John Brignell
From Wikipedia, the free encyclopedia
 |
Please see the discussion on the talk page. |
John Brignell is a retired Professor of Industrial Instrumentation who has a part-time interest in debunking the use of what he sees as poor science and false statistics in media.
John Brignell was educated at Stationers' Company's School and began his career as an apprentice at STC. He studied at Northampton Engineering College (which became the City University, London) and took the degrees of BSc(Eng) and PhD of the University of London.
He joined the staff at Northampton and was successively Research Assistant, Research Fellow and Lecturer. He worked in a number of areas including dielectric liquids and computer aided measurement, co-authoring a book "Laboratory on-line computing" in 1975. He was for ten years Reader in Electronics at the City University and held the Chair in Industrial Instrumentation at Southampton for twenty years from 1980.
Brignell retired in the late 1990s from his academic career and now devotes part of his time to his interest in debunking what he asserts to be the use of poor science and false statistics common in much of today's media. In this context, he has expressed controversial opinions on many subjects. In particular, he has disputed the reality of anthropogenic global warming, questioned the relationship between second-hand smoke and lung cancer, and suggested that the hole in the ozone layer may have existed before the rise in the use of chlorofluorocarbons.
Contents |
[edit] Career and awards
Brignell has researched and written extensively in the area of sensors and their applications, and in 1994 co-authored a book with Neil White on "Intelligent sensor systems". He had an extensive private consultancy practice for many years and has advised some of the larger international companies, as well as many small ones in the UK, on all aspects of industrial instrumentation. He pioneered the use of a number of technologies in sensing, such as thick film, and latterly turned his attention to the considerable possibilities of micro-engineering. He was elected Fellow of the Institute of Physics, the Institute of Measurement and Control, the Institution of Electrical Engineers and the Royal Society for the encouragement of Arts, Manufactures and Commerce. In 1994 he was awarded the Callendar Silver Medal by the Institute of Measurement and Control. He served on the ISAT Committee of Institute of Physics from its inception and was the founding chairman of the first joint professional group of the Institution of Electrical Engineers (J1), having served on both its predecessors (E1 and C11).
[edit] Number Watch
Brignell believes that the quality of scientific research, particularly in some newer fields of science, has fallen below an acceptable level. Brignell's response has been to publish a series of books on the phenomenon as he sees it, and to support those books in July 2000 he created the Number Watch website. It which describes itself as being "devoted to the monitoring of the misleading numbers that rain down on us via the media" [1]. On that site he regularly draws attention to media articles that fail to meet his standards of scientific or mathematical credibility. Number Watch also contains commentary on general political issues.
Brignell has self-published two books purporting to debunk the mathematics behind media scares, Sorry, wrong number! and The epidemiologists: Have they got scares for you!, under the name Brignell Associates. Brignell appears to doubt the usual view on ozone depletion: "...hole in the ozone layer that was probably always there and an unproven theory as to how it was caused" [2].
Brignell's methodology is to draw attention to media articles which he says are based on bad science or statistics and explain what is wrong with them. There are a number of instances where he has drawn attention to published papers that contradict each other. Critics of Brignell have suggested he shows bias in his choice of targets. Supporters argue that this merely reflects the prevalence of particular kinds of arguments in the media.
Brignell has expressed delight with the feedback from the "encouragement and support I have received from some of the giants of the pro-science movement in the USA -- in no particular order Steven Milloy, Alan Coruba [3] [sic], ... Bob Carroll, Michael Fumento and S. Fred Singer." [4].
