Saturday, April 24, 2010

One Picture Is Worth 2.5 Billion Dollars

"Mystic Mountain" in the Carina Nebula as viewed by the Hubble Telescope.  (We got our money's worth.)

Thursday, April 08, 2010

Tao

I recently read the following anonymous quote in an SAT reading comprehension passage:

There are those who believe that science consists entirely of disproving alternative hypotheses, as if when you eliminate the alternative views, the one you have left is right. The problem is that there is no way to think of all the possible hypotheses that nature can devise. More than that, you have to prove which is the most reasonable. But any hypothesis can, with a limited data set, be reasonable. There is at least a touch of truth in the idea that any variable affects another. If you look long and determinedly enough, you will find that almost any variable element you chose to examine apparently affects the behavior you are studying.

Now, other than helping to explain a lot of the current mania surrounding carbon dioxide's effect on global warming, this passage exhibits a few inadequacies which I believe are reflective of our Western culture:

1) It still allows that one hypothesis (independent variable) alone is most often sufficient to explain the behavior of a dependent variable. In my experience "most reasonable" may mean that this hypothesis explains say only 50% of dependent variable behavior. Another independent variable may explain 45%. Ignoring this second one might be catastrophic.

2) Multivariate analysis seems to be eschewed in Western culture possibly because of its complexity. Witness how we tend to gravitate to a single "silver bullet" approach to many of our conundrums. Whereas we desire taking a single pill to cure a health problem, Chinese herbalists will compound a time-tested mixture of many herbs, minerals, and even animal parts to cure even a simple headache. They seem to understand that variables do indeed interact. Independent variables "A" and "B" might explain 90% of a dependent variables behavior. But "A", "B", and "C" might explain only 40% due to "C"'s deleterious effect on "B". Alternatively, "B"'s effect might be magnified by "C". This is comparable to the effect that a catalyst has on a chemical reaction.

3) There is a poorly understood concept in statistics called "homoscedasticity". This means that some independent variables are equivalent in explaining the behavior of the dependent variable. Think about both cake and ice cream being blamed for the increase in diabetes.

So we see that Western cultures have the unfortunate tendency to propound phantom truths often with devastating consequences ... witness the banning of DDT (sacrificing million of human lives so to protect the population of peregrine falcons), the false belief that thimerosal causes autism (and probably setting back investigations into its real cause), and, of course, the pejorative rhetoric surrounding carbon dioxide (a life-form essential gas). This may be why the Chinese seem to be laughing up their sleeves over the U.N.'s attempt to reduce carbon dioxide emissions.

I know that it seems treasonous, but perhaps we do have something else to learn from Eastern cultures ("the way").