Science journalism excoriated

Via Clifford: Ben Goldacre writes a mostly great Guardian column about the sorry state of science journalism.

Almost everything he complains about in Britain also goes for America, though I'm not familiar with the quaint subgenre of "scientists have found the formula for..." stories, which must be a British thing. American papers would have a deep horror of printing the formula for anything, no matter how flippant.

The core problem he describes is that it's extremely difficult to convey a quantitative notion of uncertainty if the audience is assumed to know nothing of mathematics, and the idea of science as the self-correcting pursuit of qualified knowledge gets largely lost as a result. He also touches on the problem of bad organizational press releases, which particularly infuriate me: if the PR office of the scientist's own institution can't get the description of the work right, what chance does an outside journalist have?

I do think Goldacre's attack on the humanities toward the end is unnecessarily smug, based in part on a parody vision of his own, and the knock on "cultural relativism" is mistargeted. Cultural relativism is, as far as I know, a term of art in anthropology, and its introduction in the mid-20th century by Franz Boas and others was a genuine step forward that, at least initially, was not an antiscientific reaction at all but an attempt to work around observer biases when figuring out somebody else's culture. That sterile forms of extreme ontological relativism were extrapolated from it later and exploited in romantic attacks on science shouldn't blind us to that.

Math and computers

One of the statements I used to hear during the 1970s and 1980s that would most reliably reduce me to agitated ranting was "I could never learn to use a computer; I'm no good at math." As far as I could tell, not even programming a computer required much in the way of mathematical skill. For me, the big mathematical stumbling block in elementary school had always been arithmetic, and the whole point of computing machinery is that it does that for you. As "computer literacy" courses always emphasized at the time, the concept of a program is familiar to anyone who has ever written or followed a formalized list of directions, such as a recipe or a knitting pattern. And, of course, most users of computers didn't actually program them, something that is even more true today.

But these days I find myself using my mathematical chops all the time. The important skills are mostly the ones taught in high-school and introductory college-level algebra and calculus classes: the ability to do simple algebraic derivations, to visualize functions, estimate quantities and think about extreme cases, and some matrix algebra.

There's not much call for the really advanced material I got, the point-set topology and differential equations and such. But every so often, combinatorics and graph theory and orthogonal functions make an appearance. My proudest moment as a grad-school survivor was the day I actually applied the concept of a differential form in my work (only to find years later that I'd just rediscovered something that was already written up in Graphics Gems).

Also, doing quantum mechanics got me the level of comfort with linear operators and matrix algebra that comes only from years of regular practice.

So I'd say this: you don't need to be good at math to program a computer. But if you are good at math, that is an extremely useful and marketable skill that will set you apart from the crowd of coders.

(In case you're wondering why I'm going on about this, it was because I was reading this Crooked Timber discussion of education and credentialism and thinking about how my educational background has in some ways helped me more than that of the people who actually majored in what I do. It's not that the skills they learned aren't useful; they know some important things that I don't. But their skills are also more readily available in the market precisely because of their presence in the computer-science curriculum. It was a complete accident.)