This article applies to England and Wales.
What evidence tribunal judgments describe, 2017 to 2025
In brief: Across thirty-six named types of evidence, substantial employment tribunal judgments mention payslips, fit notes, screenshots and WhatsApp more than they did, and formal warnings, diaries, photographs and medical reports less. The contract of employment is flat at half of all judgments. The medical shift is confined to disability claims; the fall in warnings is not confined to anything.
Last updated: September 2026
By Yerty | Analysis of published Employment Tribunal decisions, 2017 to 2025.
Ask what evidence a tribunal judgment talks about and the answer has changed in nine years, but not simply towards the phone. Pay paperwork and fit notes are up, formal warnings and medical reports are down, and the contract of employment has not moved.
This analysis takes thirty-six named types of evidence and counts how often each appears in substantial merits judgments in 2017–19 against 2023–25. It reports them one at a time, because each figure is one exact phrase and nothing is grouped. The companion piece on email, WhatsApp and letters covers the correspondence channels; this one covers everything else a party puts in front of the tribunal. The tribunal's own process documents (bundle, list of issues, case management orders) are not evidence and are left out.
Summary
- The digital rise is WhatsApp. WhatsApp appears in 16% of substantial judgments, up from 3.5%; screenshots in 3%, up from 1%. Text messages (6% to 7%), voicemail and social media barely moved.
- Formal warnings are described less, in every kind of claim. Written warnings fall from 17% to 13% of judgments and the fall appears in all ten claim types with enough cases. The contract of employment is flat at half of all judgments.
- Pay paperwork is up. Payslips 7% to 11%, timesheets 3% to 5%, bank statements 2% to 3%; they concentrate in the wages and working-time claims, which grew.
- Medical evidence changed form, not volume, and only in disability claims. Within disability discrimination judgments medical reports fall from 19% to 7% while fit notes rise from 15% to 23%. Outside disability claims the medical rows are flat.
- What the phone displaced: diaries (7% to 5%) and photographs (8% to 5%).
All figures are England and Wales, single-claimant merits judgments of 20,000 to 60,000 characters, one judgment per case: 2,029 in 2017–19 and 1,738 in 2023–25. A judgment counts if it mentions the evidence at least once, which is not the same as relying on it.
Evidence by claim type
One exact phrase from each evidence group, by claim type. Each cell is the share of that claim type's judgments in 2023–25 that mention the phrase, with the change since 2017–19 in points. Claim types overlap (a judgment with an unfair dismissal claim and a wages claim sits in both columns), so columns do not sum to the population. Claim types with fewer than 85 judgments in either period are omitted.
Share of each claim type's judgments mentioning the phrase, 2023–25. Change since 2017–19 on hover.
| Evidence | All cases n 1,738 | Unfair dismissal n 1,085 | Breach of contract n 534 | Unlawful deduction n 561 | Disability discrimination n 375 | Working time n 354 | Race discrimination n 253 | Sex discrimination n 148 | Whistleblowing n 161 | Redundancy n 103 | Age discrimination n 92 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Witness statement | 67% | 67% | 69% | 68% | 66% | 68% | 72% | 69% | 75% | 62% | 66% |
| Contract of employment | 50% | 52% | 72% | 71% | 32% | 71% | 34% | 42% | 44% | 68% | 29% |
| Written warning | 13% | 17% | 12% | 7% | 10% | 7% | 15% | 8% | 9% | 4% | 11% |
| Payslips | 11% | 11% | 18% | 22% | 6% | 27% | 10% | 10% | 5% | 19% | 10% |
| Occupational health | 17% | 17% | 11% | 12% | 53% | 10% | 13% | 10% | 16% | 7% | 16% |
| Diary | 5% | 6% | 5% | 4% | 3% | 6% | 6% | 6% | 5% | 8% | 4% |
| 16% | 17% | 17% | 25% | 12% | 23% | 16% | 19% | 19% | 9% | 16% |
One exact phrase per evidence group. Claim types overlap, so columns do not sum to the population; claim types with fewer than 85 judgments in either period are omitted. Single-hue scale, 0–80%.
The contract of employment and payslips are money-claim documents. The contract appears in 71–72% of breach of contract, unlawful deduction and working-time judgments and in 29–34% of disability, race and age discrimination judgments; payslips 22–27% against 5–10%.
Medical evidence belongs almost entirely to disability discrimination: occupational health is mentioned in 53% of those judgments and in 7–17% of every other column. The rest of the medical rows (fit note 23% against 4–10%, impact statement 19% against 0–4%) follow the same line.
Two movements are in every column. WhatsApp rises by 6 to 20 points in every claim type, most in the wages and working-time claims where shifts and pay are arranged by message. Written warnings fall by 2 to 7 points in every claim type. Neither is a claim-mix effect. The medical shift is the opposite: inside the disability column medical reports fall 12 points and fit notes rise 8; outside it the medical rows are flat.
