This article applies to England, Wales and Scotland.
In brief: Published Employment Tribunal decisions are a selected and non-random record. Many disputes never generate a published document, while some withdrawals, procedural dismissals and settled cases do. A success rate based on published judgments describes the chosen corpus and denominator, not every dispute or claim presented.
Last updated: 29 July 2026
Published Decisions Are Not All Disputes
A dispute may resolve internally, settle through ACAS, settle after an ET1, be withdrawn, be dismissed procedurally or reach a merits judgment.
Many disputes resolved without a merits decision never appear in the database. However, it is wrong to say settlements and withdrawals never produce published decisions. Some withdrawn proceedings result in a short published judgment dismissing the claim. A settlement can also lead to withdrawal and a procedural dismissal judgment.
A published document therefore does not necessarily mean the merits were decided.
Why Success Rates Disagree
| Denominator | What it may include |
|---|---|
| Cases received | Presented proceedings |
| Disposals | Merits, withdrawals and other concluded proceedings |
| Merits outcomes | Only substantive decisions |
| Published documents | Published judgments or decisions in the corpus |
| Classified outcomes | Items the chosen method could classify |
These answer different questions. Avoid saying a merits-only rate will always be higher unless the actual datasets prove it.
Official Statistics and a Published-Decision Corpus
HMCTS statistics measure claims and cases moving through the system. A private corpus measures collected and classified published documents. They are not necessarily consecutive stages of one linked funnel.
Documents, Cases and Complaints
- Document: one published judgment or decision.
- Case/proceeding: the overall tribunal matter.
- Complaint/head: one legal cause of action.
Calling every complaint a “claim” risks confusion with HMCTS terminology.
Extraction Accuracy and Validation
Any accuracy claim should disclose the field, sample size, sampling method, model or rules version, class prevalence, precision/recall, human review and duplicate handling.
A statement such as “88% to 98% accurate” is not independently assessable without that evidence.
Provisional Figures
Backfills can change both rates and counts through new publications, deduplication, document replacement, case linking and corrected classifications. Every figure should carry a computation date and corpus version.
Four Questions to Ask
- What exactly is counted?
- What is the denominator?
- What period and corpus version are covered?
- What is missing?
Frequently Asked Questions
Do settled or withdrawn cases ever appear?
Yes. Many do not, but some withdrawals generate published dismissal judgments.
Can published decisions show the settlement rate for every dispute?
No.
Are private corpus counts automatically stable?
No.
Sources
- GOV.UK Employment Tribunal decisions — https://www.gov.uk/employment-tribunal-decisions
- MoJ Tribunal Statistics — https://www.gov.uk/government/collections/tribunals-statistics
- ACAS Employment Tribunals — https://www.acas.org.uk/employment-tribunals
About this data
- Source
- Yerty analysis of the published UK employment tribunal record and official HMCTS / MoJ statistics.
- Last updated
- 29 July 2026
- Methodology
- How Yerty builds and verifies the tribunal record — coverage & methodology.
- Limitations
- Figures reflect published decisions and official aggregates for England & Wales. They are not a complete census of every claim, and outcomes vary by claim type, evidence, and facts.
- Not legal advice
- This article is information only and does not constitute legal advice.