May 17, 2010

The Loneliest Number

A few weeks back, upon receiving my ever-treasured Wired in the mail, I read and fell hopelessly in love with a piece by Clive Thompson, “Why We Should Learn the Language of Data.” It is a call for the importance of understanding the barrage of statistics we hear daily, and ends in a beautiful line: “Statistics is the new grammar.”

Writing about statistics seems to be an increasingly popular nerd sport. Yesterday’s New York Times magazine included another thought-provoking piece on the numbers we choose to represent non-numerical human phenomena. It pleas:

Unless we know how things are counted, we don’t know if it’s wise to count on the numbers. The problem isn’t with statistical tests themselves but with what we do before and after we run them. First, we count if we can, but counting depends a great deal on previous assumptions about categorization. Consider, for example, the number of homeless people in Philadelphia…Is someone homeless if he’s unemployed and living with his brother’s family temporarily?

This does not contradict Thompson’s entreaty to better understand statistics; both articles boil down to the public need to identify BS among the information provided to you. That was the also the gist of the extensive statistics education I was forced to endure long, long ago at Ohio State as a sophomore undergrad. We endlessly played with data sets and statistical methods in order to produce whatever preconceived results that we desired, in order to prove to us that we were all capable of making statistics lie in our favor. The message stuck with me when we put together statistical evidence that Republicans were 12% happier than Democrats. But beyond just political science-y questions, Thompson issues this warning (bolding is mine for emphasis):

Statistics is hard. But that’s not just an issue of individual understanding; it’s also becoming one of the nation’s biggest political problems. We live in a world where the thorniest policy issues increasingly boil down to arguments over what the data mean. If you don’t understand statistics, you don’t know what’s going on — and you can’t tell when you’re being lied to.

This centrality of data is certainly the case in natural security research. Thompson specifically cites debates over climate change, but this hits our work especially hard in looking at energy and minerals, info on which often revolves around proven reserves, demand and price projections to indicate who has what and for how long and what that means for the United States. 

About a week back, I was in a meeting explaining to a wise man my plan to do a big – freaking huge, really – large-N study that will involve examining minerals prices since, oh, back for as long as I can get consistent data, which I estimate so far to be World War II or so. This gentleman argued that prices don’t indicate anything in the minerals world, and he stated that traders trade large quantities at single points in time, and often speculating on prices, and that this creates major distortions in minerals price data. I could not agree more, but what I will be examining is on decadal time scales, not getting into the details of daily, individual trades.

So the question is: will price data over time indicate anything? My assumption is yes, or else I would not be subjecting myself to this exercise. I am assuming that regardless of erratic daily trading, that prices over time will in a rough sense (and in the best measurable way possible) indicate underlying meaning – that prices will indeed show patterns of demand, supply, and major adjustments in human use. There is a lot of chatter about minerals and security/foreign policy right now, and not a lot of hard examination of historical trends and what they mean. I’m hoping that statistics will help in this case, though I take the NYT’s warning that the truth is obviously broader than what the numbers will show. That’s what travel is for – conducting on-the-ground case studies, which are irreplaceable in our line of work. We pledge not to get too myopic in considering what the statistics indicate as we move forward.

For our D.C. readers, I see online that there is a 100% chance of rain today, and that it is indeed raining outside. Stay dry out there, and be sure to tune into the blog for Yemen week, as resources issues are a major component of that country’s future. Happy Monday everyone.

 

This Week's Events

Tuesday, check out Brookings for a talk over Energy and Climate Change 2010: Back to the Future, beginning at 9:00am, featuring a keynote from Todd Stern, Special Climate Envoy at the State Department. Later that day at 1:00pm, CSIS will feature a series of discussions covering the global resource trade.

Wednesday you should think about dropping by Rayburn to catch a full committee hearing with the House Committee on Science and Technology over the Department of Energy’s nuclear energy research and development roadmap.

Thursday morning, at 9:30am, the Senate Committee on Energy and Natural Resources will hold a full committee hearing to receive testimony over renewable energy prospects in the California Desert Conservation Area.

If you want to know almost any angle of the effects that the Gulf oil spill could have, this is the week for you. Here is a list of Senate committee hearings covering the topic this week: