A study of the National Crime Agency's behavioural profiling unit finds it usefully illuminates a crime scene in 95% of reports, but adds a genuinely new line of enquiry in only 40%. The method's most notorious UK case shows what happens when its authority outruns its evidence.
The National Crime Agency (NCA) runs the UK's national behavioural profiling capability — formally called Behavioural Investigative Advice (BIA) — through three full-time advisers, two part-time forensic clinical psychologists and one geographical profiler.1 A 2024 peer-reviewed study, co-authored with the NCA's own lead adviser, examined 77 full BIA reports written between 2016 and 2020 across 209 completed cases, mostly murder and serious sexual offences.1
The reports met professional standards — explicit terms of reference, stated methods, peer review — at a very high rate, and 98.6% of the 1,304 individual behavioural claims made across them were grounded in stated evidence.1 But when the two researchers independently rated each report's practical value, a clear split appeared between what profiling is good at and what it isn't.1
The unit's caution has a specific origin. After Rachel Nickell was murdered on Wimbledon Common in 1992, the Metropolitan Police commissioned a psychological profile from Paul Britton, then Britain's best-known profiler, who concluded the killer had rare and specific deviant sexual fantasies.1 Acting partly on that profile, police ran a seven-month undercover "honey trap" operation to draw incriminating fantasies out of Colin Stagg, the prime suspect — who was innocent.1 The trial judge threw out the evidence as manipulative, the case collapsed, and Stagg had already spent over a year in custody.2 He was later awarded £706,000 in compensation; the actual killer, Robert Napper, was only linked to the murder by DNA in 2004 and convicted of manslaughter in 2008.4 The episode, and the Home Office review it triggered, is what pushed UK profiling from freelance consultancy toward the regulated, NCA-run advisory service that exists today.1
No study has checked BIA predictions against the offender eventually convicted, so the accuracy of the underlying claims — as opposed to their perceived usefulness — remains unmeasured.1 The NCA has had no formal external audit of BIA reports since oversight moved from the police to the agency in 2017, and the 77 reports reviewed here were a purposive, not random, sample chosen with the lead adviser's input.1 A separate, more statistical technique — linking multiple crimes to a single unknown offender by behavioural similarity — has a considerably stronger evidence base than predicting an offender's personal traits, and the two are often conflated in public discussion.3
Academics who study the field disagree sharply on how to read the evidence — not on the NCA data above, which both camps accept, but on the underlying discipline of predicting an offender's personal characteristics from crime-scene behaviour.
UK profiling has become a genuinely evidence-based, professionally governed practice: reports overwhelmingly meet formal standards, are peer-reviewed, and nearly all claims are grounded in stated reasoning rather than assertion. Proponents argue the "pragmatic practitioner" model — trained advisers combining behavioural science with investigative experience — delivers real value even without a laboratory-style accuracy score.1
A influential 2008 critique argues that predicting an offender's traits from a crime scene lacks sound theoretical grounding and has never been shown, in controlled tests, to outperform experienced detectives by a reliable margin — and that belief in profiling persists mainly through vivid anecdotes and media portrayal rather than validated results.2