by Patrick Cox
November 26, 2018
One of the most important developments of the past decade is the emergence of evidence that many—if not most—published academic studies are worthless, or worse.
This phenomenon is called the reproducibility or replication crisis. I think it should have been labeled the “academic science scam.”
The sorry state of academic research was already known to real scientists, but the topic was taboo. For obvious reasons, researchers who needed relationships with university scientists were motivated to keep their mouths shut about problems they had discovered in published research. A scientist might publish a study that contained evidence refuting a prior study, but the unofficial guild of research scientists tried to keep accusations of fraud to a minimum.
The event that broke this omertà, or code of silence, was a 2012 paper by Dr. C. Glenn Begley, who worked for the biotech company Amgen. He questioned the accuracy of the pre-clinical trials that were the pillars of what scientists believed to be true about cancer.
- His attempts at replicating those trials failed in almost nine out of ten cases.
Subsequent studies verified his alarming claims.
- The University of Virginia’s 2015 Reproducibility Project showed that about two-thirds of the studies published in the field of psychology couldn’t be replicated.
- A more recent project, Many Labs 2, showed that half of the studies published in that arena couldn’t be duplicated, which I suppose is an improvement.
Since Begley’s exposé, periodic discoveries of serious errors and fraud in published papers have kept the topic in the public eye. Most recently, the retraction of papers involving a Harvard Medical School lab have led to another round of media stories.
Bad studies and journal articles are not just a spectacular waste of the nearly $40 billion of taxpayer money devoted to basic research… they also mislead scientists and misdirect future research.
Some of the bad science being exposed can be explained by honest errors, but much of it must be viewed as purposeful manipulation of data to produce positive results.
From a scientific perspective, studies that fail to validate a hypothesis are as valuable as studies that prove a hypothesis. They are much less likely, however, to be published—and unpublished papers do nothing for researchers’ careers.
A survey published in the Annals of Internal Medicine (AIM) reported the results of polling a random sample of 522 statisticians who work in biological and medical research. This is the area I care most about because it translates into medicines that may or may not benefit our health.
- One-quarter of the biostatisticians polled reported requests by scientists for unscrupulous data manipulation or outright fraud.
I believe the AIM paper likely underreports requests for assistance in scientific deception though. Unethical statisticians probably won’t admit that they were asked to fiddle with the data.
Another factor is that many areas of science have become politicized.
Journal editors routinely deny such bias exists, but there’s evidence that many favor research that supports their own position in ideological conflicts.
- As a result, exaggerated or even falsified research that fits an editor’s agenda is more likely to be published.
The implications go beyond the fact that the TED Talk you enjoyed may be complete bunk. It means the swamp of waste and corruption that plagues government has thoroughly infected the university system.
While this is appalling and tragic, it would be far worse if it were still hidden.
The term “peer-reviewed” used to endow scientific papers with the status of a 16th-century papal encyclical.
Those days are over.
Similarly, those who criticized spending tax monies on research with no commercial benefits have been characterized as anti-science barbarians. This, too, is changing, and research funds for universities no longer rise automatically.
Fortunately, corporate funding of basic research in academia has made up the difference, and then some. But in total, corporations spend as much on applied research in their own labs as academia spends on basic research.
Critics who are biased against market forces believe for-profit research is somehow inferior to “pure scientific research,” but that’s not true.
The profit motive acts as a check on fraud because companies need the truth to create useful products and therapies. A scientist who falsifies data for a biotech company will be fired and may be sued.
To be sure, there are many academic scientists and departments that maintain the highest ethical standards. Their work is tainted by the growing scandal of junk science generated by universities, but it is the broader context that’s most worrying.
That context is the growing federal debt that rational economists lose sleep over. The worsening student loan debt—caused in large part by the failure of colleges to deliver educations that command high-paying jobs—is a major contributor to this problem.
Despite record-breaking economic growth rates, there has been little acknowledgment in political circles that the bill will eventually come due. When that day comes, there will be cuts in government funding for many programs, including academic research.
The solution will be practical science funded by investors in for-profit enterprises. Once again, capitalism will make the discoveries and fund the innovations that will take us to the next level.
— The Academic Junk Science Scandal Continues to Unfold originally appeared at Mauldin Economics.