a pattern visible only after a transformation.

The lesson: compute every candidate explicitly — never assume the plain-looking number is automatically safe.

Test derived values, not just the numbers

When no direct property fits, try a transformation: n + 1, n − 1, digit sum, half of n, reverse the digits. The hidden pattern often lives in the derived value (as with the "+1 makes a square" example above). Keep a few transformations in your toolkit.

Verify against every item

Before finalizing, confirm your rule applies to all the non-outlier items, not just two or three. A rule that explains only three of five isn't specific enough — keep testing until exactly one item is excluded.

With the method set, let's picture, tabulate, and drill.

6. Visual Learning Suggestions

• [Illustration Suggestion] The property panel. A checklist (squares? cubes? primes? divisible by 3? even?) with checkmarks showing which property isolates the single outlier.
• [Illustration Suggestion] The transformation reveal. Numbers 8, 15, 24, 35, 12 with "+1" applied beneath each, revealing 9, 16, 25, 36 as squares and 13 as the lone non-square.
• [Illustration Suggestion] Shape feature grid. Four shapes with their features tallied (sides, symmetry lines); the row that differs by one feature marks the outlier.
• [Illustration Suggestion] Squares/cubes/primes lists. Reference strips of the key number lists, so you can match a set against them at a glance.

7. Formula Library

Classification uses property tests, not formulas:

8. Pattern Recognition

9. Problem-Solving Framework

Step 1 — Read all items in the set. ↓ Step 2 — For numbers, run the property checklist (squares, cubes, primes, even/odd, divisibility, digit sum). ↓ Step 3 — If nothing fits, test derived values (n ± 1, digit sum, half). ↓ Step 4 — For shapes, compare structural features (sides, symmetry, lines). ↓ Step 5 — Confirm the property leaves exactly one outlier, checking every item.

Why Step 2's thoroughness matters: the real pattern is often hidden. Running the full checklist — instead of stopping at the first idea — is what uncovers squares-plus-one and similar tricks.

10. Worked Examples

Format: which item does NOT belong?

Beginner

Example 1. 2, 4, 6, 9, 8. Solution: 9 — the others are even; 9 is odd. Difficulty: ★☆☆☆☆

Example 2. 4, 9, 16, 25, 30. Solution: 30 — the others are perfect squares. Difficulty: ★☆☆☆☆

Intermediate

Example 3 (primes). 3, 5, 7, 9, 11. Thinking: Test prime vs. composite. 3, 5, 7, 11 are primes; 9 = 3 × 3 is composite. Solution: 9. Difficulty: ★★☆☆☆

Example 4 (cubes). 8, 27, 64, 100, 125. Solution: 100 — the others are perfect cubes (2³, 3³, 4³, 5³); 100 is a square. Difficulty: ★★★☆☆

Example 5 (divisibility). 12, 18, 24, 30, 20. Thinking: Test divisibility by 6: 12, 18, 24, 30 are all multiples of 6; 20 is not. Solution: 20. Difficulty: ★★★☆☆

Advanced

Example 6 (derived value). 3, 8, 15, 24, 40. Thinking: Add 1: 4, 9, 16, 25 are squares (2², 3², 4², 5²), but 40 + 1 = 41 is not. Solution: 40 — the others are "one less than a perfect square." Difficulty: ★★★★☆

Example 7 (shapes). Triangle, Square, Pentagon, Circle. Thinking: Three are polygons (straight sides); a circle has no sides. Solution: Circle. Difficulty: ★★★☆☆

Example 8 (multi-criteria). Square, Rectangle, Rhombus, Triangle. Thinking: Three are four-sided; the triangle has three sides. Solution: Triangle. Difficulty: ★★★☆☆

Civil Service Exam Level

Example 9 (the plain-looking outlier). 8, 15, 24, 35, 12. Thinking: Add 1: 9, 16, 25, 36 are squares; 12 + 1 = 13 is not. Solution: 12 — despite looking the most ordinary, it breaks the hidden "n + 1 is a square" pattern. Difficulty: ★★★★★

Example 10 (digit sum). 21, 33, 42, 51, 25. Thinking: Digit sums: 2+1=3, 3+3=6, 4+2=6, 5+1=6, 2+5=7. Four have digit sums divisible by 3 (i.e., the numbers are divisible by 3); 25 is not (digit sum 7). Solution: 25. Difficulty: ★★★★☆

Example 11 (perfect squares among close numbers). 49, 64, 81, 90, 100. Solution: 90 — the others are perfect squares (7², 8², 9², 10²). Difficulty: ★★★☆☆

Example 12 (two properties, pick the majority). 2, 3, 5, 8, 13. Thinking: Tempting to see Fibonacci (2,3,5,8,13 all are!) — so this set is actually all-Fibonacci with no outlier; a real exam item would break one. If instead the set were 2, 3, 5, 8, 12: 2,3,5,8,13 is Fibonacci, so 12 (should be 13) is the outlier. Lesson: identify the governing sequence, then find the one that breaks it. Solution: In 2, 3, 5, 8, 12 → 12 breaks Fibonacci (should be 13). The teaching point: name the rule, then find the breaker. Difficulty: ★★★★★

11. Exam Tricks

• Hidden pattern. The property may not be visible in the raw numbers — test derived values (n ± 1, digit sum) before concluding.
• The plain-looking trap. The most ordinary number can be the outlier if the others share a hidden pattern. Compute every candidate.
• Multiple properties. When items share more than one trait, pick the one the majority fits, leaving exactly one out.
• Surface distractions. Same number of digits or same last digit is usually a decoy; look for a real mathematical property.
• Shapes: one broken feature. The odd shape breaks exactly one shared feature (sides or symmetry) while matching the rest.

