The Enterprise Blog

Charles Murray

It’s the Disappearance of the Data That’s the Most Damning

By Charles Murray

December 1, 2009, 9:28 am

I don’t know anything about global warming but I know a lot about quantitative data analysis. The little secret—not dirty, exactly, but akin to the reasons it’s best not to watch how sausage is made—is the number of judgments that have to occur during the course of data analysis.

It isn’t a sinister process—judgments are inescapable—but the way those judgments are made can importantly affect the results. Example: I spent yesterday working through alternative ways of adjusting for “educational inflation” from 1963–2008. Only 50 percent of white adults had a high school diploma in 1963 versus 92 percent in 2008. If you’re using educational attainment as an independent variable over that period of time, how do you deal with the fact that a high school diploma meant something very different in 1963 than in 2008? Dozens of such judgments lie beneath any complicated social policy analysis you read.

That’s why a good social scientist typically runs analyses using different codings of variables to see just how sensitive the results are—and why, above all, a good social scientist must be prepared to ship the database to anyone who wants to go back to the original numbers.

That brings me to Climategate. The thousands of temperature measurements used to prove long-term warming cannot be treated as-is (“60 degrees Fahrenheit at 6:30 AM, 15 May, 1895, Cotswold station”). That “60” has to be treated in the context of time, date, location, local effects on the background temperature—and on and on—when it is analyzed.

The people who made those adjustments are, we now know, desperately invested in proving the truth of man-made global warming. And they lost the data. That’s more damning than anything else in the emails.  If you’re doing important work that you know will be controversial, you don’t lose the data. You document everything you did to the data. You make the data available to others. If you don’t do all of those things, people are right to ignore anything you have published about the data. And that’s what we should do with everything these men have published about man-made global warming.

Comments are closed.