Wednesday, August 26, 2015

Fraud... in Science

The BIG news in Science, at least from where I sit, is Science Fraud.  A behavioral science Ph.D. student at UCLA teamed up with a distinguished prof at Columbia University to publish a paper last December in one of the most prestigious Journals in science, called Science.  When some other Ph.D. students questioned his numbers, it came to light that he had fabricated most of the data for the paper, and, not surprisingly, had also fabricated several aspects of his academic resume.

A perhaps more important instance of science fraud that didn't get much attention was the case of a Duke Uni cancer researcher named Anil Potti.  Potti apparently faked a large body of important cancer research that misled many other researchers and resulted in numerous paper withdrawals and Duke's firing of Dr. Potti who now practices oncology in North Dakota.  North Dakotans beware!

I don't want to give the impression that these are two isolated cases, one Japanese anesthesiologist, Dr. Yoshitaka Fujii, currently holds the current record for scientific retractions. Dr.  Fujii, was found to have made up data in 172 studies. He now has, in the past four years, 183 retractions when you combine his science fraud with his failure to obtain ethical approvals to perform his research.

These frauds are scary, but sadly this is the tip of the iceberg.  We scientists are tempted to:
1) Collect tons of real data, then discard most of it in order to pick out the findings that support our hypotheses.
2) Ignore negative findings in order to report only positive results- part of this is a result of pressure from reviewers who tend to reject papers that found NO results of a treatment.  I call it the, "We knew it all along" syndrome. 
3) Eliminate participants or data that keep us from obtaining "statistical significance". 
4) Find things that support our hypotheses so that we can keep that grant money coming our way!

This last one is the foundation for the whole science fraud enterprise I think. There is tremendous pressure on scientists to:
a) Publish a LOT of papers in prestigious journals, b) get grant funding and c) keep grant funding.

Science is a tricky business.  It is hard to foresee everything that should be foreseen- so why not "foresee it" after the fact.  After all, a science paper is a historical document of something that happened one time in the past.  Who's to know?

"But science is self-correcting!  We replicate studies and ultimately find the truth!"  Yeah right.  One of my own students was just bemoaning to me yesterday the almost impossibility of publishing a study which merely replicates a previous study.  "Why publish that reviewers say, we already know that!"  I do not recall ever seeing a paper published which was an exact replication of a previous paper.  Doesn't happen because it won't be published.

So, in the end, it turns out, surprise of surprises, that scientists are NO MORE trustworthy than anyone else in society. Scientists lie and cheat, and that's just the nature of mankind.  I am a cynic and skeptic, true enough, but I come by it honestly (maybe).

So, want to put all your faith in scientists compared to those huckster religious types??

Think again.

No comments:

Post a Comment