Workplace Bullying Statistics UK: 2026 Facts, Data & Key Insights
Workplace bullying statistics UK
What the evidence says about harmful behaviour at work, its effects on wellbeing and what happens when people raise concerns.
Evidence checked 17 October 2026.
Bullying, harassment and wider workplace conflict are different measures. Their populations, questions and dates stay beside each figure.

reported bullying in the preceding three years
UK employees; research published January 2020. A historical estimate, not a 2026 prevalence figure.reported workplace conflict
CIPD UK working-adult survey, December 2025–February 2026. Broader than bullying alone.reported stress, anxiety and/or depression as an effect
Among 1,943 conflict-experienced respondents in Great Britain; one combined self-reported impact, not a diagnosis.Different questions describe different experiences.
Workplace bullying can involve repeated humiliation, undermining or intimidating behaviour. Research also asks about harassment and the broader experience of conflict. Those questions should not be treated as interchangeable.
A survey’s recall period is another important distinction. A percentage referring to the previous three years is not directly comparable with a percentage referring to the previous 12 months. Differences can also reflect recruitment, question wording and which workers were included.
This guide presents bullying-specific historical evidence alongside newer findings about workplace conflict. It does not turn the newer conflict total into a current bullying-prevalence estimate or join unlike results into a trend.
The specific question matters.
A respondent reporting bullying is answering a narrower question than one about any conflict at work. Different recall periods also change what a percentage means.
Survey and legal definitions
A survey answer describes a reported experience. Legal harassment has its own tests and depends on the facts and the applicable law.
A broader experience
Conflict can include disagreement and difficult relationships as well as harmful behaviour. A conflict total is not a count of bullying cases.
Why an older bullying estimate remains useful
CIPD’s research published in January 2020 asked about bullying and different forms of harassment during the preceding three years. The figures remain useful as dated evidence about those particular experiences.
CIPD’s September 2024 report drew on more than 2,000 employers and more than 5,000 workers. Its 25% figure concerned employees reporting conflict or abuse in the preceding year, a broader outcome.
The 2024 total must not be divided into the 2020 bullying and harassment percentages. The studies have different questions, periods and respondents, and some categories overlap.
Who answered the current CIPD survey?
The 2026 UK Working Lives survey had 5,074 working-adult respondents, collected between December 2025 and February 2026. It used a quota sample weighted to ONS workforce characteristics.
The report sometimes uses the word employees in its narrative, but its figure bases include working adults, including self-employed respondents. The selected results do not print every valid-item base, so the full survey size is not attached to every conditional percentage.
The report says it retained the conflict questions used in 2024, when the module referred to the preceding 12 months. It does not reproduce the exact questionnaire wording in the 2026 chapter. These results remain separate from the older three-year bullying questions.
What the specific historical questions show.
The historical CIPD questions distinguish bullying from sexual harassment and other harassment. The categories can overlap, so their percentages are not parts of a single total.
The January 2020 research reported 15% experiencing bullying, 8% other harassment and 4% sexual harassment over the previous three years. These are reported experiences rather than legally adjudicated cases.
Among people who had been bullied or harassed, 40% said their manager was responsible for their most recent experience. This combines bullying and harassment, and does not mean that managers caused 40% of all current bullying cases.
The 2024 conflict research separately described experiences including humiliation or undermining, shouting and verbal abuse. Harmful behaviour can also happen through messages or online platforms; it is not limited to face-to-face encounters.
The newer 2026 Good Work Index reports a 25% overall conflict share. The selected experiences below apply to all working respondents, including self-employed people, rather than only that quarter who reported conflict. They overlap and are not an exhaustive breakdown.
Bullying and harassment: the historical questions
What did employees report experiencing over the previous three years?
| Group / measure | percent |
|---|---|
| Bullying | 15% |
| Other harassment | 8% |
| Sexual harassment | 4% |
Historical survey, published 21 January 2020; n=2,211 employees. The three categories can overlap and must not be added. These are self-reported experiences, not legal findings. Numeric confidence intervals are not supplied. This is not a 2026 prevalence estimate.
