http://www.latimes.com/news/opinion/commentary/la-oe-wilson-social-sciences-20120712,0,4672619.story
As far as defenses against 'hard science' bullies go, this one's pretty soft. The idea that social sciences are inferior because we can't run randomized experiments is ridiculous. And it's not just ridiculous because some social scientists do run randomized experiments with successful results. It's mostly ridiculous because randomized experiments would be generally an inferior way of going about research on the social world.
Biological and physical sciences take as their object of study units of analysis that have no agency. And by that I mean that they create no meaning. Cells, organs, atoms, molecules, elements... these things are what scientists say they are. They mean what scientists say they mean. So of course it's possible to run repeatable, randomized experiments. They are examining a physical reality which they have defined in the first place.
Attempting to run experiments on social actors is inherently problematic precisely because social actors are in the business of making meaning. They interpret the world around them according to culturally prescribed logics, vocabularies, rules, categories, and experiences. To try to "repeat" an experiment on a social actor would require holding an infinite number of variables constant. And further, to 'repeat' an experiment on the same actor is literally impossible because that actor has been fundamentally transformed by the initial experimental attempt.
Social actors can also sustain qualitatively different relationships to their involvement in experimental activity. They can engage a situation ironically, as a joke, as a test, as a competition, as a performance, as a rehearsal, and so on. To try to control these "frames" of meaning through which actors understand their engagements would be impossible. The post-experimental questionnaire is rendered useless if one party engaged the experiment as a race and the other as a test.
So if physical and natural scientists find their work more valuable because they can perform repeatable experiments, they shouldn't strain to pat themselves on the back too much. It's not because they have more rigorous scientific procedures, but because they have objects of study which are without the agency to find and make meaning. And in that way, their task is simpler. They study units of the universe which they themselves define, we study the processes through which social actors give their worlds meaning. And those meanings are always temporary, contingent, and negotiated.
If we can't predict where the next rampage shooting will be, it's not because our procedures our poor. Rampage shootings are not so passive as to allow themselves to be predicted. Prediction is simply the wrong goal. Good social science yields understanding, empathy, and empirical knowledge of how people live meaningful lives. To stay true to our field of inquiry is to remember that we study subjects, not objects.
