Today I posted up under the Active Smoking menu page some pdf files sent to me by JLtrader a week or so back. The pdfs are an assortment of historical documents casting doubt on whether smoking causes lung cancer.
One example is a 1963 article by the statistician Joseph Berkson. Here are a few samples from it:
And still hasn’t happened.
Fifty years on, it remains true that animal studies have failed to produce lung cancer. And if the simple theory of a chemical carcinogenic origin of lung cancer had been ‘given up’ by 1963, it is surprisingly alive and well 50 years later. In fact, it’s what more or less everybody believes.
Maybe there never was an increase in lung cancer in the 20th century?
It is interesting, isn’t it, that it’s the medical profession which has been relying on statistical arguments to support its war on smoking, while statisticians (with exceptions like Berkson and Sir Ronald Fisher) are silent? But what do doctors know about statistics?
Anyway, I hope that there’s interesting reading there. And I hope more will be added.
A few days ago I posted up a video showing how convection currents were set up when balloons of air expanded.
Today I have a variant video, in which instead of expanding, packets of air are speeded up on the right hand side of the model, and slowed down on the left hand side.
The same convection currents appear, as can be seen from the motion of the single red packet. The effect is the result of one side of the system being denser and heavier, and the other side being less dense and lighter, causing the unbalanced system to rotate. The difference in densities is apparent in the snapshot above, where there is a dense wedge of bodies in the lower left (coloured blue above), and a larger triangle of more widely spaced bodies (coloured yellow above). These might be regarded as providing on their tops two slopes down which balls can roll.