# Crash course in precision and uncertainty, in advance of that climate debate, free for Mr. Pruitt

July 13, 2017
By

(This article was originally published at Big Data, Plainly Spoken (aka Numbers Rule Your World), and syndicated at StatsBlogs.)

Scott Pruitt, the EPA chief, continues to make innumerate comments about his personal views on climate change science. His chief accusation - chanted often - is that we need more "precision". In his view, achieving 100% precision is necessary because it removes all uncertainty, allowing lawmakers to take action. This post is inspired by his latest interview in which he is encouraging a TV debate event to air out the imprecision of climate science, among other grievances (Link comes with a warning that it is quite exasperating to read through.) I previously commented on his prior statement about precision here.

Mr. Pruitt, it's time to learn some statistics.

In statistics, we make a distinction between precision and uncertainty. Those are distinct concepts. They are related but not in the direction that seems natural. More precision in fact is tied to higher uncertainty, not lower!

One way to see this is to think about the train schedule. One can give you a 100% precise estimate by saying that the train shall arrive in exactly 4 minutes. To be a bit facetious, some people (or software machines) will be more than 100% precise by saying the train shall arrive in exactly 4.01578726787634 minutes. Any idiot will react to that by saying the estimate is stupid, and will be wrong. It is an extremely precise but clearly unreliable estimate.

If one were to say the train shall arrive in between 2 to 6 minutes, there is a good chance that the train shall arrive within that time interval. One component of numbersense is understanding this basic statistical concept: you buy more certainty by sacrificing precision.

***

Regarding that TV debate that Pruitt is proposing, we should practice representational democracy - instead of pretending that there are equal numbers of scientists on each side of the debate, the size of each team should represent the proportion of scientists supporting each side of the debate. If we follow what the politicians did during the Presidential election, then if you are a minor party with low levels of public support, you don't even get to participate!

Please comment on the article here: Big Data, Plainly Spoken (aka Numbers Rule Your World)

Tags: , , , , , ,

 Tweet

Email: