Climate Change Science: Type I and Type II Errors
What if I am wrong? Those five words sum up the true beauty of science.
| Climate Resources Greenhouse Gases General Circulation Models Reliability of General Circulation Models |
In technical terms, the double check step is know as error analysis.
Error analysis grows out of the fact that scientific experimentation, broken down into its constituent parts, is really the language of statistical probabilities. In the course of any scientific experiment, the scientist is not trying to prove the correctness of his hypothesis, he's trying to reduce the probability that his hypothesis is incorrect. Confidence is the language of statistical analysis, and the most aggressive language a scientist uses to explain his experimental results is that he's confident within a certain probability, say 98%, that his hypothesis is still reasonable.
Error analysis steps in at this point. Regardless of whether a scientist disproves or fails to disprove an hypothesis, there's always that margin for error. Scientists consider error in two distinct ways.- TYPE I errors are those where scientists assume a relationship exists (in the variables they are studying) where none exists.
- TYPE II errors are those where scientists assume no relationship exists when in fact it does.
Which is the more egregious error, Type I or Type II? There is no easy answer. It depends on the nature of the experiment.
Climate change theory basically reduces to the hypothesis that anthropogenic (human induced) emissions of certain gases called greenhouse gases (GHGs) causes the heating of the atmosphere, and consequently changes in global climate patterns.
A Type I error in climate change means scientists assume that a relationship between anthropogenic GHG emissions and climate change exists, when in fact it is not true.
A Type II error in climate change thinking means scientists assume no relationship between anthropogenic GHG emissions and climate change exists, when in fact there is a relationship.
Which is worse, thinking there's a relationship between anthropogenic GHG emissions and climate change, when no relationship exists, or rejecting the idea that a relationship between the two exists, when in fact such a relationship does exist?
A climate change advocate might suggest that if you are going to be wrong on climate change, Type I errors would be preferable to Type II errors.
Consider the following line of reasoning. A Type I error means, contrary to the hypothesis, anthropogenic GHG emissions are not the primary factor responsible for increasing temperatures around the world. Instead, natural causes such as solar flares or volcanic eruptions provide either better or additional explanatory power for documented changes in global temperatures.
States implementing policies based on results from the climate science community, whose results are based on Type I errors, would, at worst, be wasting their time and money trying to solve the problem. At the very least, GHG reduction policies would help improve overall air quality.
Type II errors, on the other hand, are the least preferred type of error to commit because they turn the logic around. If climate change skeptics (those scientists and policy advocates who attribute current weather patterns to natural factors) are wrong, the worst possible outcome for policies aimed at maintaining status quo GHG emissions levels, risks worsening an already problematic situation.
Recent research on the Arctic Ocean, for example, predicts that it will be ice free in two generations. Polar bears and other Arctic wildlife face threats of extinction by the changes (Santa would also be evicted).
While there is still debate in the scientific community over the exact relationship between climate change and the increasing number and intensity of hurricanes in the Atlantic Ocean and Gulf Coast areas of the United States, research indicates that continued GHG emissions will increase water temperatures and the probability of more hurricanes.
Climate change research has reached the point where scientists now claim a high degree of confidence in their results. Despite the confidence, they still ask themselves, what if I'm wrong? We'd all do well to follow their lead and consider the same question.
© 2000-2007. Patricia A. Michaels
