Things that used to be on a blog.

Confirmation bias, denialism and Morton's Demon.

November 30, 2012

Anyone who has ever done research will be familiar with the problem of confirmation bias – hearing what you want to hear. Anybody doing research in the social sciences has to be constantly aware of the possibility of confirmation bias, of selecting results and readings that fit the hypothesis and either ignoring or eliminating things that don't quite fit. I don't mean rejecting obvious outliers where the observations are so far from the rest that a mistake can be assumed – I mean shaving the results to suit what the experimenter expects to find. This may not even be a conscious act, because doing it consciously approaches fraud and most people are basically honest.

The classic case in the hard sciences of confirmation bias is cold fusion. Ponds and Fleischmann found what they wanted to find and then stopped looking. In the social sciences there was Cyril Burt's just-too-good statistics about separated twins and Margaret Mead's willingness to believe whatever some young girls told her. In medicine there was William McBride's work on Debendox. I don't think any of these people started out to do the wrong thing, but they all did it anyway because they wanted to confirm their beliefs.

Confirmation bias is rife in paranormal research, largely because this research is carried out by true believers. While there have been cases of deliberate fraud, the most common problem is testing until some anomalous situation arises and then stopping, claiming evidence of psychic or paranormal powers. I was adjudged the most psychic person in the room at a skeptics' function once, and I did this by correctly calling a coin toss seven times in a row. To a paranormal researcher this could be seen as evidence of my super powers but with about 120 people in the room you would expect to take six or seven tosses to eliminate everyone.

The next stage is denial, where results or data which contradict beliefs are rejected. Again, this can be a totally unconscious matter, but to be true denial it has to be deliberate.

One aspect of denial is that often people will be presented with evidence that conflicts with what they already believe, but if it still agrees with their general belief system they will accept it and consequently hold two contradictory opinions at the same time. The ability to hold two contradictory positions simultaneously and believe them both to be true is described by the word George Orwell invented for 1984, "doublethink", defined as "the power of holding two contradictory beliefs in one's mind simultaneously, and accepting both of them". Anyone who can practice doublethink is going to be resistant to conflicting information because they can assimilate it without rejecting what they already know.

I want to finish by offering a light-hearted explanation of this phenomenon of resistance to conflicting ideas. It's called "Morton's Demon" and was first described by Glenn Morton in 2002 as a means of explaining why creationists will listen carefully to what you say and then completely ignore it.

In 1871, James Clark Maxwell suggested a thought experiment, now referred to as "Maxwell's Demon". Given two rooms separated by a molecule-sized door, a demon at the door could allow fast molecules to go from room A to Room B and allow slow molecules to pass from B to A. This would eventually cause a temperature difference between the rooms which could be exploited to do useful work. If the demon used no energy this would be a form of perpetual motion machine, violating the second law of thermodynamics. While this might convince someone who didn't know how the universe works, it was soon challenged on the basis that the demon would in fact use energy to observe the molecules. This is an example of how science works – if something is proposed which defies what we know then the first thing to look for is why it might be wrong.

Glenn Morton expanded the idea of Maxwell's Demon to explain the resilience of nonsensical or wrong beliefs. He was particularly concerned about young Earth creationists but his demon applies to a much wider class of people.

His demon sits at the front of the mind and filters incoming ideas, only letting in those with which the person agrees and blocking the rest. This is much more powerful than any system where the ideas are tested for compliance by the mind and then rejected – they don't even get considered in the first place. I have seen people repeat the same faulty arguments within minutes of being informed, with evidence, that they are wrong. And I do mean minutes. As these people appear to otherwise be functioning human beings who can even tie their own shoelaces it seems reasonable to infer that the counterarguments are not even being perceived, let alone being evaluated and rejected.

Hmm. I think I see a thesis in cognitive psychology somewhere here.

This article was published as the Naked Skeptic column in the May 2011 edition of Australasian Science