08 August 2016

Scientism - Its Fatal Flaw

Well, *I* believe in SCIENCE! 

Heard that before? It's certainly nothing new - it goes back to at least Voltaire.

Scientism is an expression in use for most of the 20th Century, and is often used to refer to science applied in excess - or applied unreasonably. The term scientism can generally apply in either of two ways:
  1. To indicate the improper usage of science or of scientific claims,  
  2. To refer to a belief that methods of natural science form the only proper elements in any inquiry.
In a broader sense, scientism is also used to describe the invocation of science as a focus of worship, generally by people who would describe themselves as atheists. It's sort of like Methodism, or Daoism, or... you can fill in the blanks here.

Two recent articles, "Lies, Damned Lies, and Medical Science", and "Trouble at the Lab" draw some obvious and frightening conclusions about this approach or life view.

John Ioannidis, a physician and mathematician, published two seminal articles in 2005. They are among the most-cited papers in all of modern science - and they are incredibly embarrasing to scientists. In the first paper, Ioannidis convincingly showed why 80 percent of non-randomized scientific studies turn out to be wrong. Fully 25 percent of supposedly gold-standard (and thus far more expensive) clinical trials gave incorrect results. It is from studies like this that the medical doctors that you and I seek help from base their diagnoses and treatment protocols. Our lives depend on these being correct. These incorrect results include recommendations to use hormone-replacement therapy in post-menopausal women, that mammograms and PSA tests are critical for extending lives, that anti-depressants such as Prozac, Zoloft, and Paxil can help depression, that doing puzzles will ward off Alzheimers Disease, and that drinking lots of water during intense exercise is helpful.

The problem? Not one of these turns out to be true. THOUSANDS of stories in magazine articles have been written based on these published studies. The number of studies that contradict other studies of the same thing are so high that The Week magazine actually has a section called "Health Scare of the Week".

Many physicians, on their own, have discovered that just taking a patient off of every drug can improve their health immediately. 

The truly glaring problem: the large majority of these studies cannot be replicated. This means that other groups cannot repeat the same experiments and get the same results. Amgen, an American drug company, tried to replicate 53 landmark studies in basic research on cancer. They were able to reproduce the results on just 11 percent of the studies. In a separate study done by Bayer, the German pharmaceutical company, only 25% of published results could be reproduced. These analyses aren't being published by disgruntled scientists, but by editors in the premier of all science journals: Nature. Dr. Ioannidis says that between a third to a half of  medical research results has been shown to be untrustworthy. He suggests that physicians, when faced with all this potentially lethal error and confusion... simply ignore them all!

Ioannidis' second paper explains why these flawed studies happen and get published in peer-reviewed journals. Without belaboring the details (you can read them yourself), it comes down to many things - but things that compound themselves:
  • The "publish or perish" ethos for young scientists to get tenure or grants
  • Ignorance of what constitutes statistical significance among most scientists
  • Ego
  • Fear of reprisals by peers or superiors
  • The tendency of scientific journals to publish almost exclusively "new" and "exciting" discoveries
  • Bias in research study design, bias in analysis, self-serving interpretation
  • Fraud.
This latter issue is interesting, and when identified firmly it is supposed to lead to retractions of published articles. However, a University of Edinburgh study of 21 confidential surveys of scientists worldwide found that only 2 percent of them admitted to falsifying or fabricating data - but 28 percent said they knew of colleagues who engaged in these practices! If that difference hints to you at a broader problem, then give yourself three stars. 

The problem with Scientism is that it falls for the oldest trick in the book: it worships at the feet of the Golden Calf. One of several modern versions of the Golden Calf is Science. But like all man-made things, their faith is based on something that is fatally flawed. It is very, very human.

Are we advocating that people not trust science? Absolutely not - just don't bet your life on it, and certainly don't pour your faith and belief into it! Science is still far better and more honest than the The Talking Heads and Corporate Paid Pundits on talk radio or "fair and unbiased" news channels.