Why her case matters for the Letby review
The Royal Statistical Society’s published framework on statistical evidence in criminal trials is a direct consequence of what went wrong in Sally Clark’s case. Every argument advanced by independent statisticians — Prof. Richard Gill, Prof. Peter Green, the triedbystats.com team — about the Letby shift-rota chart rests on that post-Clark framework. The framework exists because the same pattern of error the Clark jury accepted is the pattern of error the Letby jury was asked to accept.
The facts of her case
- Solicitor from Cheshire. Her first son died at 11 weeks; her second at 8 weeks.
- Convicted of both murders in November 1999.
- Central statistical argument at trial: the paediatrician Sir Roy Meadow told the jury the probability of two cot deaths in one middle-class family was 1 in 73 million. He calculated this by squaring the probability of a single cot death.
- The calculation was wrong. Squaring a single-event probability treats the events as independent; biological events in the same family with shared genetics and environment are not independent.
- Pathology evidence from Home Office pathologist Dr Alan Williams omitted microbiology findings suggesting the second son had died of a Staphylococcus aureus infection.
- First appeal rejected in October 2000. Second appeal succeeded in January 2003.
- Died in March 2007 of acute alcohol intoxication.
What the Royal Statistical Society said
In October 2001 the RSS publicly corrected Sir Roy Meadow’s statistical calculation. Its subsequent guidance — the canonical UK reference on statistical evidence in criminal trials — crystallised several principles:
- Selection effects must be disclosed and accounted for.
- Independence assumptions must be justified; biological events in the same context are typically not independent.
- Base rates must be given. The prosecutor’s fallacy must be avoided.
- Statistical reasoning should be presented by a qualified statistician, not reverse- engineered by a medical witness.
How her case applies to Letby
The Letby shift-rota chart fails at least two of the RSS principles on its face:
- Selection effect. The 25 events were selected partly because Letby was there; her shift attendance against the selected set generated a visual match by construction.
- Independence. The collapses occurred on a unit with a concurrent outbreak, staffing strain and infrastructure failures. They are not independent trials on a fair coin.
For the full treatment, see our Sally Clark parallel analysis.
Why this biography is on the site
Sally Clark is not connected to the Letby case except by parallel structure. We include her biography because the RSS framework her case produced is load-bearing in the independent-statistician critique of the Letby shift-rota chart. Readers unfamiliar with the Clark case cannot fully weigh Prof. Gill’s or Prof. Green’s critiques of the Letby statistics. This page supplies the reference.
The personal cost of the Clark miscarriage of justice is also a specific, material reminder of what wrongful convictions do to the people subjected to them. Mrs Clark died four years after her 2003 acquittal. Wrongful convictions served under severe sentences have a documented pattern of producing catastrophic effects on the convicted person. That is part of why timely review of contested convictions matters.