Workplace Harassment Statistics UK: 2026 Facts, Data & Key Insights

Workplace harassment statistics UK

Understanding reported experiences, the response to concerns and the legal duties that shape safer working environments.

Evidence checked 17 October 2026.
Survey experiences, legal definitions and sector-specific evidence stay separate. Each figure keeps its population, timeframe and limits.

Three colleagues having a respectful conversation around a table in a welcoming office
CIPD · 2026 survey3%

of women reported unwanted attention of a sexual nature

UK working-adult survey, December 2025–February 2026; the corresponding figure for men was 1%. This is one measured behaviour.
Civil Service · 2025 survey9.0%

reported bullying or harassment in the previous year

Aggregate respondents in participating Civil Service organisations. A combined measure, not harassment alone or a UK workforce rate.
Civil Service · conditional group43.5%

said they reported their experience

Among respondents who experienced bullying or harassment and answered the reporting question. Reporting was not restricted to formal grievances.

What is being measured?

Workplace harassment can affect wellbeing, working relationships and employment. Understanding its scale starts with the question asked: sexual attention, bullying, harassment and wider conflict are not identical measures.

A survey may combine bullying and harassment, ask about a particular behaviour or use a broad definition of conflict. It may cover all working adults, one sector or a selected group of employees. A percentage only describes the population and experience specified in that question.

This guide brings those perspectives together without adding overlapping percentages or joining unlike surveys into a trend. The four charts retain their own dates and eligible groups. The legal discussion distinguishes reported experiences from the tests used to determine unlawful harassment.

Survey experience

The question defines the result.

Unwanted attention, bullying and harassment can be measured separately or together. A respondent’s experience is not a tribunal finding.

Population

Keep the group in view.

An organisation-wide survey describes its participating staff, not every worker in the country. Targeted inclusion surveys have their own eligibility.

Law and timing

Jurisdiction matters.

Great Britain and Northern Ireland have different equality frameworks. Future reforms must not be described as duties already in force.

What the historical three-year figures mean

CIPD research published on 21 January 2020, using a survey of 2,211 UK employees in August–September 2019, found that 4% reported sexual harassment, 8% other harassment and 15% bullying during the preceding three years.

The experiences can overlap. They do not add to a combined 27% prevalence rate. Nor should the 4% be compared directly with a newer question about unwanted sexual attention: the wording, recall period and respondent population differ.

CIPD’s 2024 research separately reported 25% experiencing conflict or abuse in the preceding year. That broader measure is not a replacement estimate of sexual harassment.

A specific question about unwanted sexual attention.

The 2026 Good Work Index reports a specific behaviour: unwanted attention of a sexual nature. The published figures are 3% of women and 1% of men in its working-adult survey.

These percentages apply to the relevant working respondents, not only to people who had already reported conflict. The survey includes self-employed people as well as employees. They are not estimates of every form of sexual harassment or findings that a legal test was met in each case.

The report publishes whole percentages without subgroup confidence intervals. Dividing the rounded values would not provide a precisely estimated relative risk. The chart shows the source’s published precision without adding an uncertainty range that the report does not supply.

United Kingdom · Working adults, including self-employed people · December 2025–February 2026

Unwanted attention of a sexual nature

What did women and men report in the 2026 working-adult survey?

Women3%
Men1%
0%5%

One specific behaviour, not all forms of legally defined harassment. The full survey had 5,074 respondents; exact subgroup item bases and numeric confidence intervals are not supplied. The report retains the earlier past-12-month conflict module without reprinting its full wording. Whole percentages are the published precision; do not derive a precise relative risk from them.

The current CIPD survey and its reference period

The 2026 UK Working Lives survey had 5,074 respondents, with fieldwork from December 2025 to February 2026. It used a quota sample weighted to ONS workforce characteristics. The full sample is not the valid-response denominator for either sex group.

The report retains the conflict questions introduced in the 2024 module, which referred to the previous 12 months, but does not reproduce the complete question wording in its 2026 chapter. That limitation remains in the downloads.

Its overall conflict figure is 25%. Among respondents with conflict, 51% said they had let it go and 32% said it was fully resolved. These are separate conditional questions, not a sequence showing what happened to every person.

