Thursday, November 26, 2009

That's Science

Mr. Ovshinsky knows as much or more about the development and production of alternative energy as anyone on the planet. He developed the technology and designed the production method that made it possible to produce solar material “by the mile.” When he proposed the idea years ago, based on the science of amorphous materials, which he invented, he was ridiculed.

I ran into that quote in a recent New York Times opinion piece about alternative energy. It demonstrates something about science that I feel is not generally understood. That is, every scientist is ridiculed for his ideas. This is normal, this is how science works.

Spend any time at all in science seminars and you will pick out a few people, every department has at least one, who will attack any idea presented to them. They are almost always quite reasonable people, but the vast majority of ideas in science are wrong. It is not the ability to produce ideas that separates science from philosophy or religion, its the ability of scientists to throw out bad ideas. These people are absolutely necessary for science to function. If an idea cannot produce solid statistics from reproducible well controlled experiments it absolutely deserves to be ridiculed. Every idea starts out this way, therefore every idea is ridiculed until someone can set up a good experiment that silences critics.

The case in the article shows how this works quite well. Scientist comes up with idea, idea gets attacked, scientist goes back to the lab with this criticism in mind and produces a device that anyone can test, idea gets accepted.

The biggest issue I see that people uneducated in the ways of science have is that they simply don't understand the standards scientists use. Scientists know that if they want an idea accepted they must set up proper controls, blind or double blind the experiment if possible, produce results that use valid statistics, and write up a procedure that can be duplicated by its critics. The biggest disputes between scientists and non-scientists come from areas where the use of these methods leads to a clear conclusion that the effect does not exist, but not knowing about these methods makes it easy to fall into the trap.

Astrology makes a wonderful case in point. It is really quite easy to use the scientific method to test astrology. Go find astrological data for people, than give people either the correct astrological prediction, or one for another person, and compare the results. If the statistics then showed that people's lives were better predicted by astrology charts made for them, than ones made for strangers than science would have no choice but to conclude there is something there.

Sure enough these experiments have been done. For example one study gave half of people their own astrological reading, and the other half the reading for the birth date of a famous serial killer. They then asked the subjects how well the astrological reading applied to them. Perhaps unsurprisingly, people said that astrology readings were quite accurate. Only they said they were just as accurate when they were given their own reading, or when they had been given the reading for a serial killer. The only reasonable conclusion any scientist would make is that the descriptions astrologers make are so generic that they describe everyone equally well. Scientists have known this for hundreds of years, but the general public just knows too little about science to ever even realize that just because an astrological reading describes them well, does not mean it has anything at all to do with the stars.

Similarly, The phases of the moon could predict how people behave, this is an easily testable hypothesis. Only they don't. It has been well established by several mechanisms such as watching crime rates and comparing them to the phases of the moon, comparing admissions to psychiatric clinics or emergency rooms. Sure enough, no connection. This is not surprising, there is no clear mechanism that the time of day on the moon could alter human behavior. However people are quite poor at statistics without being really careful. Normally we count the hits and ignore the misses. A crazy day with a full moon might get blamed on the moon, but a normal day will not be used as evidence against the hypothesis.

No comments: