Statistics is hugely valuable in the real world. Simply knowing how to run, and interpret, a regression is invaluable to management consultants. Statistics is now permeating the IT world, as a component of data science — and to do statistics, economists have to learn how to manage data. And statistics forces economists to learn to code, usually in Matlab.
As Econ 101 would tell us, these skills command a large premium. Unless universities want to shrink their economics departments, they have to shell out more money to keep the professors from bolting to consulting and financial firms.
If sociologists want to crack this bastion of economists’ “superiority,” they need to tech up with statistics. Sociologists do use some statistics, but in general it’s just much less rigorous and advanced than in economics. But there is no reason why that has to continue. Sociologists work with many quantitative topics. There are vast amounts of quantitative data available to them — and if there is a shortage, survey research centers such as the University of Michigan’s Institute for Social Research can generate more.
Using more and harder statistics will probably require more quantitative modeling of social phenomena. But it won’t require sociologists to adopt a single one of econ’s optimization models, or embrace any economics concepts. It won’t require giving one inch to the “imperialist” economics of Gary Becker’s disciples. All it will require is for sociologists to learn a lot more advanced statistics, and the data management and coding skills that go with it. The best way to make that happen is to start using a lot more sophisticated statistics in sociology papers. Eventually, the word “sociologist” will start to carry the connotation of “someone who is a whiz with data.” I’m sure some departments have already started to move in this direction.
I imagine this would generate a wide range of responses from sociologists. A few quick thoughts:
1. Using more advanced statistical techniques is one thing but it also involves a lot of interpretation and explanation. This is not just a technical recommendation but also requires links to conceptual and theoretical changes.
2. Can we statistically model the most complex social realities? Would having more and more big data make this possible? Statistics aren’t everything.
3. Any way to quantify this anecdotal argument? I can’t resist asking this…