[Previous entry: "Gender Bias at Psychology Today?"] [Main Index] [Next entry: "The hate contingent"]
02/07/2006 Archived Entry: "Questions to Ask Scientific Authority"
My last FOXNews/ifeminist article entitled http://www.ifeminists.net/introduction/editorials/2006/0201.html Questions to Ask Scientific Authority elicited some interesting responses. I reprint one of them below. (The column was also reprinted by LewRockwell)....
I write from England. Further to your article on Lewrockwell.com on 3rd Feb 06, I would like to follow on a point about scientists. I am a litigator in labour law in Britain, but my first degree was in Genetics/Biochemistry and I did 4 years post-grad science before moving into law.
I have always taken a working definition of a scientist as "A person in search of a grant". You are quite right to pick up in your article on this tendency. Scientists will claim that research into yeast may help provide a model for cancer in humans, by looking a cell division. That may be so, but if you want to know why your dog is sick, why start by looking at a mushroom? Relating to statistical studies, I see a more serious problem with statistical surveys. The fundamental point of a scientific experiment is to take one variable, add or remove it, and see what happens in a controlled situation. Any proper science will have, if possible, a positive control, which gives the result expected e.g. a test for the presence of copper in a liquid has a copper sulphate solution as a positive control. That will show that the experiment does "work". It will also have a negative control, say calcium sulphate, which will not give a result, and (tends to) show that the test can distingush a true result from a false one. The remainder of the tests are for the presence or absence of copper where that is an unknown. Any mathematician would say that a problem with two unknowns is insoluble.
Perhaps more importantly, a scientist should be looking for a mechanism. i.e how does a process work, what is the cause and what is the effect? Scientists have shown a "link" between lung cancer and smoking. This was by a statistical process, but that does nothing to show how lung cancer is caused, but simply that it occurs. That is like showing a link between steel production and blast furnaces. That wont help you make steel, but tells you where you may look for steel to be produced. Any research that does not provide or explore a mechanism is simply looking for a starting point, nothing more, and deserves no more attention.