Claim: “An Afghan male is 22× more likely to be convicted of rape than someone born in the UK.”
⚠️ Verdict: Misleading / Overstated
Summary: A review of the available data shows a much smaller difference once like-for-like comparisons are made. The “22×” figure is not supported by reliable evidence and significantly overstates risk.
Overview
In August 2025, Nigel Farage repeated a controversial claim on social media and in broadcast interviews. He said that Afghan men in the UK were “22 times more likely to be convicted of rape than those born here.” The statement went viral almost immediately, circulating in memes, short clips, and online discussions about immigration and crime.
The claim was presented without any context or identifiable source. Independent journalists and data analysts soon found that the figure came from a misreading of incomplete information, producing a ratio that was both exaggerated and statistically meaningless.
Where the figure came from
The “22×” statistic appears to have originated from a small blog that attempted to calculate conviction rates by nationality. It used two separate datasets: a partial list of Afghan-born offenders convicted of rape in England and Wales, and census estimates for Afghan-born residents. Dividing one by the other produced an apparent rate that was then compared to that of UK-born men.
This method contained major flaws. It treated very small sample sizes as representative, ignored missing data, and failed to adjust for demographic differences such as age. It also combined conviction data and population data that covered different years, meaning the ratio had no valid statistical basis.
What the data actually show
Ministry of Justice datasets do not provide nationality-level breakdowns for rape convictions. They only record total offences and, in some cases, the proportion of foreign nationals in custody. That broader category includes anyone without UK citizenship, regardless of immigration status or route of entry. It is not the same as “Afghan males who arrived illegally.”
When data analysts revisited the available figures, they found that even under the most generous assumptions, the difference was a small multiple of the UK-born rate—closer to three or four times higher, not twenty-two. Once adjusted for age and socioeconomic background, the difference largely disappeared.
Independent fact-checking
Sky News – Data & Forensics reviewed the claim and concluded that it was based on “a selective reading of incomplete data.” Their investigation explained that conviction data by nationality are too sparse to produce reliable ratios and that the small sample size makes year-to-year changes appear far larger than they really are.
Context and interpretation
The claim circulated during heightened debate over immigration and public safety. By implying that men from Afghanistan posed a vastly greater threat, it contributed to a distorted narrative that equated nationality with criminality. Researchers warned that such claims risk fuelling hostility toward refugee communities and misrepresenting how sexual offences are recorded and prosecuted.
Experts also highlighted that conviction rates reflect complex factors including reporting practices, access to legal support, and differences in population age profiles. None of these are accounted for in the original “22×” calculation.
Why the claim is misleading
- It compares groups of vastly different sizes using inconsistent definitions.
- It relies on very small numbers of cases, which exaggerate differences through statistical noise.
- It ignores demographic and socioeconomic factors that strongly influence conviction patterns.
Because of these problems, the “22×” figure cannot be treated as credible. It exaggerates differences that are not supported by official data and misleads readers about the relationship between nationality and crime.
How to check similar claims
- Verify that a data source is cited and accessible.
- Check whether the comparison controls for group size, age, and time period.
- Be cautious of large round numbers and multipliers that appear without context.