The full grid, thirty-six evidence types by claim type, is broken out claim by claim in the series below, each with a guide to the evidence that claim turns on.
Evidence by claim type: unfair dismissal, disability discrimination, wages and working time, race, sex and age discrimination, whistleblowing, redundancy.
Witness evidence
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| Witness statement | 59.2% | 66.7% | +7.5 |
Two judgments in three now refer to a witness statement, up from three in five, and the rise is in every claim type (+4 to +12 points). It is the one row here that is as much about the tribunal's process as the parties' evidence; it is kept because a witness statement is evidence.
Employer documents
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| Contract of employment | 49.3% | 49.9% | +0.6 |
| Written warning | 17.3% | 13.1% | −4.2 |
| Final written warning | 12.7% | 9.7% | −3.1 |
| Dismissal letter | 12.6% | 11.2% | −1.4 |
| Job description | 12.3% | 11.1% | −1.2 |
| Disciplinary policy | 12.2% | 11.1% | −1.1 |
| Investigation report | 10.7% | 10.1% | −0.6 |
| Resignation letter | 8.8% | 7.0% | −1.8 |
| Appeal letter | 6.0% | 5.8% | −0.3 |
| Employee handbook | 4.2% | 4.5% | +0.3 |
| Grievance letter | 4.5% | 4.0% | −0.5 |
| Staff handbook | 3.5% | 2.9% | −0.6 |
| Personnel file | 2.8% | 1.6% | −1.2 |
Half of all substantial judgments mention the contract of employment, in both periods. Everything else the employer produces is flat or slightly down. The clearest movement is in formal warnings: written warnings fall from 17.3% to 13.1% of judgments, final written warnings from 12.7% to 9.7%. Judgments describe the formal disciplinary ladder less often than they did.
Two readings fit: fewer cases turn on a prior warning, or judgments summarise the disciplinary history more briefly. The claim-type split below rules out the simplest version of the first.
Pay and time paperwork
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| Payslips | 7.3% | 11.3% | +4.0 |
| Rota | 8.4% | 10.1% | +1.7 |
| P45 | 6.4% | 5.2% | −1.2 |
| Timesheets | 3.3% | 4.6% | +1.2 |
| Bank statements | 1.9% | 3.5% | +1.5 |
Pay and time paperwork is the one class of employer-side document that rises. Payslips go from one judgment in fourteen to one in nine; bank statements nearly double from a low base; timesheets and rotas both rise. They are the evidence the wages and working-time claims turn on, and those claims grew as a share of the population.
Medical evidence
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| Occupational health | 17.8% | 17.3% | −0.5 |
| Occupational health report | 6.5% | 6.5% | 0.0 |
| Fit note | 6.7% | 8.9% | +2.2 |
| Medical records | 3.4% | 4.4% | +1.0 |
| Impact statement | 2.3% | 4.1% | +1.8 |
| Medical report | 4.2% | 2.3% | −1.9 |
| Prescription | 1.9% | 1.6% | −0.4 |
Medical evidence as a whole did not rise. It changed shape. Fit notes rose by a third while medical reports nearly halved; occupational health, whether the report or the referral, is flat. The disability impact statement, a document that barely existed in 2017–19, is now in one judgment in twenty-five.
The pattern is towards routine, employer-held medical paperwork (fit notes, OH reports) and the claimant's own account (impact statements), and away from commissioned medical reports. Whether that reflects cost, a growing willingness to decide disability status without expert evidence, or drafting, is not something the counts can say.
Notes made at the time
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| Diary | 6.8% | 5.1% | −1.7 |
| Minutes of the meeting | 4.9% | 3.7% | −1.2 |
| Contemporaneous notes | 2.1% | 2.4% | +0.2 |
| Incident report | 2.2% | 1.6% | −0.6 |
| File note | 1.0% | 0.8% | −0.2 |
Diaries fall. So do photographs (7.6% to 5.1%) in the group below. These are the two evidence types most obviously displaced by a phone: the running diary by the message thread, the photograph by the screenshot. Screenshots rise from 0.7% to 3.3% over the same period. The word "timeline", which is what a diary becomes when it is typed up for a hearing, rises from 2.8% to 5.0%.
Messages, images and other digital material
| Document | 2017–19 | 2023–25 | Change |
|---|---|---|---|
| 3.5% | 15.9% | +12.5 | |
| Photographs | 7.6% | 5.1% | −2.5 |
| Text messages | 6.2% | 7.3% | +1.1 |
| Social media | 3.7% | 5.2% | +1.6 |
| Screenshots | 0.7% | 3.3% | +2.6 |
| Microsoft Teams | 0.0% | 2.1% | +2.1 |
| Voicemail | 2.1% | 1.7% | −0.3 |
| Phone records | 1.0% | 1.1% | +0.1 |
| Call logs | 0.2% | 0.7% | +0.5 |
WhatsApp is now in one substantial judgment in six. Microsoft Teams appears for the first time in the later period. Voicemail is flat to down and text messages barely move: the growth in messaging evidence is almost entirely WhatsApp. Photographs fall while screenshots rise, which is the same object arriving by a different route.
