Examiner Differentiation Logic in IELTS

Adjacent bands · Descriptor evidence · May 2026

Direct answer

Examiners distinguish adjacent IELTS bands by finding which public descriptor level your performance consistently proves—not by averaging impressions. Between Band 6 and 7, the decision often hinges on one criterion: shallow Part 3 caps FC, thin body paragraphs cap TR/CC, or error clusters cap GRA. Examiners ask: does this candidate sustain Band 7 behaviours under the whole test, or only flash them once?

Holistic marking: weakest criterion pulls down

IELTS scores are not four independent numbers averaged for fun—examiners notice when Fluency is 7 but Part 3 logic is Band 6. The lower evidence sets the ceiling.

Band 6 evidence Adequate communication; uneven development; errors sometimes impede

Band 7 evidence Sustained development; flexible language; errors rarely impede

Decision rule If two criteria sit at 6 and two at 7, overall skill often lands 6.5—not 7

Adjacent-band questions examiners ask

Question	Band 6 answer	Band 7 answer
Does the candidate sustain depth?	Lists, repeats, or circles	Develops with support
Is control flexible?	Simple-heavy mix; errors under pressure	Range with recovery
Is task fully addressed?	Partial coverage or drift	Clear position; all parts

Why AI misses examiner differentiation

AI often scores surface proxies. Read why AI and examiner scores disagree and examiner mismatch causes.

Train like an examiner

1. Score one criterion per pass

Listen once for FC only; reset; score LR only.

2. Find the floor criterion

Your lowest consistent descriptor level is your true band until it rises.

3. Re-test on fresh prompts

Memorized performance hides differentiation—use Band 6→7 shifts.

Key takeaways

Examiners decide adjacent bands on sustained descriptor evidence.
Weakest criterion often caps the skill score.
Band 6 vs 7 is usually development and control, not vocabulary display.
Train with single-criterion passes on unfamiliar prompts.

FAQ

They match performance to public descriptors; sustained development and flexible control tip to 7; uneven depth or impeding errors tip to 6.

Holistic, but bounded by descriptors—large gaps between trained examiners are rare on the same performance.

Different signals measured—AI proxies vs examiner penalties for templates and shallow development.

Score your weakest criterion first—the one examiners will use to cap you.

Get Band Reality Check →