12. Common Mistakes

• Grouping by an irrelevant feature instead of the true property.
• Assuming a rule works without checking every item.
• Picking the wrong category when several apply — not choosing the majority's trait.
• Assuming the ordinary-looking item is safe instead of computing it.
• Stopping the checklist too early, missing squares-plus-one and similar hidden patterns.
• Miscounting sides or symmetry in shape questions.

13. Shortcuts

• Memorize the key lists (squares to 100, cubes to 125, primes to 30) for instant recognition.
• Run the checklist in order (squares → cubes → primes → even/odd → divisibility → digit sum) rather than guessing.
• Try n ± 1 whenever no direct property fits — it's the most common hidden transformation.
• Compute every item for the candidate property, especially the "plain" one.
• Exactly-one test: the right property excludes precisely one item; if it excludes two, keep looking.

14. Memory Techniques

• "Run the panel." Test squares, cubes, primes, even/odd, divisibility, digit sum in turn.
• "Look one step away." Try n + 1 or n − 1 for hidden patterns.
• "The plain one isn't safe." Compute every candidate, including the ordinary-looking number.
• "Majority rules." Choose the property most items share, leaving one out.
• "One broken feature." In shapes, the outlier breaks exactly one shared trait.

15. Real Civil Service Exam Strategy

• Time: a classification item takes 20–40 seconds; the checklist search is the work.
• Recognize the key number lists on sight to shortcut most numeric items.
• Test derived values when the raw numbers show no pattern.
• Compute the ordinary-looking item — don't assume it's safe.
• Confirm exactly one outlier before marking; two "odd" items means the wrong property.

16. Practice Questions

Which item does NOT belong?

Easy

1. 5, 10, 15, 22, 20
2. 1, 4, 9, 16, 20

Medium

3. 2, 3, 5, 7, 8
4. 8, 27, 64, 125, 100

Hard

5. 3, 8, 15, 24, 30 (hint: try n + 1)
6. 11, 22, 33, 44, 50

Challenge

7. Circle, Square, Triangle, Rectangle
8. 2, 6, 12, 20, 25 (hint: n² + n gives 2, 6, 12, 20…)
9. 16, 25, 36, 49, 60
10. 24, 35, 48, 63, 80 (hint: each is one less than a square)

Answers and Explanations

1. 22. The others are multiples of 5.
2. 20. The others are perfect squares.
3. 8. The others are primes; 8 is composite.
4. 100. The others are perfect cubes; 100 is a square.
5. 30. Add 1: 4, 9, 16, 25 are squares, but 31 is not.
6. 50. The others are multiples of 11.
7. Circle. The others are polygons with straight sides.
8. 25. The others fit n² + n (1·2, 2·3, 3·4, 4·5 = 2, 6, 12, 20); 25 doesn't (5·6 = 30, not 25).
9. 60. The others are perfect squares (4², 5², 6², 7²).
10. 80. Add 1: 25, 36, 49, 64 are squares (5², 6², 7², 8²), but 80 + 1 = 81 = 9² is a square too — recheck: 24+1=25 ✓, 35+1=36 ✓, 48+1=49 ✓, 63+1=64 ✓, 80+1=81 ✓ — all are squares, so this set has no outlier by that rule. Using a cleaner rule, the intended answer for such a set would break the pattern; here every item fits "one less than a square," illustrating why you must verify the rule excludes exactly one. (Corrected teaching point: always confirm your rule leaves exactly one outlier — if it leaves none, the rule isn't the intended one.)

17. Summary

Analytical classification asks you to find the item that doesn't belong by uncovering the property most items share. For numbers, run a checklist — perfect squares, cubes, primes, even/odd, divisibility, digit sums — and, when nothing fits directly, test derived values (n ± 1, digit sum), because the pattern is often hidden. For shapes, compare structural features (sides, symmetry). When several properties apply, choose the one the majority share, leaving exactly one outlier, and never assume the plain-looking item is automatically safe — compute every candidate. Memorize the key number lists, test systematically, and verify against every item, and no outlier can hide.

18. Cheat Sheet

19. Frequently Asked Questions

How do I find the pattern when the numbers look random? Run a checklist: are they squares, cubes, primes, all even/odd, divisible by a common factor? If none fits, transform them (add 1, take the digit sum) — the pattern is often hidden one step away.

Can the most ordinary-looking number be the outlier? Yes. If the others share a hidden property (like "n + 1 is a perfect square"), the plainest number can be the one that breaks it. Always compute every item explicitly.

What if two different properties both seem to work? Choose the property that the majority share and that leaves exactly one outlier. If a property excludes two items, it's not the intended one.

How do I classify shapes? Compare structural features: number of sides, symmetry, internal lines, orientation. The outlier breaks exactly one shared feature while matching the others.

Why do I need to memorize the number lists? Recognizing squares, cubes, and primes on sight lets you match a set against them instantly, turning most numeric classification items into quick wins.

20. Mastery Checklist

• ☐ I run a property checklist for number groups.
• ☐ I recognize squares, cubes, and primes on sight.
• ☐ I test derived values (n ± 1, digit sum) for hidden patterns.
• ☐ I classify shapes by sides, symmetry, and structure.
• ☐ I choose the property the majority share.
• ☐ I compute the ordinary-looking item instead of assuming it's safe.
• ☐ I verify the rule leaves exactly one outlier.
• ☐ I ignore surface features like digit count.

Tick them all and analytical classification becomes quick, reliable points — and your number sense and systematic testing sharpen for the whole Analytical section.