Separate Northern Ireland and general harassment measures

In NISRA’s 2025 Work Quality release, 87.5% of employees aged 18 and over in Northern Ireland reported no bullying or harassment at work during the preceding 12 months. The observations cover July 2024–June 2025, and were published on 31 March 2026. The 95% confidence interval is 85.9%–89.1%.

The source’s short label, “never bullied or harassed”, refers to that previous year, not the employee’s lifetime. It combines bullying and harassment and excludes self-employed workers. It should not be ranked against the differently sampled Civil Service or CIPD results.

ONS also publishes a broader harassment measure for people aged 16 and over in private households in England and Wales. Its latest annual tables cover the year ending March 2026, with revised behaviour groupings and an August 2026 correction. These are not workplace-only figures and are not used here as a workplace prevalence estimate.

What Civil Service respondents reported.

The Civil Service People Survey offers a consistent question within one workforce. In 2025, 9.0% of respondents reported experiencing bullying or harassment at work in the previous 12 months.

The chart uses aggregate respondent percentages from the published tables. This differs from the headline organisational benchmark, which takes the median score across organisations. The headline rounded 8% and the aggregate 9.0% answer different statistical questions.

The trend is descriptive. Numeric confidence intervals and exact item-valid bases are not published for these selected results, so the small differences should not be described as statistically significant changes.

A follow-up question asks people who experienced bullying or harassment what it involved. The selected experiences below are conditional on that group and can overlap; they are not percentages of all Civil Service staff.

Participating Civil Service organisations · Aggregate respondents · 2021–2025 surveys

Reported bullying or harassment in the Civil Service

What share of respondents reported either experience in the previous 12 months?

Scale: 0–12.0%Published observations; exact values are in the data table.

The question combines bullying and harassment; this is not harassment alone or a UK workforce estimate. Aggregate respondent percentages differ from the median organisation benchmark. No numeric item confidence intervals or exact item bases are supplied; no significance claim is made. The separate Northern Ireland Civil Service survey is excluded.

Participating Civil Service organisations · Selected follow-up respondents · September–October 2025

Selected experiences within the Civil Service

What did respondents who experienced bullying or harassment say it involved?

Negative micromanagement51.0%
Treated less favourably40.8%
Ignored, excluded or marginalised37.7%
Humiliated in front of others37.6%
0.0%60.0%

Base: respondents reporting bullying or harassment who selected at least one follow-up option. Four selected experiences overlap and do not cover every response. These are not percentages of all staff. Exact item bases and numeric confidence intervals are not supplied. Separate Northern Ireland Civil Service survey excluded.

Who is covered by these Civil Service figures?

The 2025 survey ran from 23 September to 21 October. It received 343,961 responses from 584,718 invited staff across 102 organisations, a response rate of 59%. These whole-survey counts are not the denominators for each chart item.

The separate Northern Ireland Civil Service survey is excluded. The results are a government management survey, not designated Official Statistics, and should not be extrapolated to all UK workers.

The 2025 questionnaire was shorter, but the combined bullying-or-harassment question used for the trend was unchanged. The five displayed annual results come from the same current aggregate table. Reporting behaviour, participation and the composition of respondents still matter when interpreting them.

The experience-type chart covers respondents who reported bullying or harassment and selected at least one option in that follow-up. It presents four selected categories, not an exhaustive distribution. Several could describe the same person’s experience.

Reporting is a separate part of the story.

Among Civil Service respondents who experienced bullying or harassment and answered the reporting question, 43.5% said they reported it, 42.4% did not and 14.2% preferred not to say.

“Preferred not to say” is a separate answer. It should not be added to “did not report”. The question also does not restrict reporting to formal grievances, so its reported share is broader than a count of formal cases.

The three displayed figures total 100.1% because they are rounded. They describe reporting, not whether a concern was upheld, investigated, resolved or followed by a particular employment outcome.

Participating Civil Service organisations · Respondents with bullying or harassment · September–October 2025

Whether respondents reported their experience

Did people who experienced bullying or harassment say they reported it?

Reported43.5%
Did not report42.4%
Preferred not to say14.2%
0.0%60.0%

Base: respondents who experienced bullying or harassment and answered this question. Reporting is not restricted to formal grievances. Preferred not to say is separate from did not report. Displayed categories total 100.1% because of rounding. Exact item bases and numeric confidence intervals are not supplied. Separate Northern Ireland Civil Service survey excluded.

