John Brignell has an odd response (scroll down to “Hit Parade”) to some of my criticism. He doesn’t link, or dare to even mention my name, so it’s probably rather mystifying to his readers what he is responding to. Brignell goes on the Michael Fumento road, boasting about how the 2,488 hits he got on Monday vastly exceeds the 10 hits he got from me. Trouble is, he got those hits from a link in a comment in a two-day old post, so it’s hardly a meaningful comparison. For what it’s worth, his web counter shows 230k visits in five years, which is less than what I have in two years.
Earlier I wrote:
John Brignell dismisses the [Lancet] study just because:
A relative risk of 1.5 is not acceptable as significant.
Actually the increased risk was statistically significant. You won’t find support for Brignell’s claim in any conventional statistical text or paper. To support his claim he cites a book called Sorry, wrong number!. Trouble is, that book was written by. … John Brignell. Not only that, it was drafted by … John Brignell. Brignell is a crank who dismisses the entire field of modern epidemiology as some sort of plot by scientists to scare people.
Brignell’s response is:
Among the charges in the web log were that the author is not an epidemiologist, so not qualified to comment on epidemiology, and that he is innumerate for suggesting the relative risks of 1.5 are unacceptable for observational studies. The first is like saying you have not committed mass murder therefore you are not entitled to write about crime. Critics of observational studies have included great scientifically inclined epidemiologists, such as Alvan R Feinstein, Sterling Professor of Medicine and Epidemiology at Yale. The also great R A Fisher would have no truck with them at all. The second accusation is typically hyperbolic. An innumerate person would not even be able to begin discussing a concept such as risk ratio. There is a substantial body of opinion outside mainstream epidemiology that is critical of such lax statistical standards. Correspondence to Number Watch confirms that many professional statisticians are appalled by what is going on. Besides which, the proof of the pudding is in the eating. The book The Epidemiologists begins with some examples of the many completely contradictory headlines generated by popular epidemiological studies.
Actually I didn’t say that Brignell shouldn’t be commenting on epidemiology because he wasn’t an epidemiologist, but that the only support he offers for his 1.5 claim is his own opinion. He even admits that his view is outside mainstream epidemiology and still has not offered any cite or argument to support his claim. I guess Fisher might well be on Brignell’s side, since Fisher rejected the idea that smoking causes lung cancer, but very few deny this any more. And Brignell’s 1.5 risk ratio principle is innumerate. According to his principle, for example, the observed ratio of male to female births of 1.03 is not significant and we can’t conclude that male births are more likely.