Nationality is not religion. Crime data must distinguish them.
Nationality and religion are different things. Crime records often describe citizenship or where people’s families came from, not their faith. Those records cannot establish that Islam causes violence.
The examples here come from Sweden, England and Wales, Germany and the US. They do not give one crime rate for Muslims across Europe.
On this page
What a Swedish comparison found
Swedish records for 2015–2018 compared groups by whether they or their parents were born abroad. Differences in recorded suspicion of crime became smaller when researchers allowed for age, sex and economic circumstances, though some differences remained. The records described family background, not religion. A remaining gap does not tell us its cause.
What the UK child-abuse inquiry could establish
The 2025 Casey review found national ethnicity records too incomplete for a reliable national conclusion. In some local child-exploitation investigations, people recorded as Asian were overrepresented. That is an ethnicity finding in those cases, not a national crime rate for Muslims. Abuse and failures to protect victims need to be acknowledged while avoiding blame of a whole faith.
Citizenship is not faith
German crime statistics count suspects and record citizenship. They do not tell us a suspect’s religion. A group’s share of suspects also needs to be compared with its share of the population before treating it as a rate. US FBI crime records include age, sex, race and ethnicity, but not an offender’s faith; religion-related hate-crime labels describe the target or motive.
Prison records answer another question
England and Wales do record prisoners’ self-reported religion. But the number in prison depends on detention and sentence length, as well as crime. It is not a measure of every offense or a rape rate. A change in how religion was recorded in June 2025 also makes comparisons across that date harder.
Check what the records actually measure before drawing a conclusion about religion.
Keep in mind
- Suspects, people convicted and people in prison are different counts.
- Age and economic circumstances explain some differences, not necessarily all of them.
- An offender’s actions do not make other people of the same background guilty.
What does a remaining difference mean?
Allowing for known differences between groups does not reveal every cause of a gap.
A statistical comparison cannot assign an unexplained difference to a religion that was not measured. Countries also use different crime definitions and recording practices. Names, ethnicity, citizenship and family origin cannot be substituted for a person’s faith.
Sources you can check
Publication links open in a new tab, so you can keep your place here.
Registered offending among native/foreign-background residents (opens in a new tab)
Swedish National Council for Crime Prevention, 2021:9
It counts recorded suspicion of crime in 2015–2018, not proven guilt. Groups are defined by migration background, not religion.
Return to reading ↑National audit on group-based child sexual exploitation and abuse (opens in a new tab)
UK government, June 2025
National ethnicity records are incomplete. Particular local investigations cannot establish a national rate for a religious group.
Return to reading ↑Police crime statistics 2025 (opens in a new tab)
German federal government, April 2026
It records citizenship and suspects, not faith. More detail about offenses and population sizes is needed before making a rate comparison.
Return to reading ↑NIBRS technical specification 2025.0 (opens in a new tab)
FBI
A religion-related hate-crime code describes a target or motive; it does not record the offender’s religion.
Return to reading ↑Offender Equalities Report 2023/24 (opens in a new tab)
HMPPS
A prison population on one date is not a measure of every crime. Recording changed in June 2025, making before-and-after comparisons harder.
Return to reading ↑
Detailed research notes and source data
Our longer notes explain how we checked the studies, including disagreements and documents we could not get.
Download the detailed research notes →Something unclear or missing?
You can save a feedback note to share with the site owner. The form saves a file on your device; it does not send it to us.
Save a feedback note →About this draft
These pages have not completed expert review.
Research checked: . Wording reviewed: 2026-10-07. Version: draft-3.
We revisit the page when new evidence or corrections change the answer. A public history of corrections is not available yet.