Theory gets too Much Respect, and Measurement Doesn't get Enough (60-Second Lecture Video and Transcription)

In the last post, I told you about Penn's "60-Second Lecture". Mine is now completed, and we had a good time. Watch the video, and you'll have a good laugh at the unflattering opening shot, complete with a barking dog in the background, and the blinding sun in my face throughout. A rough transcription follows.

Science is advanced by just two things, measurement and theory. Their interplay pushes science forward, as each disciplines the other.

Some people believe that good research requires tightly-integrated measurement and theory, present in equal amounts.

I submit to you, first, that such views are both naive and false. Measurement and theory are rarely advanced at the same time, by the same team, in the same work. And they don't need to be. Instead we exploit the division of labor, as we should. Measurement can advance significantly with little theory, and theory can advance significantly with little measurement. Still each disciplines the other in the long run, and science advances.

And I submit to you, second, and primarily, and perhaps provocatively, that theory gets too much respect in science, and that measurement doesn't get enough. A wry observer once remarked that theorists typically have the top-floor offices, while experimentalists and statisticians are tucked away in the basement. But Lord Kelvin got it right more than a century ago, when he argued that measurement is the essence of science. And moreover, theory is largely data distillation, attempting to eliminate noise and isolate signal, and hence dependent on measurement. The theory mill grinds data.

Of course it's futile, and it's not at all my intent, to assert that one or the other of measurement and theory is somehow uniquely important, or even relatively more important. Each is equally important, and each most definitely needs the other. But again, that's the point: given their truly equal importance, they should get equal attention and respect. Instead theory gets too much, and measurement doesn't get enough. Perhaps that will change in the emerging age of Big Data.