As discussed below, Carroll's support was subsequently withdrawn. On the other hand, as illustrated by the examples below, Brignell's approach and views are closely consistent with those of Caruba, Fumento, Milloy, and Singer. However, whereas these writers have attracted controversy because of their paid work for corporate interests, there has been no suggestion that Brignell has received any outside funding. Brignell is a strong supporter of Milloy, to whom he dedicated his book, Sorry. Wrong Number![5]
Number Watch also contains extensive commentary on British politics. Brignell criticises all the major parties as being disguised Greens, but is particularly critical of the Labour Party. He supports the views of the UK Independence Party, but notes that, given its lack of political power "The bloodless Green coup that overtook the Conservative party virtually means the achievement of Orwell’s one party state."[6]
[edit] General critique of modern scientific research
Brignell cites a number of factors he believes have led to a general decline in the quality of scientific research. He believes that modern scientists do not receive the broad scientific education that they used to, and consequently do not fully understand the need for scientific rigour in their research and do not fully understand techniques derived from outside their own field. He asserts that scientists are judged on the quantity of their work, not the quality. Once, in order to become famous, a scientist needed to make a major scientific breakthrough. Now, Brignell says, the easiest path to fame is to become recognised by the media as an appropriate spokesman on some issue. Brignell further suggests that the media compound the problem through being insufficiently critical of the science they report on.
Brignell also criticises such practices as data dredging [7], misuse of significance tests [8] and linear extrapolation of trends.
[edit] Statistical Significance
In preparation for the publication of his book Sorry, Wrong Number! Brignell spent two years collecting relevant articles from The Times. He found "hundreds of contradictions in published claims", that is, reports of studies that contradicted each other. (Brignell, Sorry, Wrong Number!) He suggests that when two studies come to mutually contradictory conclusions, one of them must be false. Others have come to the same conclusion, and it has been suggested that a large number of published studies are false.[9]
Brignell suggests that the cause of such contradictions is the reliance on statistical significance alone as an indication of truth. He suggests that this reliance is undermined by a combination of four main factors:
- the use of 95% confidence levels ('the one in twenty lottery');
- the acceptance of low relative risks (RR<2.0 or RR>0.5);
- the absence of randomisation;
- small studies.
(Brignell, Sorry, Wrong Number!)
[edit] Relative risk
Brignell states: "most scientists [...] take a fairly rigorous view of relative risk (RR) values. In observational studies, they will not normally accept an RR of less than 3 as significant and never an RR of less than 2. Likewise, for a putative beneficial effect, they never accept an RR of greater than 0.5." Brignell applies this claim (and several others, such as the use of P<0.05) to dismiss studies of the risks of passive smoking, citing Steven Milloy. [10]
The notion of relative risk proposed by Brignell is supported by some epidemiologists in the context of single studies, particularly those identifying new risks without a biomedical basis (see relative risk) and it is further supported by empirical observations and a theoretical underpinning[11]. In general, most academic journals in epidemiology (as in most other fields) rely on tests of statistical significance and other forms of hypothesis testing.
[edit] Speed Cameras
Brignell states "One of the notable areas for fraudulent numbers is in policy on speed control on the roads" "The latest toy for ministers and bureaucrats is the speed camera."
He asserted that a government report on the topic was "a startling example of statistical chicanery from figures of road fatalities issued by the bureaucrats to justify a massive investment in these devices, following a trial introduction in 2000." "They claim a 'huge' reduction of 49 fatalities between 1999 and 2000 when the cameras were introduced. We are not supposed to notice that there was an almost equally huge increase of 44 fatalities between 1998 and 1999, for no apparent reason at all. Thus the change over two years was less than 1%, monumentally insignificant." ([12],'And from the Ministry of Truth').
Brignell argued that politicians were taking advantage of regression to the mean to make claims about the effectiveness of speed cameras. His argument was that the number of accidents occurring at a particular location is a random variable. When these numbers are sorted, the locations at the top of the list will be those that have produced an above average number of accidents. In subsequent surveys these locations will produce lower numbers of accidents. The installation of speed cameras at these locations is coincidental to this trend.
