ELTiverse

Search Terms

Search for ELT terms and concepts

Moderation

Assessmentprofessional-developmentAssessment Moderation

Moderation is the quality assurance process of checking, adjusting, and calibrating marking across different raters, markers, sites, or institutions to ensure fairness and consistency. Where rater training prepares examiners before assessment, moderation monitors and corrects after or during the marking process.

The fundamental question moderation answers: Would this candidate receive the same score regardless of who marked their work?

Types of Moderation

Statistical Moderation

Scores from different raters or groups are adjusted statistically to compensate for systematic differences:

  • A rater whose mean score is 0.5 bands below the overall mean is identified as severe; scores are adjusted upward
  • A school whose internal assessment results are consistently higher than external exam results has its scores scaled down

Statistical moderation preserves the rank order of candidates within each group but adjusts levels to a common standard. It requires sufficient data — small sample sizes make statistical adjustment unreliable.

Social Moderation (Consensus Moderation)

Raters meet to review and discuss sample scripts/performances together. They:

  1. Score samples independently
  2. Compare scores and discuss discrepancies
  3. Reach consensus on appropriate scores
  4. Agree on standards for borderline cases

This is the most common form in institutional contexts and closely resembles standardisation meetings used in rater training.

Expert Moderation

A senior examiner or team leader reviews a sample of marked work to verify that the rater has applied the rating scale correctly. If systematic issues are found (e.g., a rater consistently ignoring one criterion), the rater receives feedback and may have their marks adjusted.

Moderation in Practice

ContextModeration approach
IELTS WritingDouble-marking for a proportion of scripts; statistical monitoring of each examiner's patterns
Cambridge examsTeam leader reviews sample of each examiner's marking; statistical post-hoc analysis
University departmental essaysSocial moderation meetings; blind second-marking of borderline scripts
Language school end-of-course testsTeam discussion of anchor scripts; spot-checking across markers

Key Principles

  1. Blind marking first — Raters should score independently before seeing others' judgments; discussion comes after
  2. Use anchor scripts — Pre-agreed exemplar performances at each level provide a shared reference point
  3. Focus on the criteria — Moderation discussions should be grounded in band descriptors and rubric language, not personal preferences
  4. Act on findings — Moderation that identifies problems but does not lead to score adjustment or further training is performative
  5. Document decisions — Record the rationale for any score changes for transparency and accountability

Moderation vs Standardisation

The terms are sometimes used interchangeably, but a useful distinction:

  • Standardisation = establishing shared standards before marking (proactive)
  • Moderation = checking that those standards were applied during and after marking (reactive)

Both are needed. Standardisation without moderation assumes training was perfectly effective. Moderation without standardisation has nothing to moderate against.

Why It Matters

Without moderation:

  • Rater severity/leniency differences directly affect candidates' scores and life outcomes
  • Raters who drift from standards over time are not identified
  • Institutional certificates and grades lose credibility
  • Inter-rater reliability is assumed rather than demonstrated

Key References

  • Weigle, S. C. (2002). Assessing Writing. Cambridge University Press.
  • Lumley, T. (2005). Assessing Second Language Writing: The Rater's Perspective. Peter Lang.
  • Wyatt-Smith, C., Klenowski, V. & Gunn, S. (2010). The centrality of teachers' judgement practice in assessment. Assessment in Education, 17(1), 59–75.

See Also

Related Terms