Why her book matters for Letby
Math on Trial (2013) is the canonical accessible reference on how statistical fallacies produce wrongful convictions. It walks through ten cases including Sally Clark and Lucia de Berk, and catalogues the specific mathematical errors that put innocent people in prison. It is the book a CCRC reviewer needs on the shelf when considering a case like Letby.
Schneps’s framework identifies several recurrent mathematical fallacies: the multiplication fallacy (Meadow’s squared-cot-death probability), the selection- effect fallacy (de Berk, Letby), the prosecutor’s fallacy (confusion of conditional probabilities), and the ecological fallacy (aggregating individual cases in ways that distort population-level reasoning).
Professional background
- Mathematician, Institut de mathématiques de Jussieu, Sorbonne Université.
- Research speciality: Galois theory, number theory, the inverse Galois problem.
- Co-author with her daughter Coralie Colmez of Math on Trial (Basic Books, 2013).
- Long-standing public-engagement work on mathematics in legal contexts, including commentary on UK and continental European miscarriage-of-justice cases.
The Schneps framework applied to Letby
Four specific Schneps fallacies apply to the Letby shift-rota chart:
- Selection-effect fallacy. Events selected as suspicious because Letby was present, then presented as evidence of her guilt.
- Prosecutor’s fallacy. P(shift overlap | innocent) treated as P(guilty | shift overlap).
- Independence fallacy. Cluster events on the same unit in the same period treated as independent events for probability calculation.
- Denominator-suppression fallacy. Numerator (events where Letby was present) shown; denominator (events where she was not, including her non-shift days) not shown.
Each of these is catalogued in Math on Trial. The Letby chart demonstrates all four.
The international-comparative dimension
Schneps’s academic work is rooted in French and American mathematical and legal contexts as well as UK. Her perspective is therefore genuinely international in the sense the Shoo Lee Panel is: not UK-specific pleading, but statistical-community consensus that crosses jurisdictions. That international-statistical-community consensus is itself evidence: a statistical framework that does not hold up across different legal systems is unlikely to hold up within any one of them.