Friday, November 5, 2010

On quantitative methods and best practices

Following Texas In Africa's weeklong discussion of what makes social scientists scientists and not advocates (highly recommended), Wronging Rights responds with why lawyers are not social scientists. And it's a fairly compelling case that doesn't rely on denigrating social science:
As Laura explains, social scientists are primarily concerned with using evidence to explain why certain events occurred and to predict future events. Lawyers use evidence to prove that certain events occurred and sometimes to prescribe or proscribe future action.

Consequently, lawyers (and those who employ their methodology) and social scientists are interested in patterns in evidence for very different reasons. When human rights researchers collect evidence, they are trying to provide enough support to legitimize a claim that serious violations are occurring.... The evidence-gathering methods of the human rights movement are extremely well suited to the strategy they were designed to support: advocacy for the cessation of violations of human rights laws.
Blattman (one, two) on the problem of using matching as an identification scheme. Basically, if you're matching on observable X, you aren't solving the identification problem any better than with regressions. What you are doing is reweighting the data: regression puts more weight on outliers and data points that are less well explained; matching puts more weight on similar data points and removes points that maybe shouldn't be compared. But that doesn't get rid of selection issues or unobserved variable bias.
The cardinal sin of political science dissertations? Jumping to a case and data collection before you have a research design. You should have your research design, tests for your assumptions, and a complete write-up plan before you ever collect a single variable. One of the great things about randomized control trials is not that they eliminate many sources of bias, but they force you to follow the scientific method (to a degree).
Yale's James  Scott against a life focused on peer reviewed articles. Relatedly, a call for students who want to be challenged to push back on professors who aren't challenging them.

No comments:

Post a Comment