[edit] DDT
On his website Brignell criticises what he calls the "deadly legacy of Rachel Carson" in curtailing the use of DDT. He endorses arguments that have been made by critics of the ban on agricultural use of DDT, but rejected by USAID, that poor countries have been blackmailed into abandoning the use of DDT by the threat of withdrawal of aid. [13]. Brignell argued that the resurgence of malaria in Sri Lanka was a case in point.[14]
[edit] Counting the dead
In the discussion that followed a study published in The Lancet estimating the number of people killed in Iraq since the invasion, Brignell wrote that the use of a relative risk figure of 1.5 was inappropriate. "A relative risk of 1.5 is not acceptable as significant," he wrote. As noted above, Brignell argues that risk increases of less than 100% should be ignored for studies in observational science, based on his view that there are practical limits to the accuracy of such analysis. [15]
[edit] EPA and Second Hand Smoke
Brignell has criticised the US Environmental Protection Authority finding that second hand smoke caused lung cancer, primarily on the basis of arguments about relative risk, but also because of concerns about the use of a one-tailed test and his belief that meta-analysis is not appropriate in epidemiology.
- "The EPA applied to environmental tobacco smoke (EPS) in 1992 a statistical technique called meta-analysis in which the data from multiple small studies are combined to make one large study. Most respectable scientists would agree that meta-analysis might just be appropriate for combining carefully controlled animal experiments or well-regulated clinical studies, but not for epidemiology! However, our heroes did not stop there because it did not produce the right result. Let us be quite clear about it, by all the conventions of science and applied statistics the EPA study established that there is no significant risk of lung cancer due to passive smoking. What did they do about it? They moved the goalposts. They actually reduced the confidence limit from the normally accepted lower limit of 95% to 90%, equivalent to doubling the chances of being wrong to one in ten. It was even worse than this, but I don't want to get too deep into details of statistical methods. There are many other serious defects in the EPA report; such as ignoring about twenty confounding factors and choosing to ignore a study that produced a statistically significant decrease in risk. So after all this knavery, what result did they come up with? They calculated a risk ratio of 1.19. [...] As a basis of comparison a report in The Times (Aug 6, 1997) quotes a study in Uruguay that produces a risk ratio of 4 for the development of lung cancer in heavy consumers of dairy products. This is an increase of 300%, compared to the EPA's 19%, which presumably means that some 50,000 Americans will die each year of eating rice pudding. As a correspondent in the next day's Times asks "What are my chances if I am regularly in the same room as rice pudding eaters?" The most extraordinary thing about the EPA-ETS fiasco is that, after four major statistical fiddles, they could only come up with a pathetic risk ratio of 1.19. Even four out of ten of our imaginary scientists in Chapter 2 did better than that on a non-existent effect." (John Brignell, Sorry, Wrong Number, p.128)
[edit] Criticism
Brignell's statistical arguments and views on environmental issues have been criticised by, among others, Tim Lambert. On DDT, Tim Lambert argues that Brignell's assertions about Sri Lanka are historically incorrect. [16].
Responding to Brignell's criticism of the Lancet study, Lambert argued the increased risk meets the conventional criterion of statistical significance, and that it is appropriate to apply the same standards of significance to observational studies as is applied to more tightly controlled experiments:
- Suppose we had perfect records of every death in Iraq and there were 200,000 in the year before the invasion, and 300,000 in the year after. Then the relative risk would be 1.5 and Brignell would dismiss the increase as not significant even though in this case we have absolutely certainty that there were 100,000 extra deaths. [17]
The ensuing controversy reflected the more general debate about relative risk.
Bob Carroll (author of the Skeptic's Dictionary) initially accepted Brignell's argument against the EPA. However, he changed his mind on the basis that the "scientific principle" (relative risk less than 2) that Brignell used to reject the finding was not recognized by epidemiologists or statisticians.
[edit] Books
- John Brignell, Sorry, wrong number!, Brignell Associates, September 2000. ISBN 0-9539108-0-6
- John Brignell, The epidemiologists: Have they got scares for you!, Brignell Associates, July 2004. ISBN 0-9539108-2-2.
