Four letter-clusters have been given, out of which three are alike in some manner and one is different. Select the one that is different. The odd one out is not based on the number of consonants/vowels or their position in the letter cluster.

Correct — C. Three of the four clusters walk through the alphabet in exactly the same way.

Rule: +2, +2, +1 on alphabet positions.

KMOP: K(11) → M(13) → O(15) → P(16)
NPRS: N(14) → P(16) → R(18) → S(19)
EGIJ: E(5) → G(7) → I(9) → J(10)

UVXW runs U(21) → V(22) → X(24) → W(23), i.e. +1, +2, −1. Its last letter steps backwards while the other three clusters only move forward, so option (c) is the odd one.

Why the others are wrong

• (a)
  KMOP — KMOP steps K → M → O → P, gaps +2, +2, +1. That is the same walk as NPRS and EGIJ, so it sits inside the group of three rather than outside it.
• (b)
  NPRS — NPRS steps N → P → R → S, gaps +2, +2, +1 — identical to KMOP and EGIJ. Every letter moves forward, which is what the outlier fails to do.
• (d)
  EGIJ — EGIJ steps E → G → I → J, gaps +2, +2, +1. Having two vowels does not make it the outlier; the stem rules vowel counting out and the gap pattern still matches.

Odd-one-out on letter clusters is almost never about meaning. Convert each letter to its alphabet position and the four clusters become four number strings.

Then read the gaps, not the letters. Three strings will share one gap pattern and the fourth will break it.

Here the stem removes the two shortcuts candidates reach for first, vowel count and vowel position, which leaves the gap walk as the thing to test.

Key facts

• Alphabet positions in KMOP: K=11, M=13, O=15, P=16 — gaps of 2, 2 and 1.
• UVXW reads U=21, V=22, X=24, W=23, so its third step goes backwards by one.
• The stem bars grouping by number or position of vowels and consonants.

Common traps

• Seeing U, V, W, X and calling UVXW alphabetical without reading the order it is printed in.
• Grouping by vowel count, which the stem forbids and which here gives 1, 0, 1 and 2 — no three-and-one split anyway.
• Computing gaps from the wrong end, which turns every +2 into a −2 and hides the break.

No directly related past PYQ was found.