Tuesday, March 15, 2016

Measurement is not the way to make a science quantitative

"The full and intimate quantification of any science is a consummation devoutly to be wished. Nevertheless, it is not a consummation that can be effectively sought by measuring." -- Thomas Kuhn, "The Function of Measurement in Modern Physical Science"

What Kuhn is getting at here is that the ability to make meaningful measurements is the end product of the quantification of a science, and not the road to its quantification.

Social scientists would do well to remember this point, especially when they try things like "measuring" happiness by asking a bunch of people to rate their own happiness on a scale of 1 to 10. They have no theory of what, exactly, they are measuring, of how their "device" purports to measure it, or as to what sort of quantitative relationships they expect it to have with other measurements.

Physical scientists took over a century of experimenting with thermometers to figure out what they were measuring and how it would relate to, say, the volume of a gas. Social scientists have to often jumped straight to slapping numbers on things and then running regressions on "the numbers."

1 comment:

  1. Indeed. Excellent point. Every day I see people treating ordinal data like interval data, and claiming what the numbers prove!



"If your approach to mathematics is mechanical not mystical, you're not going to go anywhere." -- Nassim Nicholas Taleb