Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
The Story of Sally Clark Group 2 - Jing Kai (Presenter), Nabilah, Soon Guan Reading Sally Clark (5 February 2017). Retrieved 21 March 2017, from http://en.wikipedia.org/wiki/Sally_Clark The statement by Professor Dawidon: SALLY CLARK APPEAL. Retrieved from http://www.statslab.cam.ac.uk/~apd/SallyClark_report.doc Overview Background Information Conviction for Murder Analysis of Statistical Evidence Aftermath Conclusion Background Information • Born an only child in Wiltshire, South West England) • Married solicitor Steve Clark (1990) • Moved to Cheshire, North West England in order to practice law with her husband (1994) Conviction for Murder • Two separate incidences where Sally Clark’s baby passed away due to unknown causes (Sudden Infant Death Syndrome, SIDS) • Sally Clark was found alone at home on both occasions • No evidence to prove or disprove her involvement • Sally Clark was convicted and sentenced to life imprisonment on 9th November, 1999 The Antagonists Dr Alan Williams • Withheld microbiology test results for second baby • Tests showed the bacterium Staphylococcus aureus in multiple sites of the baby → Baby could have died due to health concerns! The Antagonists Professor Sir Roy Meadow • Bayes analysis: 1 in 73 million chance (1/8543) x (1/8543) of two children from an affluent family suffering cot death • Conclusion that since this probability is very low, the probability of murder is very high Analysis of Statistical Evidence 1) Assumption of independence → Flaws in calculation: • Lurking variables: Gender, genetic or environmental factors • Chances of a second death increases by 5-10 fold after the first • Uncertainty with probability of double cot death most likely understated Analysis of Statistical Evidence 2) Relevance of data: • Fallacious logic • Probability of double cot death (Prosecutor’s fallacy) • Double cot murder is a less likely phenomenon • Relative likelihood should be used Probably of innocence! Relative likelihood P (2 SIDS deaths in a family) = 1 in 5 million P (2 murdered babies in a family) = 1 in 15 million P (2 SIDS deaths in a family) : P (2 murdered babies in a family) = 3 : 1 Aftermath • Unearthing of microbiological tests by Sally Clark’s husband • Along with the case for flawed statistical evidence, Sally Clark was acquitted in January 2003 • Review of hundreds other cases → 3 similar cases overturned Aftermath • Roy Meadow was struck off the medical register in 2005 for serious professional misconduct but subsequently reinstated after appeal • Dr Alan Williams was banned from Home Office pathology work and coroners' cases for three years • Sally Clark diagnosed with psychiatric problems associated with posttraumatic disorder and alcohol dependency syndrome → Passed away on 16th March 2007 due to acute alcohol intoxication "Throughout my review, I was horrified by the shoddy fashion in which these cases were evaluated. It was clear that sound medical principles were abandoned in favour of over-simplification, over-interpretation, exclusion of relevant data and, in several instances, the imagining of non-existent findings." Dr Sam Gulino, forensic pathologist for the State of Florida, USA. Conclusion • Logical and accurate use of statistical information with explanations required • Data needs to be precise and contextspecific • Failure to make appropriate inferences from statistics can result in fatal consequences Thank You!