John Bibby wrote: 18 July 2005 19:19
Subject: Re: Roy Meadow
Having now looked at this URL, I am flabbergasted at how crass Roy Measows'
evidence is. His statistical errors should have been spotted by an average
GCSE student. (Thus of course, one should not be surprised that the lawyers
present failed to spot them. What about Sally Calrk's defence lawyer - why
did he/she not pick it up?)
---------------------------
Quite true. My original mailing (Wed, 10 Nov 1999 16:17:02 on ALLSTAT) was posed as a question because it seemed so blindingly obvious that this was a "travesty of method".
The statistical evidence was dismissed at the first appeal. The judges ruled that it was incorrect but had not unduly influenced the jury, and could have been challenged in the original trial as "lies, damned lies and statistics". I take particular exception to this "institutional prejudice against statistics" but the RSS did not find it appropriate to highlight the issue to the Lord Chancellor.
The conviction was at first upheld, on the grounds that there were numerous other strands of evidence. Reading the ruling, it is clear that each piece of evidence was interpreted on a prior assumption of guilt, an assumption that was subsequently reversed without changing the other evidence. While juries may be stupid and prejudiced, it is alarming that our most learned judges can get away with the same false logic.
Meadow claimed that his conclusions were based on original research. He was an acknowledged leading expert when he gave evidence, but subsequently backtracked and claimed he was merely quoting the Bristol study. Most seriously, to me, he claimed to have destroyed his own research case notes - shades of Cyril Burt.
Regarding the charge for which he has been struck off, I would emphasize the error was not *statistical* but *scientific*. It was - to me as a medical outsider - blindingly obvious that deaths within one family could have common causes that might be genetic, environmental or social, quite apart from the common presence of one person. I find as a statistician that it is far more valuable to concentrate on the exact meaning and provenance of data than to accept the data and try to model any observed irregularity. I am not unique, but the attitude is novel to those (inside or outside statistics) who expect statistical analysis to concentrate on formulae and doing sums.
Meadow was not the only paediatrician who applied prejudice in this case. David Southall (http://www.msbp.com/forum/post-758.html as an immediate link for background) accused Sally Clark's husband, until Southall too was dragged before the GMC.
Another current case that raises issues of interpretation of data is that of Sino Jenkins, currently awaiting a decision on whether to proceed to a third trial. The presence of minute traces of blood on his clothing does not seem to me consistent with the claim that he bludgeoned anyone to death, but the case seems to revolve around the probabilities of this or other explanations. Starting from any prior other than "guilty", it does not seem to me that any case is proven "beyond reasonable doubt."
Allan
***********************************************************************************
This email and any attachments are intended for the named recipient only. Its unauthorised use, distribution, disclosure, storage or copying is not permitted. If you have received it in error, please destroy all copies and notify the sender. In messages of a non-business nature, the views and opinions expressed are the author's own and do not necessarily reflect those of the organisation from which it is sent. All emails may be subject to monitoring.
***********************************************************************************
******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
*******************************************************
|