MARC HAUSER has plenty of company when it comes to scientific misconduct.
Hauser, you’ll recall, had built a brilliant career at Harvard. He directed a primate lab and published a long list of scientific papers on topics like the cognitive nature of morality, and the similarities between human and animal behavior. He was a popular teacher, and author of the hit book “Moral Minds.” And then it turned out that he was taking liberties with his scientific data. One paper was retracted, others were corrected, and, earlier this month, he left the university.
Also this month, federal authorities announced that a cancer researcher at the Boston University School of Medicine was inventing data. Two papers have been retracted. The scientist, and I use the term loosely, was shown the door.
These two cases are part of a remarkable flood of scientific retractions. Between 2001 and 2010, the number of retractions increased more than 15-fold, according to a recent investigation by the Wall Street Journal. There were 22 retractions in 2001, and 339 last year, according to the Journal, over a period of time when the number of publications increased by only 44 percent.
It would seem a grim development, this sudden scourge of epic sloppiness and outright fraud in the halls of science. But it’s actually news we should all welcome: We are not witnessing an explosion of misconduct, but a new openness about it.
There are some forces, including easy access to image manipulation software like Photoshop, that are making it easier to fake results. But the problem has festered for decades, and now, finally, science is beginning to get serious about dealing with it.
The most spectacular recent case of scientific fraud came out of South Korea. In early 2004, researchers there announced that they had cloned a human cell, earning front-page headlines around the world, and tantalizing the public with the prospect of future disease treatments. Invitations to collaborate poured in from top biologists. The South Korean government ensured that the lead scientist, Hwang Woo-suk, had every resource at his disposal. He was a national hero.
But it was all a work of fiction. Hwang had wasted people’s time and money, bringing disrepute to biology, and to science as an enterprise.
Scientific fraud can also have more direct human costs. One of the heavyweight champions in this category is Andrew Wakefield, the British researcher who published a 1998 paper suggesting autism was caused by vaccines. The paper was an utter fraud, with invented patients. But the scare pushed down vaccination rates, causing outbreaks of measles. And his theory has persisted among many parents of autistic children, reducing support for urgently needed (genuine) research.
This is one of the most pernicious aspects of the problem: Bad ideas have a way of taking on a momentum of their own, even among scientists. One 2007 paper in the Journal of the American Medical Association looked at high-profile discoveries that were later contradicted by more reliable follow-up studies. Even years later, it reported, the initial (erroneous) “discovery” was still being widely and favorably cited.
One of the biggest obstacles to dealing with fraud in science is the sense, shared by scientists, that science is somehow exceptional as an enterprise. But science is a human endeavor and scientists are no better or worse than the rest of us. One sophisticated analysis suggested that at least 2 percent of scientists have cooked their data at least once – a statistic that is alarming, but also has the ring of truth to it.
Chastened by the Hwang scandal and others, scientific journals are more diligent in pursuing problems, and the Internet has made it easier for other researchers to check papers. But the scientific establishment’s efforts to discover and correct errors are scattershot and overly time consuming. Harvard’s handling of the Hauser case was almost laughably opaque. Scientists still don’t know what to make of many of his papers.
We need to do better. Science is a quest for the truth. And to know what is true, one must know what is false.