Probabilities and Possibilities

Yesterday, we were talking about statistical probabilities of two infants, in the same family, dying ‘accidentally’ (aka, from ‘Sudden Infant Death Syndrome). In the Sally Clark case, Dr. (?) Meadows made a serious statistical error by multiplying the statistical improbability of such an event occurring, so that the improbability was in the range of 1:30,000,000. (The detailed numbers are unimportant – it is the enormity which is important)

From previous investigations, I know that, in 2010, the number of SIDS deaths in the UK was 130 out of a total of live births of around 700,000. That gives a ratio of some 1:5400. Because Sally Clark had two such deaths in quick succession, Meadows multiplied the improbability. He said 1/5400 times 1/5400 equals 1/30,000,000-ish. Only on appeal was this use of statistics queried.

It isn’t really the multiplication which is the problem, even though it is wrong to do so in any individual case. What is wrong is to use statistics AT ALL. The statistics prove only the rarity of the event of a SIDS death. They have no relevance to whether or not the death of an infant was accidental (SIDS) or murder. In the Sally Clark case, this fact was not recognised. Meadows was allowed to say that the rarity itself was reason for suspicion. I do not disagree with that, but there is a world of difference between ‘reason for suspicion’ and ‘proof of guilt’. I would say that the probability of 1:5400 is big enough to arouse suspicion, and that the multiplication of the odds to 1:30,000,000 adds very little to suspicion of foul play. In other words, despite the math, the suspicion of foul play in the second incident was no greater than the suspicion in the first incident. The improbability of the event(s) was irrelevant..

====

The above has led me to think, once again, about the McTear Case (see sidebar). My thinking has also been affected by discussions at Leg Iron’s place (see sidebar) about atheists. I opined there that, since God does not do “the 9 pm Heavenly News”, as some sort of hologram which appears in the sky, then it is a matter of faith that the deity exists. But, at the same time, the simple fact that God does not do ‘the 9 o’clock heavenly news’ does not mean that God does not exist. Thus, Atheism is just as much a matter of faith as is Theism. Other than being a believer in God, there is only one other possibility for anyone who can think, and that is to be ‘agnostic’.

====

The above considerations led me to think once again about Doll’s ‘Doctors Study’. Is the belief that smoking causes ‘smoking related’ diseases any more than a faith? Is the ‘Doctors Study’ really objective science, or is it a manipulation of statistics?

Deaths from lung cancer are rare in comparison with other causes. You could say that nearly everyone dies for some other reason before they can get lung cancer.

I know, I know that what I am saying is conjecture, so, once again, I shall show my derived graph of deaths from Doll’s survival graph:

Doll graph again

(The heaviest smokers are depicted by the line on the left. Non-smokers are the line on the right)

As you can see, all the doctors died eventually, although smokers seemed to have died, on average, before non-smokers. Even so, in their 80s and 90s, there were still smokers alive. Note also that the parallel shape of the lines indicates that, once the doctors in the study started to die off, they died at the same rate, whether they were smokers or not. Doctors in their 60s, 70s and 80s died at the same rate. Thus, it is possible to say that it was the non-smokers who were ‘out of step’ – but they all died anyway.

Why did the non-smokers die? Why, especially, did a doctor who was never even exposed, in any significant way to SHS, die? He should be immortal.

======

It is becoming more and more obvious every day that ‘guilt by denunciation’ is just as present now as it was in the Spanish Inquisition. The Savile affair is a case in point. The Smoking Ban appeared from nowhere precisely because of ‘guilt by denunciation’. When Parliament voted for the Smoking Ban, it descended to the level of the Spanish Inquisition. The statistics which were used were akin to the trickery involved in the Sally Clark trial. The difference, however, is that there is no possibility of appeal.

=====

It is for these reasons that THE PEOPLE have rebelled. The politicians cannot understand and are nonplussed. But the REAL Government understands and will easily rebuff the wishes of THE PEOPLE. The REAL Government are those people who decide who can stand as leader of the conservative party, labour party of libdem party. Who decided that Cameron could stand for the position of leader of the conservatives, but Joe Blogs could not? Who decided that both of the brothers Miliband  could stand as leader of the labour party? Who decided that Clegg was just the man to lead the Libdems?

WHO MADE THOSE DECISIONS AND HOW?

Everything seems to be the wrong way round. ENTERPRISE AND COMPETITION should decide upon smoking bars and non-smoking bars. The law can exclude children, but it would be better if parents were in control.

 

 

Advertisements

4 Responses to “Probabilities and Possibilities”

  1. garyk30 Says:

    All of the ASH crowd spew statistical evidence that they are not qualified to judge.

    They quote experts or studies; but, the experts themselves are seldom testifying.

    When they do, as in the McTear event, their testimony is very seldom conclusive.

  2. garyk30 Says:

    This got left off the top of my last post.

    ‘ the trickery involved in the Sally Clark case ‘

    Dr. Williams was not an expert at math when he gave his expert staistical testimony.

    Defense should have noted that and had that part of his testimony thrown out.

  3. garyk30 Says:

    A chart that you could make from the data in the ‘Doctors Study’ would be a bar chart showing the percentage of non-smokers that died from the diseases ’caused’ by smoking vs the % of smokers that died from those same diseases.

    Such a chart would show equality.

    non-smokers = 16 per 19 total deaths/1,000 = 84%

    smokers = 30 per 35.5 total deaths/1,000 = 85%

    As your graph shows, smokers may die younger, on average, than non-smokers; but, non-smokers have about the same probability of dying from a smoking ’caused’ disease as do smokers.

    Or we could say that non-smokers have 99% of a smoker’s chances of dying from a disease said to be ’caused’ by smoking.

  4. Junican Says:

    The methods used by the Tobacco Control Industry follow a formula. Let us follow a typical case:
    1. ASH produce the results of a study which they say that they originated. They claim ‘ownership’ of the study. As far as is known, no one but ASH were involved in organising the study. In fact, it becomes clear, on thinking about the subject of the study, that ASH, being just a ‘media manipulation’ outfit would never know how to go about organising such as study. Thus, we can be reasonable sure that some other outfit (perhaps the College of Physicians) organised the study.
    2. The effect of this system is to take the Royal College and the actual researchers out of the firing line. Any queries are fobbed off by ASH. “Never apologise, never explain”
    3. The system then has the peculiar effect of permitting members of the RCP, along with the BMA and another party, to issue quotes in newspapers about the study, knowing full well that they themselves are protected, since it was apparently a study which they not had anything to do with.

    But in recent years, that system seems to have been abandoned. ASH seem to have become almost completely concerned with surveys, mostly via YouGov, and publicity. Now, the academics who oversee the studies carry the can, but the REAL organisers, the RCP et al, still keep their heads down while using the results of the studies to lobby government.

    The result has been that ‘iffy’ studies have been given the status of ‘science’, totally without foundation.

Comments are closed.


%d bloggers like this: