Are refugees more likely to commit crime?

Episode Summary

Executive Summary: This episode debunks three viral or disputed claims: asylum seekers in Dorset allegedly causing 44% of sex offences, a YouGov "quiet revival" church-going surge, and a claim that blue-tit chicks need 20,000 caterpillars each. It also corrects a prior More or Less error about Hinkley Point fish-protection costs. Across all stories, the show emphasizes careful sourcing, limits of available data, and how AI and survey fraud can distort public claims.

Main Topics: Dorset asylum seeker crime claim (Priority: 5/5): The programme investigates a social-media claim that asylum seekers in Dorset were responsible for 44% of alleged sex offences. It finds the framing misleading and unsupported by the cited evidence. Limits of crime statistics by immigration status (Priority: 5/5): A discussion with Madeleine Sumption explains that UK crime data rarely records asylum-seeker status, so nationality is only an imperfect proxy and can’t directly answer claims about asylum seekers. YouGov's retracted 'quiet revival' poll (Priority: 5/5): The episode revisits a poll claiming a big rise in church attendance, especially among young men, and explains that YouGov later retracted it after identifying fraudulent respondents. AI, bots, and polling fraud (Priority: 4/5): The show explores how non-probability online surveys can be contaminated by bogus respondents and how agentic AI may make detection harder by mimicking human behavior. Correction to Hinkley Point fish-cost claim (Priority: 4/5): More or Less admits a prior error: the "quarter of a million pounds per fish" line was misleading because it should have referred to endangered fish saved over 25 years and involved disputed cost assumptions. Fact-checking David Attenborough’s blue-tit caterpillar claim (Priority: 4/5): The team verifies a BBC documentary claim and concludes the figure was mis-stated per chick; about 20,000 caterpillars is closer to a nest total, not each chick.

Key Arguments: The Dorset claim that asylum seekers make up 44% of alleged sex offences is unsupported; the cited police data instead showed eight asylum-hotel residents investigated for sex crimes versus 808 sex-crime reports in Bournemouth over the relevant period. Police do not routinely record crime data by immigration status, so nationality-based statistics cannot cleanly prove claims about asylum seekers. Foreign nationals accounted for 7% of sexual offence crimes in Dorset in 2024-25; adding dual British nationals raised this to 8%, but those figures are not equivalent to asylum-seeker rates. The YouGov church-attendance survey was retracted because fraudulent respondents were found in the sample, undermining the headline claim of a quiet revival. Non-probability surveys are especially vulnerable to bogus responses, and AI agents can increasingly imitate human survey-taking behavior, including timing, text, typos, and mouse movements. The Hinkley fish-protection cost claim was oversimplified because the £700 million figure bundles multiple measures and uncertain operational costs, while the fish-saved estimate was only for certain endangered species over an assumed 25-year period. The blue-tit caterpillar claim was likely a per-nest figure rather than per chick: scientific estimates suggested roughly 1,800 caterpillars per chick, around 18,000 for a 10-chick nest.

Data Points: Asylum seekers in Dorset population share: 0.8% - Figure used in the viral post, specifically referring to asylum seekers living in hotels in Dorset Claimed share of alleged sex offences: 44% - The social-media claim being investigated Sex offences investigated among asylum-hotel residents: 8 - Dorset Police update for 1 Jan 2025 to 31 Mar 2026 Sex crimes reported in Bournemouth: 808 - Reported in the same period as the updated investigation figure Sex offences previously charged among migrant-hotel residents: 1 - Last year’s Dorset police FOI response, most recent data at the time of the post Sexual offence crimes by foreign nationals: 7% - Dorset Police data for 2024-2025 Sexual offence crimes by foreign nationals plus dual British nationals: 8% - Expanded figure from Dorset Police data Males awaiting prosecution cited by Grok: 116 - A figure Grok used in its answer, though it could not be found as a standalone official statistic Males from migrant hotels among those awaiting prosecution: 51 - Grok cited this as 44% of the 116 males Fish protection measure cost: £700 million - Fingleton Review case study on Hinkley Point fish-protection measures Acoustic fish deterrent cost: £50 million - Part of the fish-protection package Fish escape pipe system cost: £150 million - Part of the fish-protection package Low-velocity intake heads cost: £500 million - Part of the fish-protection package; disputed as to how much is truly fish-protection cost Endangered sea trout estimate: 0.028 per year - Example estimate from the Fingleton Review Endangered river lamprey estimate: 6 per year - Example estimate from the Fingleton Review Endangered fish category estimate: 18 per year - Another example estimate from the Fingleton Review transcription Assumed lifespan used in calculation: 25 years - Used in the original More or Less estimate for fish-saved cost Blue-tit caterpillars per chick (corrected estimate): about 1,800 - Derived by Dr. Malcolm Burgess from feeding rates and nestling period Blue-tit nestlings’ time in nest: 16 to 20 days - Used to estimate caterpillar demand per chick Blue-tit chicks per nest: about 10 - Used to convert per-chick caterpillar estimate into a nest total Caterpillars per nest: about 18,000 - Approximate total for a typical blue-tit nest, closer to Attenborough’s quoted 20,000

Pivotal Quotes: "The stat comes from a Dorset Police Freedom of Information FOI disclosure... The 44% figure, as originally presented, is misleading and inaccurate in that framing." — Grok: Grok’s on-air correction of its earlier answer about Dorset sexual offences "The quiet revival wasn't plausible back in my first report. I'm glad to be right... it turns out that it was indeed a matter of bogus responses on the survey." — David Vos: Reaction to YouGov retracting its church-attendance survey "They should have said 20,000 caterpillars per nest rather than per chick." — Tim Harford: Conclusion of the blue-tit fact-check after consulting RSPB scientist

Implications: The episode shows how misleading claims can spread when statistics are stripped of context, and how fragile online polling has become in the age of bots and AI. For listeners, the lesson is to demand sources, definitions, and cross-checks before trusting headline numbers.