Covert recordings are not reported here for the reason given in the companion piece: the phrase list has not been hand-checked.
All evidence types, all cases · share of judgments, ■ 2017–19 (n 2,029) → ■ 2023–25 (n 1,738)
Sorted by change. Rows under 3% in both periods are omitted; the full figures are in the table below. A mention is not reliance.
Method and limits
Population and matching. Single-claimant merits decisions with written reasons, England and Wales, one judgment per case, 20,000 to 60,000 characters (n = 2,029 in 2017–19, 1,738 in 2023–25). Each figure is the share of judgments containing the exact phrase at least once, matched whitespace-tolerant and anchored at both ends, so "written warning" excludes "written warnings". Where tribunals ordinarily write a different phrase for the same document, the phrase they write is the one measured (fit note singular; impact statement without "disability"; occupational health with or without "report"; WhatsApp as the bare word).
A mention is not reliance. No negation handling is applied; "there was no investigation report" counts the same as "the investigation report showed".
Exact phrases only. A combined count of every wording for a document would be higher than any single row.
No category totals. Sums across document types are not given because the term groups behind them were found to contain words that do not mean what the group is named for. Those will follow once the groups are rebuilt.
England and Wales only. Scottish judgments use different drafting conventions: "witness statement" appears in 8% of Scottish judgments against 67% in England and Wales, because evidence in Scotland is led orally. A Great Britain figure would blend the two.
Population. Substantial single-claimant merits judgments only; the length band retains a falling share of the population over the period (49% in 2017, 30% in 2025), which the companion piece discusses.
Frequently asked questions
What evidence is most often mentioned in employment tribunal judgments?
In substantial merits judgments from 2023–25: witness statements (66.7%), the contract of employment (49.9%), occupational health (17.3%), WhatsApp (15.9%), written warnings (13.1%) and payslips (11.3%). All figures are England and Wales, single-claimant, one judgment per case, and count judgments that mention the evidence at least once.
Is WhatsApp evidence more common in some claims than others?
Yes. In 2023–25 WhatsApp appears in 25% of unlawful-deduction judgments and 23% of working-time judgments, where shifts and pay are arranged by message, against 12% of disability discrimination and 9% of redundancy judgments. In 2017–19 no claim type was above 7%.
Are tribunals seeing more medical evidence?
Not more; different. Fit notes rose from 6.7% to 8.9% of judgments and impact statements from 2.3% to 4.1%, while medical reports nearly halved (4.2% to 2.3%) and occupational health was flat at around 17%. Within disability claims, medical reports fell from 19% to 7% while fit notes rose from 15% to 23%.
Which employer documents appear most?
The contract of employment, in half of substantial judgments in both periods. Written warnings (13.1%), dismissal letters (11.2%), job descriptions (11.1%) and disciplinary policies (11.1%) follow, all slightly lower than in 2017–19. Formal warnings show the clearest fall: written warnings from 17.3%, final written warnings from 12.7% to 9.7%.
Are payslips used more as evidence?
Payslips are mentioned in 11.3% of substantial judgments, up from 7.3%; bank statements 3.5%, up from 1.9%; timesheets 4.6%, up from 3.3%. They are the one class of employer-side document that rises, and they are the evidence that wages and working-time claims, a growing share of the population, turn on.
Have diaries and photographs declined?
Yes. Diaries fall from 6.8% to 5.1% and photographs from 7.6% to 5.1%, while screenshots rise from 0.7% to 3.3% and "timeline" from 2.8% to 5.0%. These are the evidence types most directly displaced by a phone: the running diary by the message thread, the photograph by the screenshot.
Does a document being mentioned mean the tribunal relied on it?
No. The counts include judgments that rejected a document, gave it little weight, or noted that it was missing, because no negation handling is applied. They describe what judgments discuss, not what decided the case. Reliance would need a hand-coded sample, which this analysis has not done.
Analyse the data
The grid is one exact phrase per group. Query any of the thirty-six terms by claim type, year or industry.
Intelligence Hub →Manage your evidence
Different claims turn on different documents. Organise yours by what your claim has to prove.
Case Hub →Sources
Analysis of published Employment Tribunal decisions, 2017 to 2025, drawn from the public register on GOV.UK under the Open Government Licence v3.0. Population: single-claimant merits decisions with written reasons, England and Wales, one judgment per case, 20,000 to 60,000 characters; 2,029 judgments in 2017–19 and 1,738 in 2023–25.
Related reading: what evidence do employment tribunals look at, tribunal evidence is moving from paper to screen, can WhatsApp messages, emails and recordings be used as evidence at an employment tribunal, and employment tribunal claims: what you need to know.