Selected experiences of conflict in the current survey
Which experiences were reported by working adults in the 2026 survey?
| Group / measure | percent |
|---|---|
| Undermined or humiliated | 12% |
| Shouted at or heated argument | 9% |
| Verbal abuse or insult | 8% |
| Discriminatory behaviour | 5% |
These selected, overlapping experiences apply to all working respondents, not only those who reported conflict. Conflict is broader than bullying. The 2026 report retains the earlier past-12-month module but does not reproduce its full wording. Exact selected item bases and numeric confidence intervals are not supplied.
People, incidents and overlapping experiences
A respondent may report more than one kind of behaviour or more than one person involved. A percentage of respondents is not a count of incidents, and several percentages should not be added unless the source defines mutually exclusive categories.
Manager involvement can make a reporting route outside the immediate management line important. The survey does not establish why a particular manager acted as they did or what happened in an individual case.
In the 2020 employer research, 34% of the 160 employer respondents whose organisations reported barriers identified managers’ lack of confidence in challenging inappropriate behaviour as a leading barrier to conflict management. That is an employer-reported barrier, not an explanation established for every instance of bullying.
Reported impacts and research associations.
Bullying and harassment are associated with poorer mental health. Research can identify patterns without establishing the cause of a particular person’s condition.
A 2024 BMC Public Health study analysed England’s 2014 Adult Psychiatric Morbidity Survey. It found an association between workplace bullying or harassment and common mental disorder after adjustment for measured characteristics. The survey is called APMS 2014; its fieldwork ran from May 2014 to September 2015. The later publication date does not make it a current observation.
Its cross-sectional design cannot establish the sequence of every experience or exclude all unmeasured influences. Reported stress, anxiety, sleep difficulties or other distress also should not be treated as a clinical diagnosis.
Newer conflict research adds a different perspective: what people themselves attribute to a conflict they experienced. Those answers describe perceived effects among the respondents asked, rather than a clinical assessment or the effect of bullying alone.
The Acas findings apply to 1,943 adults who had experienced conflict in their main or only job during the previous 12 months. They were aged 16–65 and had worked during that period. Several effects could be reported, so the five bars are not shares of a single total.
Bullying, harassment and mental health
How did past-week common mental disorder differ by past-year bullying or harassment?
| Group / measure | percent |
|---|---|
| Reported past-year bullying or harassment | 29.4% |
| Did not report past-year bullying or harassment | 12.9% |
CMD means common mental disorder assessed by CIS-R, not a prior recorded diagnosis. This historical cross-sectional association does not establish causation. Exact item-valid bases and prevalence confidence intervals are not supplied; odds-ratio intervals must not be used as percentage whiskers.
Reported effects of workplace conflict
Which effects did respondents attribute to conflict in their main or only job?
| Group / measure | percent |
|---|---|
| Stress, anxiety and/or depression | 57% |
| Drop in motivation or commitment | 49% |
| Drop in productivity | 25% |
| Sickness absence | 10% |
| Resigned | 10% |
Base: 1,943 respondents who experienced conflict in the preceding 12 months. Selected impacts overlap. Stress, anxiety and/or depression is one combined self-reported effect, not three diagnoses or a bullying-only estimate. Numeric confidence intervals are not supplied. A January 2026 correction affected a different table.
Reading the historical mental-health comparison
The study included household residents in England aged 16–70 who had done paid work in the past month, including employees and self-employed people. It asked about bullying or harassment during the previous 12 months and assessed common mental disorder in the past week using the CIS-R survey interview.
The chart shows weighted percentages of 29.4% among those reporting bullying or harassment and 12.9% among those not reporting it. These are descriptive group estimates, not the adjusted odds ratios from the research.
There were 3,838 participants in the analytic sample. The table gives group counts rather than separate valid-item denominators, and does not publish prevalence confidence intervals for these two percentages. A correction published on 16 June 2026 changes labels in Table 1; it does not turn the Table 4 comparison into new fieldwork.
Work-related ill health is not a bullying total
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.
These totals cover all reported work-related causes, not bullying alone. They cannot be used to calculate how many days bullying caused or to attribute the whole national total to poor workplace relationships.
Conflict may be associated with absence, turnover and other costs, but individual and organisational outcomes have multiple influences. A survey percentage is not an estimate of a particular employer’s legal liability or productivity loss.
Why stress statistics are not a sector bullying ranking
HSE identifies public administration and defence, health and social work, and education as having above-average rates of work-related stress, depression or anxiety over 2022/23–2024/25.
Those figures measure work-related ill health, not bullying prevalence. Workload, hierarchy, support and reporting arrangements vary within sectors as well as between them. They do not provide a defensible league table of the sectors with the most bullying.
Why the Acas and CIPD totals are kept separate
Acas defines conflict broadly enough to include disagreements, disputes and difficult relationships with colleagues, managers, reports or people outside the organisation. It can be an isolated incident or ongoing.
The 44% prevalence in the Acas survey and the 25% in CIPD’s 2026 report use different sampling and definitions. They are not evidence of a rise or a contradiction. The impact chart uses Acas’s explicit question base of 1,943.
Acas fieldwork ran from 22 August to 21 September 2025. Its January 2026 correction concerned a different table; the displayed impact values were not changed. A September 2026 qualitative follow-up used 30 interviews and is not a replacement prevalence survey.
Reporting is different from resolution.
Whether a concern was raised, how it was raised and whether it was resolved are separate questions. Each percentage needs its own eligible group.
In CIPD’s 2020 research, 53% of those bullied or harassed did not report the most recent incident. The 2024 report separately found that 47% of employees experiencing conflict had simply let it go. These are different questions and should not be treated as a measure of change.
The 2024 research also found that 36% of employees reporting conflict considered it fully resolved. Meanwhile, 81% of employers felt they did enough to prevent bullying and harassment, and 81% felt they did enough to manage it. These were two separate questions, both excluding sexual harassment. Employer perceptions and employee experiences come from separate respondent groups.
In the 2026 report, 51% of people who experienced conflict said they had just let it go. The chart shows other selected responses from that same conditional group. A discussion with a manager or HR is not necessarily a formal grievance, and looking for a new job is not the same as leaving one.
How respondents dealt with conflict
What did people who had experienced workplace conflict say they did?
| Group / measure | percent |
|---|---|
| Just let it go | 51% |
| Manager or HR discussion | 31% |
| Discussion outside work | 26% |
| Discussion with the other person | 22% |
| Looked for a new job | 8% |
Selected responses overlap. The denominator is respondents with conflict, not all workers; its exact item base and numeric confidence intervals are not supplied. A manager/HR discussion is not necessarily a formal grievance. Job-searching does not mean resigning, and letting it go does not establish resolution.
What silence can and cannot tell an employer
Few formal complaints do not by themselves establish that a workplace has little harmful behaviour. People may use informal routes, decide not to raise an issue or feel uncertain about the response.
The January 2020 research found 24% of employees thought bullying and harassment were swept under the carpet. That is a perception from a dated survey, not a direct audit of every employer’s practice.
Reporting mechanisms need to be understandable and accessible, including when a concern involves a line manager. Monitoring how concerns are handled adds information that a complaint count alone cannot provide.
The current resolution results
CIPD’s 2026 report found 32% of respondents with conflict said it was fully resolved, 34% partly resolved and 33% not resolved. The rounded figures sum to 99%; that is not evidence of a missing response category.
The 2024 figure of 36% fully resolved remains dated context. The reports alone do not establish why the percentage differs or show that every unresolved situation will remain unresolved.
Acas uses five resolution categories in a different survey: 27% fully, 23% largely and 19% partly resolved, with 14% mainly not and 17% not at all resolved. The categories cannot simply be mapped onto CIPD’s three-way breakdown.
Prevention needs more than a policy on paper.
Survey definitions and legal tests are different. The legal position also varies across the UK, so the jurisdiction and effective date matter.
In Great Britain, there is no standalone statutory definition of workplace bullying equivalent to the Equality Act’s harassment provisions. Depending on the facts, discrimination, health and safety, employment or other laws may apply. Northern Ireland has a separate equality-law framework.
Under the Equality Act 2010, harassment can include unwanted conduct related to a relevant protected characteristic that violates dignity or creates an intimidating, hostile, degrading, humiliating or offensive environment. Sexual harassment is unwanted conduct of a sexual nature; it does not require a link to a protected characteristic.
The preventative duty and the October 2026 change
Since 26 October 2024, employers in Great Britain have had a duty to take reasonable steps to prevent sexual harassment. If a tribunal upholds a sexual-harassment claim and finds this preventative duty was breached, compensation can increase by up to 25%.
From 30 October 2026, the duty becomes one to take all reasonable steps. That change is still forthcoming on this guide’s evidence-check date of 17 October 2026.
