Highest ever number of referrals to anti-extremism scheme Prevent recorded
8,778 individuals have been referred to Prevent, the government's anti-extremism scheme, in the year to March 2025.
It's the highest number ever recorded since records began in 2015, when the statutory Prevent duty was placed on authorities.
It is also a 26% increase from the 6,922 referrals the year before.
Where are referrals coming from?
The education sector made the highest share of referrals as it has done in all reported data (3,129 -36%).
But this represented a small decrease in proportion from 40% of referrals the year before.
The second-highest share of referrals was from police forces (2,631; 30%), which increased by 37% compared with the previous year (1,921; 28%).
Referrals from local authorities had the largest percentage increase (54% from 582 to 895).
By contrast, referrals from the community and from friends and family were the only groups to show decreases, falling by 25% and 7% respectively.
New data on ethnicity
For the first time, data about those referred to the Prevent scheme now includes details about their ethnicity.
For those referrals where ethnicity was specified, 65% (2,747) of referrals were recorded as White, 19% (798) Asian, 8% (320) Black and 8% (338) other.
More people referred for right wing extremism than for Islamist extremism
'Extreme Right-Wing' concerns accounted for 21% (1,798) of referrals, higher than those related to 'Islamist Extremism' (10%; 870).
Compared with the previous year, the proportion of referrals for 'Extreme Right-Wing' concerns increased (up from 19%; 1,314 of 6,921), while the proportion for 'Islamist Extremism' decreased (down from 13%; 913 of 6,921).
One third of referrals have mental health conditions
The data also shows one third of Prevent referrals had at least one mental health or neurodiversity (2,995 out of 8,778).
Autistic Spectrum Disorder (ASD) was the most common condition. But two thirds of all referrals had no mental health information recorded - either because no concern was identified, or data wasn't captured.