OK, two posts for today, because I can’t help myself. Yeah, I am a social scientist. Which means that people either think I run control experiments on various populations (an idea that freaks me out)*, or they think that I have no method to my research at all – I just sort of run around, talk to a few people until I get bored or run out of money, and then come back and write it up.
Of course, both views are crap.  Good social science is founded on rigorous fieldwork and data whose validity can be verified.  How one collects that data, and verifies that validity, varies – it depends on what you are studying.  For whatever reason, though, people have a hard time understanding this.  Quick story: a former chair of my department, during a debate about field methods, actually once asked me if it was really possible to teach someone to do interviews and participant observation.  My response: “I didn’t pop out of the womb able to do this, you know.”  End of discussion, thankfully.
But now I have found someone who has written this up nicely – Wronging Rights (absolutely hilarious, and totally awful, all at the same time – just go read for a bit and then feel bad about yourself for laughing.  Everyone does) has a great post on the subject that links to a series of even better posts at Texas in Africa that covers it (see the Wronging Rights post link to connect to the relevant Texas in Africa posts).
Social scientists, get to reading.  Journalists, read this and understand why you are not social scientists.  Especially you, Thomas Friedman.  And the rest of you . . . never, ever ask me if you can teach someone to do social science . . .
*controlled experiment: what, am I supposed to pick two identical villages (no such thing), and then start to work with one village while studiously ignoring the other village no matter what happens to that community (i.e. drought, food insecurity, disease, what have you) because I need to preserve the integrity of my control group?  There are other ways to establish the validity of one’s results . . .