"Mike Lyle" wrote in message
...
Duncan Heenan wrote:
Sure, sure. As a statistician you understand the problems. The clot
who says "My grandad smoked forty a day and lived to 97", however,
does not understand the statistical basis of epidemiology, and is
usually drawing an inference you would not draw. You know that
knowledge advances unevenly, and with set-backs; and there are indeed
intellectual fashions; but even the least solemn medical science does
discover facts.

Mike.

I think you use the term 'statistical bias' loosely. Eliminating bias,
particularly sampling bias is at the root of statistics, and therefore many
aspects of epidemiology.

Medical science stated as a fact that full fat milk and sunshine were good
for me as a child. It now states as a fact that they are not. When the
orthodoxy (or fashion as I unkindly called it) changes, the next idea will
be stated as a fact. It is a brave person working in a profession who is
actually prepared to stand up and shout 'The Emperor is wearing no
clothes!". Which is more or less where I can in I think.