The Silly Isles
Well, fancy giving money to
Might as well have put it down the drain.
Fancy giving money to the Government !
Nobody will see the stuff again.
Well, they've no idea what money's for –
Ten to one they'll start another war.
I've heard a lot of silly things, but, Lor'!
Fancy giving money to the Government!
A P Herbert
August, and the silly season is in full swing throughout the British Isles. Cancer cures and causes pour out of the media. For this month only – cow’s milk good, peanut butter bad. Even incense is now carcinogenic. Number Watch has a suggestion that might save a few hundred trees. Why not just list the substances that are not supposed to cause cancer? Reducing cholesterol in the elderly could now be a Bad Thing. Bacteria are pouring in from outer space. More people voted in a "reality" TV programme than in the European parliamentary election. This show, of immaculate banality, dominates the front pages of the tabloids and the question of the moment is whether two of the participants copulated.
Unbelievably, the Foot and Mouth crisis is still with us and the slaughter of healthy and innocent farm animals in Britain continues unabated. The Prime Minister, who before the election took personal responsibility for the holocaust, has now lost interest and his spin-doctors are seeking to deflect the blame onto greedy farmers. The disinfection of farms has been suspended on the grounds of cost. There is no agreement as to how many billions the whole fiasco has actually cost, but it is certainly greater than the annual income of the entire farming industry.
A new bunch of animal welfare SIFs have arrived from America. They call themselves Peta. Have they come to put a stop to the widespread cruelty and killing on farms? Have they hell. Their mission is to put an end to Britain’s most popular outdoor recreation, angling. It is remarkable how, through the ages, people have derived more satisfaction from stopping the pleasures of others than from simply doing good.
People who ate lamb in the early 1990s may have been exposed to BSE, according to UK Government officials (BBC). What evidence do they have for this? According to Nigel Hawkes of The Times, precisely zero. The new Food Standards Agency has reverted to type and become the Food Scares Agency, which puts Number Watch in its place for excessive and uncharacteristic optimism about the outbreak of common sense in its early life (see September). The BBC appears to have abandoned any appearance of balanced reporting in favour of the promotion of scares. For example, it never misses an opportunity to pump up the post Kyoto hysteria. A whole Radio 4 programme was devoted to an attack on American industry, not for their naughty emissions but for their daring to question the true faith. The decline of the BBC, and its worldwide reputation for truth and honesty, is one of the saddest aspects of the rise of the new establishment.
When numbers are not enough
The British media are at last beginning to face up to reality and are using the R word. Characteristically, they flip from one extreme to another and switch from affirming that everything in the garden is lovely to declaring that all is disaster. The numbers (e.g. manufacturing output, The Times August 7th ) show that the 1998/9 recession reversed during 1999, reappeared briefly in 2000, reversed again and then came back with a vengeance in 2001. As we have frequently observed in these columns the temporary recoveries were largely illusory and fuelled by the communications bubble. This is amply indicated by the case of Marconi, discussed in our July number of the month. While the numbers in the internet field, for example, were showing relentless expansion, seasoned observers were saying things like “That is all very well, but how are they going to get their money back?” The numbers said one thing, common sense another.
One of the inventions in Sorry, wrong number! was Spreadsheet Man, who has become a dominant force in our society. Users of packages such as Excel and those familiar with business plans know how easy it is to slip in chains of assumptions. The results of such documents often have more to do with the motivations of their authors (both conscious and unconscious) than what is likely to happen in the real world.
The same applies to economists and their models. It is not just that the science base of such large models is inadequate (as in the case of global warming) but in human affairs sentiment is such a powerful influence that the models become meaningless. Thus the notorious UK Treasury model has consistently produced optimistic forecasts for UK economic growth, which encouraged the Chancer of the Exchequer to stick the national neck out with big spending plans. There is a marked contrast between the actions of The Fed, which seems to have its feet on the ground, and its European counterparts. The Bank of England at last made a token reduction in interest rates, too little too late as usual, while the European Central Bank sails on blithely oblivious of what is happening in the real world. Most of the economic “experts” in the media, whose writings are based wholly on chart-gazing, side with the establishment and many believe the cuts were unnecessary. Typically, they dismiss manufacturing industry as mere “metal bashing” and a diminishing portion of the economy. This is characteristic pigeon-holing and failure to understand interrelations, as well as over-estimating the robustness of the service sector.
There is a form of number-blindness that is analogous to snow blindness. The omnipresence and beauty of the numbers overwhelms the senses and totally obscures the inevitable collision ahead. As with British farmers in the wake of Foot and Mouth, the search for scapegoats will begin among the victims, and certainly not among the politicians, bureaucrats and journalists.
GLENDOWER: At my nativity
The front of heaven was full of fiery shapes,
Of burning cressets; and at my birth
The frame and huge foundation of the earth
Shak'd like a coward
HOTSPUR Why, so it would have done at the same season, if your mother's cat had but kittened.
Henry IV, part I
Here is a lovely silly season story from the New Scientist that was picked up widely across the media, which illustrates one of the more common statistical fallacies
|June babies have higher risk of anorexia|
|19:00 08 August 01|
Anorexic women are most likely to have been born in the spring or early summer, reports a researcher in Scotland. The finding raises the possibility that a common winter infection, such as flu, may predispose an unborn baby to the condition.