Older reporting and employer-perception findings

In CIPD’s 2020 research, 53% of people who experienced bullying or harassment did not report the most recent incident. This describes reporting to an employer; it does not show that the same people withheld their experiences from the survey. It cannot be used to calculate a hidden prevalence total.

In the 2024 research, 81% of employers thought they did enough to prevent bullying and harassment, and 81% thought they did enough to manage it. These were two distinct questions, both excluding sexual harassment.

The 36% of conflict-experienced employees who considered their conflict fully resolved came from a different respondent group and question. Comparing those answers does not establish individual employers’ motives or match each employee’s experience to an employer’s opinion.

Complaint numbers are useful operational information, but their interpretation depends on the routes available, whether people use them and how cases are recorded. Survey experiences and case records provide different information.

Current duties and forthcoming changes.

Great Britain and Northern Ireland have different equality-law frameworks. The following Equality Act explanation concerns Great Britain and the position checked on 17 October 2026.

Under section 26 of the Equality Act 2010, harassment can involve unwanted conduct related to age, disability, gender reassignment, race, religion or belief, sex or sexual orientation, with the prohibited purpose or effect. This includes violating dignity or creating an intimidating, hostile, degrading, humiliating or offensive environment.

Pregnancy and maternity, and marriage and civil partnership, are protected elsewhere in the Act but are not relevant characteristics for this particular harassment provision. That distinction does not imply that other forms of mistreatment are lawful.

The Act separately covers unwanted conduct of a sexual nature. It also covers less favourable treatment because someone submitted to or rejected sexual harassment, or harassment related to sex or gender reassignment. Bullying need not be related to a protected characteristic; sexual harassment does not require a separate link to one either.

The preventative duty: 2024 and the forthcoming 2026 change

The Worker Protection (Amendment of Equality Act 2010) Act 2023 introduced a proactive duty in Great Britain from 26 October 2024. As at this guide’s review date, employers must take reasonable steps to prevent sexual harassment.

Acas states that the duty strengthens to all reasonable steps on 30 October 2026. This is still a future change on 17 October 2026. Prevention concerns the organisation’s risks and working arrangements, rather than only responding after an incident.

The existing statutory defence to liability for an employee’s discrimination already requires an employer to show it took all reasonable steps to prevent the conduct. That defence is distinct from the preventative duty introduced in 2024.

When a tribunal upholds a sexual-harassment claim and finds the preventative duty was breached, it can increase compensation by up to 25%. This is not an automatic penalty or a standalone tribunal claim for every alleged failure to prevent harassment.

NDAs: an announced reform is not immediate commencement

The Employment Rights Act 2025 provides for restrictions on agreements preventing workers from speaking about relevant harassment or discrimination. Government consulted on details and exceptions in 2026, with implementation planned for 2027.

The July 2025 announcement did not itself create an immediate blanket ban. Existing limits on confidentiality clauses still apply, including protection for qualifying whistleblowing and certain permitted disclosures by victims of crime.

The law, timing and circumstances determine whether a particular clause is enforceable. It would be inaccurate to describe every confidentiality clause as void or the forthcoming restrictions as already applying to every form of workplace misconduct.

FCA conduct rules and creative-industry standards

FCA rules and guidance effective from 1 September 2026 clarify when serious work-related bullying, harassment and violence fall within conduct rules in covered non-banking financial-services firms. The relevant scope, seriousness, disciplinary and reporting conditions still apply.

This is not a requirement for all 37,000 firms to report every allegation. It should not be presented as a universal employment-law rule for organisations outside the framework.

CIISA’s 2025 standards address behaviour across the creative industries, including film, television, theatre and music. These sector standards do not replace employers’ legal duties or create a statutory regulator with the same powers as a court.

The consequences need their own evidence.

Workplace bullying and harassment are associated with poorer mental health. The evidence supports taking experiences seriously while keeping clinical, survey and legal conclusions separate.

A 2024 BMC Public Health study analysed England’s 2014 Adult Psychiatric Morbidity Survey, fielded from May 2014 to September 2015. It found an association between past-year workplace bullying or harassment and common mental disorder assessed for the past week.

The study does not establish that discriminatory harassment is always more harmful than other bullying, that every affected person develops PTSD or that a particular workplace experience caused an individual condition. Its cross-sectional design cannot establish every sequence of events or eliminate all unmeasured influences.

Deloitte’s 2024 UK disability-inclusion findings add a different perspective. In its selected sample, 40% reported microaggressions, harassment or bullying at work in the previous 12 months. That combined measure should not be shortened to harassment alone.

The selected inclusion sample and national ill-health totals

Deloitte’s UK findings used 500 respondents, with fieldwork from January to April 2024, published on 3 December 2024. Eligibility included workers with disabilities, chronic health conditions or neurodivergence, or qualifying difficulty on an adapted disability screen. These are targeted inclusion findings, not prevalence across all UK workers.

HSE estimated 22.1 million working days lost to work-related stress, depression or anxiety in Great Britain in 2024/25. These conditions accounted for 52% of work-related ill-health cases. The figures include many work-related causes and cannot be attributed entirely to harassment.

The latest verified HSE annual publication is dated 20 November 2025; the next annual update is due in November 2026. The older observation period reflects the release cycle, not an estimate of days lost in October 2026.

Prevention in everyday work

Acas guidance connects prevention with the organisation’s risks, clear expectations, reporting routes and appropriate training. A policy’s existence does not by itself establish that it works in practice.

Separate support from conclusions

A person can need support while the facts of a complaint are being established. Survey percentages cannot decide an individual case, replace a fair process or determine a diagnosis.

Our Mental Health First Aid courses help organisations build the management skills and psychological safety needed to address harassment effectively.

Reading the figures with care.

What percentage of UK workers experience harassment?

No single figure covers every definition and population. The 2026 CIPD question measures unwanted sexual attention; the Civil Service question combines bullying and harassment in one workforce. Older three-year figures remain explicitly dated rather than being presented as current annual prevalence.

Why does the Civil Service chart say 9% when the headline says 8%?

The chart uses the aggregate respondent score, 8.994%, rounded to 9.0%. The published headline uses the median organisation score, 7.531%, rounded to 8%. A typical organisation’s score and the aggregate respondent percentage are different statistics.

Does “reported” mean a formal grievance was raised?

Not in the Civil Service question shown here. It asks whether the person reported the experience without limiting the answer to a formal grievance. Nor does reporting establish that the concern was resolved or upheld.

Are bullying and harassment the same legal concept?

No. The Equality Act harassment provisions have specific tests. Bullying does not have to concern a protected characteristic, and other employment or health-and-safety duties may be relevant. A reported experience is not itself a legal finding.

Is the stronger prevention duty already in force?

As checked on 17 October 2026, the current Great Britain duty requires reasonable steps. The all-reasonable-steps preventative duty starts on 30 October 2026. It is separate from the existing all-reasonable-steps defence to liability for an employee’s discrimination.

Does the Northern Ireland figure describe lifetime experience?

No. The NISRA label “never bullied or harassed” means none in the previous 12 months. It covers employees aged 18 and over, not all residents, self-employed people or every UK worker.

Methods and responsible reuse

This guide presents published evidence rather than a new survey. Each chart preserves the measure, population, observation period and source precision. Exact item bases and numeric confidence intervals are left unspecified where the source does not supply them.

The Civil Service charts use aggregate respondent results, not medians of organisation scores. Selected experience categories overlap; reporting categories are mutually exclusive but rounded. The CIPD sex comparison concerns a specific behaviour in a different survey and is not connected to the Civil Service series.

Downloads retain raw estimates, source-table cells, fieldwork dates and limitations. Statistical associations do not establish individual causation, clinical diagnoses or legal liability. Legal information is dated and jurisdiction-specific.

Workplace harassment charts.
Keep the evidence and its limits together.

Download SVG and PNG charts, CSV data and methodology. Retain each measure’s population, dates, definitions and limitations.

Mental Health First Aid Course (2026). Workplace Harassment Statistics UK. Evidence checked 17 October 2026. https://www.mentalhealthfirstaidcourse.co.uk/blog/workplace-harassment-statistics-uk

Sources

Primary evidence and methodology. Checked 17 October 2026.

Written by Mental Health Experts. This guide was produced by the team at Mental Health First Aid Course, to help HR professionals, managers, and safety leads understand the current legal landscape and the mental health consequences of workplace harassment.