Sunday, April 1, 2007

The importance of outliers

In almost any natural distribution there is a “bell curve” - a central pile around the average and then a few outliers dribbling out at the high and low ends of the distribution - a few data points that are way out away from the great grouping around the average. For example, measure the height of 10,000 adult American males and most will pile up in the 5-6 foot range. But a small number will fall neared 4’6”, and an even smaller number will fall at 4’ or less. On the other end, there will be a small number at 6”6”, and a few out at 7’ or above. These are the outliers in the distribution.

I was thinking about this the other night when I read about a couple of scientists prominent in their own fields of study, though not in climate studies, who still don’t believe there is a global warming issue despite all the evidence. My first thought, of course, was to dismiss them on the grounds that there are always a few people at the fringes, no matter what the issue. I suppose if one looked hard enough one could still find people who believe the earth is flat, or that the earth is at the center of the universe and the sun and all the stars revolve around it.

But then I realized that outliers are important. Most of the time they are just plain wrong, and off the wall. But every now and then they are the voice of sanity when the rest of the crowd has gone off in the wrong direction.

In nature, mutations are always producing a few plants or animals that differ from the average in one way or another. Most of these mutations are not of any use, and many are in fact fatal. But every now and then they save the species. For example, some birds have timed their mating and hatching to occur just when the population of some species of insect is at its peak, so that there is a plentiful supply of food. A few birds every year get it wrong and their young hatch too early or too late, with a much lower chance of survival. But now that climates are warming the insect populations are peaking earlier, and the few outliers among the birds who hatch their young earlier are the salvation of that bird species, most of whom are now hatching their young too late for survival.

One of the dangerous things we have done with our food supply is to rely on monocultures – a few genetically uniform species of corn or wheat or rice planted throughout the country, even throughout the world. In the name of efficiency and high production, we have weeded out the outliers that might otherwise have saved our food supply when the next blight comes along.

So in the world of ideas I have developed an appreciation for the opinions of outliers, of people with unconventional, even unpopular, views. I usually don’t agree with them, and in fact many are off the wall and just plain wrong. But still, outliers of any sort are an important natural safety mechanism among humans as well as in nature, and we would be foolish indeed to eliminate them.