"It's not the whole answer," says John Eagles of the Royal Cornhill Hospital in Aberdeen. But it could be an unrecognised cause of anorexia nervosa, which affects around one per cent of girls in the US.
The team studied 446 women who had been diagnosed as anorexic and observed that 30% more than average were born in June. As the average monthly births is about 37, the June number must be 48. At first sight this looks like a significant result (at least by epidemiological standards). Applying the binomial distribution to a random selection from a population of 446 with a probability of 1/12 indicates that the probability of getting 48 or more in a random month is about 3%. But that is not what they are doing! They are making twelve such selections and then picking the biggest. Application of the theory of the statistics of extremes tells us that the probability of the largest of twelve such selections being 48 or greater is 30%.
Let us illustrate with a simulation. Using the random number generator in Mathcad, 446 objects were placed at random into twelve categories, and a typical result is as shown in the histogram of births against months below.
The largest value was 51 (in November on this occasion). When the simulation was run ten times the largest values were: 51,41,42,46,48,46,43,53,42 and 46; giving an average of 45.8.
In order to obtain the distribution corresponding to selecting the largest value from twelve samplings we have to perform the largest value transformation on the (cumulative) distribution:
Under the conditions of the "research" the distribution for a random month is as shown below in red, while the distribution for the highest month of twelve is shown in blue. The move to the right and the reduced scatter are typical.
Some readers have asked for an explanation of the largest value transformation. Here is a simple proof from first principles:
By definition the distribution of a random variable v is
If two selections are made from the same population, the distribution F2(x) of the largest value of the two is defined as the probability that the largest is less than or equal to x. This, however, is identical to saying that both fulfill that condition. Furthermore, the probability that two independent events occur is the product of their individual probabilities. Thus:
By the same argument this can be extended to the largest of any number of samples:
The simulation and the theory agree and both demonstrate that the largest value of 48 observed for June is totally unremarkable.
This sort of nonsense appears at regular intervals. No mechanisms are offered, or at best ludicrous and far fetched ones. "Researchers" take some affliction, such as schizophrenia, match them against birth months and, lo and behold, they inevitably find that one month (or quarter) produces more cases than average. A survey in the village pub established that ordinary punters find this sort of thing ludicrous, so why do journalists and even self-styled scientific journals fall for it?
The whole is greater than the sum of its parts
That scourge of the global warming frauds, John Daly, has spotted a beauty from The University of East Anglia Climate Research Unit. Their latest charts show that the northern hemisphere has cooled during 2000. They also show that the southern hemisphere has cooled during 2000. Yet their chart for the whole world shows that it has continued to warm. Not really surprising, as the bureaucrats have poured many millions of taxpayers' pounds into Norwich to produce evidence for global warming, and in the modern world what you pay for is what you get. Anyway, it is so passé to expect logical consistency in post-modern physics.
Another global warming myth bites the dust. A study by Tillinghast – Towers Perrin shows that, when actual insured hurricane damages are adjusted to reflect current property values and the increase in the number of people living toward the coast, insured damages in the 1990s were not unusually high compared to other decades in the 20th century.
To everything there is a season
Mid August saw the Silly Season at its height, with scares and breakthroughs in abundance. Drinking hot tea prevents skin cancer. Data dredges involving twins are always a good bet. Boy twins are likely to be heavier if their twin is a girl. Having twins reduces the risk of ovarian cancer.
Silly risk ratio of the season came from a UK Government backed study. Those living near landfill sites containing hazardous waste were at a 1% higher risk of having a baby with congenital defects.
Breakthrough launch of the season from the Royal Brompton “These results are probably the biggest breakthrough in the treatment of coronary heart disease, since the introduction of aspirin in the 1980's.” With all the breakthroughs we have had it is amazing that there is any heart disease left. The vCJD breakthrough was based on one case, a woman who could have had the disease.
It was inevitable that one of the tackiest pieces of “research” is about passive smoking. It involved the usual anecdotal evidence and even the Trojan number of people in the study was a pathetic almost 300.
According to the BBC: The researchers wrote: "Passive smoking has been a public concern for a while, because it is a major indoor pollutant to which a substantial segment of the population is exposed." They added that the results of the study was further evidence that smoking in the workplace should be far more tightly controlled.
There was a time when academic research was more than just politically correct propaganda. Of course, SIF group Ash welcomed the report and the egregious Clive Bates had his say – "It's not surprising that asking people to sit in an environment filled with substances known to toxic is bad for their health. There's absolutely no justification for having anything but non-smoking workplaces.”
The silly season is also a time for meeting old friends. The British media were full of stories claiming that greenhouse gases were killing people. It transpires that this story emanated from Devra Lee Davis, who was nominated in Sorry, wrong number! as one of the great apostles of scare. It was she who predicted eight million deaths from particulate pollution and the frightening slump in male births (by 0.02% in 44 years). Now CO2, a substance vital to life on earth, is included in causes of heart disease and, of course, thousands more will die. The story originated in the increasingly disreputable journal Science.
Number of the month - 40
40 is the number of staff at the Climate Research Unit of the University of East Anglia. A large portion of the funds available for scientific research in the UK have been diverted to the establishment and maintenance of this unit. One is John Daly. Compare and contrast. In honour of the achievements of CRU our section on Chartmanship has been updated.
you are nuts. salty ones. :)
Richard Thompson [firstname.lastname@example.org]