The story the headlines get backwards
Schizophrenia is endlessly tied to violence in the press. The evidence tells a different, better-documented story: most people never offend, the modest extra risk is driven mostly by substance misuse and untreated illness — and people with schizophrenia are far more often the victims of crime than its cause.
People with schizophrenia are at least 14 times more likely to be the victim of a violent crime than to be arrested for committing one.
In a meta-analysis of 15,446 people with psychosis, around 78% had no record of any criminal offending — violent or non-violent.
is attributable to serious mental illness. The absolute risk the public faces is small.
have schizophrenia — and that rate rises and falls in step with the general homicide rate, pointing to shared social causes.
Offending is the exception
The raised risk is real, but it is modest — and almost entirely explained by factors that have nothing to do with the diagnosis itself.
Most people never offend
Share with any criminal offending, violent or non-violent (Yee et al., 2020)
Substance misuse is the real driver
Lifetime violent-crime conviction rate (Fazel et al., 2009, Sweden)
Offending is the exception, not the rule
A 2020 meta-analysis pooling 15,446 people with psychosis found about 22% had ever offended in any way — meaning roughly 78% never had contact with the justice system as offenders at all. Most people with schizophrenia never commit a crime.
Yee et al., Schizophrenia Research (2020).
The elevated risk is real — but modest, and heavily mediated
In a 25-year Australian study of 2,861 people with schizophrenia, 21.6% had a criminal conviction versus 7.8% of matched controls. Convictions for violence rose only modestly over the period (6%→10%), mirroring a parallel rise in the comparison group (1%→3%). Researchers found no evidence that community care itself drove offending.
Wallace, Mullen & Burgess, Am J Psychiatry (2004).
Substance misuse is the single biggest driver
Sweden’s national registry (8,003 patients) found 13.2% of people with schizophrenia had a violent-crime conviction vs 5.3% of controls. But split by substance misuse the picture changes completely: 27.6% with comorbid substance misuse versus just 8.5% without — only marginally above the general population. Take substance misuse out of the equation and most of the excess risk disappears.
Fazel et al., JAMA (2009).
Shared family background explains much of the rest
When the same Swedish patients were compared with their own unaffected siblings — controlling for genes and upbringing — the raised risk shrank sharply (adjusted odds ratio fell from 4.4 against the general population to 1.8 against siblings). Much of the association reflects shared social and familial adversity, not the diagnosis itself.
Fazel et al., JAMA (2009).
Homicide is rare — and tracks society, not diagnosis
Across developed countries about 6% of homicide offenders have schizophrenia, and the annual risk for any one person with schizophrenia is roughly 1 in 10,000. Crucially, this rate rises and falls in step with the overall homicide rate, indicating the same social drivers — poverty, deprivation, substance misuse — are at work.
Large et al., Schizophrenia Research (2009); Wallace et al. (2004).
Untreated early psychosis is the window of risk — and treatment closes it
The highest risk falls before anyone gets help: in untreated first-episode psychosis the homicide rate is about 15.5 times higher than after treatment begins, and roughly 4 in 10 homicides by people with psychosis happen before any treatment. Antipsychotic treatment cuts the risk of violence by around 63% (risk ratio 0.37). The answer to risk is faster, better care — not fear.
Nielssen & Large, Schizophrenia Bulletin (2010); meta-analytic trial data.
The danger runs the other way
By every measure, people with schizophrenia are far more likely to be harmed than to harm. This is the real public-safety story.
0× more likely
to be the victim of a violent crime than to be arrested for committing one. The threat people with schizophrenia face in the community vastly exceeds any threat they pose to it.
Victims, far more than the public
Violent-crime victimisation in the past year (Teplin et al., 2005)
Harmed more across every category
How many times more often, vs the general public, by crime type
The danger runs the other way
A three-year US study established that community-dwelling people with schizophrenia are at least 14 times more likely to be the victims of violent crime than to be arrested for one. The threat they face vastly exceeds any threat they pose.
Brekke et al., Psychiatric Services (2001).
A landmark survey put numbers on it
Using the same national survey instrument as the US government, researchers found that 25.3% of people with severe mental illness had been a victim of violent crime in the past year — 11.8 times the general-population rate of 2.79%. Their annual rate was 158.2 violent incidents per 1,000 people, against 39.9 per 1,000 in the public — a four-fold volume of harm.
Teplin et al., Archives of General Psychiatry (2005), N=936.
Exploitation and theft are routine
Victimisation is overwhelmingly likely across every category — theft, robbery, assault and sexual violence. Financial exploitation is rife: one study found a majority of patients had been financially victimised in the previous month, and many were left without money for food, housing or medication, pushing them toward survival-driven offences.
Teplin et al. (2005); case-linkage victimisation studies.
Often it is a hate crime — and often it is dismissed
In a UK survey of 361 people with severe mental illness, 37% felt a crime against them was motivated by the offender’s attitude to their identity, and 25% considered it a hate crime targeting their mental health. Victims report being disbelieved or dismissed by services more often than the general public — so much harm never even reaches a record.
UK Victim Support, “At Risk, Yet Dismissed”.
The answer to risk is care, not fear
The evidence points the same way every time: risk concentrates around untreated illness and substance misuse, and it falls sharply once people get help. Fast, well-resourced treatment protects everyone — and it protects people with schizophrenia most of all. Sensational crime reporting does the opposite, deepening the stigma that keeps people from seeking care.
How to report this responsibly“Public safety initiatives and media coverage must shift from sensationalising rare acts of violence to addressing the severe victimisation crisis these patients face.”
Every figure on this page was fact-checked against the original peer-reviewed studies — Yee et al. (2020), Wallace, Mullen & Burgess (2004), Fazel et al. (2009), Large et al. (2009), Nielssen & Large (2010), Teplin et al. (2005) and Brekke et al. (2001) — together with the UK Victim Support report “At Risk, Yet Dismissed”. Where a source summary disagreed with the published paper, the published figure was used. These are real published statistics, not illustrations.
Set the record straight.
If the headlines got it backwards, help turn it around.