NDAs and the separately scoped FCA rules
The Employment Rights Act 2025 provides for restrictions on non-disclosure agreements concerning workplace harassment and discrimination. Government implementation is planned for 2027, following consultation on details and exceptions. This is not a blanket ban already in force on all workplace-misconduct NDAs.
Existing rules already prevent NDAs from blocking protected whistleblowing and certain permitted disclosures. The specific agreement and legal context matter.
From 1 September 2026, FCA rules and guidance extend the conduct-rule treatment of serious work-related bullying, harassment and violence in non-banking firms within scope. This does not require every regulated firm to report every allegation; the conduct-rule, disciplinary and reporting criteria matter.
Clear expectations and trusted routes
CIPD recommends clear behavioural expectations, routes for raising concerns and support for managers. Processes need to work when the concern involves someone with authority over the employee.
Training within a wider response
Training is one part of prevention and response. The research does not establish it as the single most effective intervention in every workplace. Policy, managerial behaviour and how concerns are handled also matter.
Our Mental Health First Aid courses equip managers and employees with the skills to build psychologically safe workplaces, have difficult conversations early, and respond effectively when concerns are raised.
Reading workplace statistics accurately.
What percentage of UK employees experience bullying?
The figure depends on the source, question and period. CIPD’s January 2020 research reported 15% over the preceding three years. It should remain dated; a newer percentage for broader conflict is not a replacement bullying estimate.
Can bullying and harassment percentages be added together?
Not when the same people can report more than one experience. The categories in the historical CIPD findings overlap. Adding them would double-count some respondents.
Does a mental-health association prove causation?
No. Cross-sectional findings identify associations, and adjusted analyses can account only for measured characteristics. They cannot establish every sequence of events or the cause of an individual’s condition.
Does no complaint mean no problem?
No. Formal complaints are only one source of information. People may raise matters informally, use another route or decide not to report an incident. Reported resolution also needs its own definition.
Which sector has the most workplace bullying?
The work-related stress figures discussed here cannot answer that question. A defensible comparison would need a consistently defined bullying question, comparable populations, dates and adequate samples.
Is every form of workplace bullying legally harassment?
No. Legal harassment has particular tests and the facts matter. Other employment or health-and-safety duties may be relevant even when those harassment tests are not met. Great Britain and Northern Ireland have different frameworks.
Methods and responsible reuse
This is a presentation of published evidence, not a new survey. The downloads retain the exact population, observation period, question wording, reported estimates and published uncertainty where available.
Historical bullying and harassment questions are kept separate from newer broader conflict questions. Conditional percentages apply to the respondents asked that question, not automatically to every worker.
Missing confidence intervals or item-specific bases are not treated as zero uncertainty. Survey reports and mental-health associations are not clinical diagnoses, adjudicated cases or proof of an individual causal pathway.
Workplace evidence, ready to share.
Keep the question with the number.
Download SVG and PNG charts, CSV data and methodology. Keep the population, dates, definitions and limitations with every reused figure.
Mental Health First Aid Course (2026). Workplace Bullying Statistics UK. Evidence checked 17 October 2026. https://www.mentalhealthfirstaidcourse.co.uk/blog/workplace-bullying-statistics
Sources
Primary evidence and methodology. Checked 17 October 2026.
- CIPD 2020 report: survey findings and methods
- CIPD 2024 bullying and harassment report
- CIPD Good Work Index 2026
- Acas 2025 workplace-conflict survey and technical annex
- Acas current workplace-conflict research collection
- HSE work-related stress, depression and anxiety 2024/25
- Acas harassment and the law
- Acas current sexual-harassment prevention duty
- Acas 30October 2026 duty change
- Acas bullying definition and legal context
- Bunce et al., workplace bullying, harassment and mental health, 2024
- Corrected primary article in PubMed Central, Table 4
- BMC correction to Table 1 column headings, June 2026
- Government NDA implementation announcement
- Government NDA consultation, scope and exceptions
- FCA current non-financial misconduct rules
- FCA final guidance announcement
- FCA Handbook COCON 4.3, effective September 2026
Written by Mental Health Experts. This guide was produced by the team at Mental Health First Aid Course. We publish data resources like this one to help employers, HR professionals, and safety managers understand the true scale of workplace bullying and its mental health consequences.