Log in

No account? Create an account
Thoughts Online Magazine
Collected Articles on Culture & Politics
Global Foolishness 
2nd-Feb-2006 05:04 am
And, from today's Wall Street Journal ( subscription required), an illustration of what happens when sloppy science meets bad journalism:

To calculate the total power generated over a storm's lifetime, Dr. Emanuel multiplied each hurricane's maximum sustained wind speed by itself and then multiplied that result by the wind speed again, a calculation known as cubing. Then he factored in how many hours the storm lasted.

Dr. Emanuel says he used scientifically accepted formulas to adjust for years when wind-speed data are most likely to contain errors, particularly in Atlantic storms from 1949 to 1969, when it is thought speed was overestimated. The calculation showed that the intensity of storms had essentially doubled in the past 30 years. He attributed growing hurricane intensity and destructive power to rising water temperatures that he said were "at least partially" the result of human activity.

The marginal notes: "known as cubing" indicates the reporter didn't pass junior high math classes. Raising the wind speed exponentially also increases the "variation" that most statistical tests will see: the spread is bigger between storms of similar speed, e. g. 130 MPH is suddenly much different from 150 MPH, and the variation looks significant. This "doubling of intensity" is at least partly, if not wholly, an artifact of the method he's using to measure it. Sadly, global warming enthusiasts will seize on the conclusion without doing any checking of the basic science, and assert that "global warming has worsened hurricanes."

Meanwhile, I really wish scientists who published these "studies" would learn statistics.

UPDATE: As it turns out, even when they argue with each other, they still don't know how to do statistics:

``The methodology is fine. There's no problem with the way they analyzed the data,'' said Landsea, who is science and operations officer at the hurricane center. ``The problem is with the data itself.''

As pointed out above, the methodology is flawed in itself. As noted above, statisticians who review climate science can be appalled at commonly accepted techniques. This would be one. Again, where were the math classes that these people took?
This page was loaded Apr 19th 2018, 5:01 pm GMT.