Interview With Prof. John O'Quigley On Lucy Letby

Professor John O'Quigley, honorary professor of statistics at University College London, presents original published research demonstrating that the shift-statistics chart is statistically inert. Using a proportional hazards model and a beta-binomial distribution fitted to the non-accused nurses' presence data, he shows that the probability of Letby being present at 25 or more suspicious events — given how frequently she worked — is approximately 10%, not the near-100% implied by the chart. He explains that the chart's continuous row of dots for Lucy Letby tells the observer nothing meaningful, because the data it represents is contaminated by selection bias: suspicious events were defined partly by her presence. He reconstructs the statistical picture using only the data from non-accused nurses and shows that when hours worked are controlled for, the apparent anomaly disappears entirely for the hardest-working nurses. He also reads from Rachel Aviv's New Yorker article and Nick Johnson's cross-examination to demonstrate the centrality of the chart to the prosecution's case.