My quandary shows how authority and science are used to influence the political process.
We expect scientists (and pollsters can be considered scientists because they’re applied mathematicians) to be both correct and savvy. So, when I read that the morning sample was based upon a sex ratio of roughly 60:40 (female/male), I automatically figured that whomever did the polling simply adjusted the results to the expected sex ratio of the electorate, multiplying each male vote by approximately 1.1 and each female one by 0.9 (assuming a 50–50 electorate). In other words, I trusted the “experts” to know what they were doing, and to report it to the public in a clear manner. Most other people do the same thing. For instance, when they hear about some forecast of dire climatic changes made at some prestigious university or federal laboratory, they assume that the people who put the thing together are so smart that they would have compensated for any systematic problems with their methods.
Here’s an example from global warming: As a matter of convention, most of our computer models for our climatic future assume that carbon dioxide — the main global warming gas — is increasing at a rate of 1 percent per year. The concentration of carbon dioxide in today’s atmosphere is roughly 375 parts per million (ppm). An increase of 1 percent in a computer model for next year’s climate would raise that concentration to 378.8 ppm, and to 382.5 the following year.
But in reality, that’s not what’s happening. In the last three decades, the percent change per year has averaged 0.39, 0.41, and 0.51 per cent, respectively. What the computer does is more than double the rate of increase that is actually occurring.
The amount of warming produced by those models is directly proportional to the rate of carbon dioxide increase. In other words, the models are compelled to calculate twice as much carbon dioxide‐related warming as could possibly occur in coming decades.
Back during the Clinton presidency, climate scientists produced a so‐called “national assessment” of the effects of climate warming for the 21st century. They used models that did exactly what I describe above, and the report showed dramatic (and false) results. The Bush administration has since used the false material from the earlier report for its own study of climate change. The Bush document, in turn, served as the basis for climate change legislation by John McCain (R‐Ariz.) limiting our net emissions of carbon dioxide. This can’t be accomplished without actively discouraging energy consumption, i.e., dramatically raising the price of gas.
McCain is a consummate political animal, positioning himself for a 2008 presidential run. He authored that legislation for one simple reason: He sees political advantage in claiming to care about global warming. After all, most of his Republican competitors are going to be on the other side, against regulation.
Politically, it is profoundly easy to demagogue any climate anomaly into global warming. Remember September’s hurricanes? A coalition of scientists — “Scientists and Engineers for Change” — exploited those disasters by plastering central Florida with billboards claiming that re‐electing President Bush would make hurricanes worse because he’s not doing enough about global warming. Their scientific basis? The same computer models used by McCain, with the wrong increase in carbon dioxide.
In this case, they failed. They were not able to persuade enough voters to eke Florida into the Kerry column. And, for that matter, neither did the exit polls produce a large enough effect to turn the nation.
Matt Drudge is a sharp guy. So are the people at Slate Magazine, who kept the erroneous exit poll numbers up all afternoon. (Drudge took them down for a while.) The poll results were leaked with the full knowledge of their political effect — just as scientists know that a computer model that must overestimate global warming will also stir things up.