Global Warming: The Scientists versus the Conservatives

Global warming is an important but confusing issue for most people because this issue has economic significance, but perhaps not as much statistical significance. Conservatives claim that statistical significance of the relationship between human action and global warming cannot be demonstrated. Much of the scientific establishment opposes this viewpoint. However, the scientific community needs to rethink how it approaches this issue.

The first thing that must come to the mind of any scientist when confronting an issue that is economically significant but statistically insignificant, is the concept of POWER. Lack of power is called “type 2 error” in statistics.

Type 2 error measures the probability of “failure to reject the null hypothesis,” given the alternative hypothesis is true. In the case of global warming, "lack of power" is the error related to the failure to reject the null hypothesis that “no relationship between human action and global warming exists," if the alternative hypothesis that "human action is causing global warming" is true.

Scientists don’t have to prove that human action is causing global warming at 95% level of confidence, even if this can be proven. All they have to do is demonstrate the “lack of power" in the conservative argument. Just because some hypothesis is statistically not significant at 95% level of confidence, does not make that unimportant in an economic sense. And given what is at stake for humanity, the huge economic significance cannot be ignored, even if the effects of global warming are statistically significant only at say 50% level of confidence.

An economically significant event that is statistically insignificant at 95% level of confidence still calls for an effective response. For example, if there is only a 40% chance that one may die after eating food in a restaurant, one would still not eat there, regardless of the fact that the null hypothesis that "the restaurant food does not cause death" cannot be rejected at 95% level of confidence. Economically significant events don’t have to be always statistically significant at 90% or 95% level of confidence for them to deserve an effective response.

The global warming debate is different from debating whether one should devote one's entire life in preparation for not getting hit by lightening. In this case, the probability of being hit by lightening is virtually zero, and so it is unwise to devote significant time and energy to deal with such an occurrence.

I don't think any climate scientist thinks that the probability of the entrapment of greenhouse gases causing severe damage to earth is so small, that no action is warranted. And yet "no action" is the typical recommendation by most conservatives and their think tanks. These conservatives are hiding behind the claim of statistical insignificance, while cleverly evading the type 2 error related to economic significance of the severe damage that can result from entrapment of the greenhouse gases.

The intellectuals are divided as “small selves” on the issue of global warming. The conservative small selves are obsessed with statistical insignificance, while the environmentalist/scientific small selves are obsessed with statistical significance. Both need to come together by focusing on the issue of “lack of power” or the type 2 error to resolve this debate. Just like buying insurance or creating conditions for elimination of terrorism (and not creating more terrorism as being done at present) are right responses for statistically insignificant but economically significant issues, taking concrete steps to significantly reduce greenhouse gases is urgently needed to avoid a potential catastrophe in the making.