Tuesday, August 28, 2012

Matt Ridley: The perils of confirmation bias - part 2

What keeps scientists accurate is rivals' scepticism, not their own

If, as I argued in the first of these columns, scientists are just as prone as everybody else to confirmation bias to looking for evidence to support rather than test their ideas ­ then how is it that science, unlike cults and superstitions, does change its mind and find new things?

The answer was spelled out by the psychologist Raymond Nickerson of Tufts University in a paper written in 1998: "It is not so much the critical attitude that individual scientists have taken with respect to their own ideas that has given science the success it has enjoyed... but more the fact that individual scientists have been highly motivated to demonstrate that hypotheses that are held by some other scientist(s) are false."

 Most scientists do not try to disprove their ideas; rivals do it for them. Only when those rivals fail is the theory bomb-proof. The physicist Robert Millikan, (who showed minor confirmation bias in his own work on the charge of the electron by omitting outlying observations that did not fit his hypothesis) devoted more than 10 years to trying to disprove Einstein¹s theory that light consisted of particles (photons). His failure convinced almost everybody but himself that Einstein was right.

The solution to confirmation bias in science, then, is not to try to teach it out of people, for that goes too much against the grain of human nature. Dr. Nickerson points out that the history of science is replete not only with examples of great scientists tenaciously persisting with theories "long after the evidence against them had become sufficiently strong to persuade others without the same vested interests to discard them" but also with brilliant people who remained wedded to their pet hates.

Galileo rejected Kepler's lunar explanation of tides; Huygens objected to Newton's concept of gravity; Humphrey Davy detested John Dalton's atomic theory; Einstein denied quantum theory.

No, the reason that science progresses despite confirmation bias is partly that it makes testable predictions, but even more that it prevents monopoly. By dispersing its incentives among many different centers, it allows scientists to check each other's prejudices. When a discipline defers to a single authority, and demands adherence to a set of beliefs, then it becomes a cult. Medicine did this with Galen and psychoanalysis with Freud.

A recent example is the case of malaria and climate. In the early days of global-warming research, scientists argued that warming would worsen malaria by increasing the range of mosquitoes. "Malaria and dengue fever are two of the mosquito-borne diseases most likely to spread dramatically as global temperatures head upward," said the Harvard Medical School's Paul Epstein in Scientific American in 2000, in a warning typical of many.

Carried away by confirmation bias, scientists modeled the future worsening of malaria, and the Intergovernmental Panel on Climate Change accepted this as a given. When Paul Reiter, an expert on insect-borne diseases at the Pasteur Institute, begged to differ-pointing out that malaria¹s range was shrinking and was limited by factors other than temperature-he had an uphill struggle. "After much effort and many fruitless discussions," he said, "I ... resigned from the IPCC project [but] found that my name was still listed. I requested its removal, but was told it would remain because 'I had contributed.' It was only after strong insistence that I succeeded in having it removed."

Yet Dr. Reiter has now been vindicated. In a recent paper, Peter Gething of Oxford University and his colleagues concluded that widespread claims that rising mean temperatures had already worsened malaria mortality were "largely at odds with observed decreasing global trends" and that proposed future effects of rising temperatures are "up to two orders of magnitude smaller than those that can be achieved by the effective scale-up of key control measures."

The IPCC, in other words, learned the hard way the value of letting mavericks and gadflies challenge confirmation bias.

Matt, an acclaimed author and former Science and Technology Editor for the Economist blogs at This is the second of three columns on the topic of confirmation bias - first published in the Wall Street Journal.

No comments: