Thursday, September 25, 2003
I’ve always been of this opinion (despite being trained in the “hard” sciences, which usually have a very low opinion of such things)
It’s art out of necessity. You can’t really perform experiments. Experiments are either simulations or do not reflect the macroscopic scope of a real political system. You can collect data. But data collection is often poor and data only can be collected in certain conditions. You can’t collect data in a counter ruled by a totalitarian dictator. Furthermore, the data reflects to total aggregate composition of all factors. It’s extremely difficult to isolate factors.
Thus we are left with high level pattern recognition. It’s the only practical resource with which to base policy.
P.S. I majored in engineering.
P.P.S. The course I took was in representative government. We studied the various different effects representative systems have, mainly parliamentary systems vs. the two party system. It’s quite interesting since you see that the way an election comes out is heavilly dependent on whether there is a prisoner’s dilemma in the election system. In other words an election isn’t identical to the will of the people. Thus an election where millions of votes are cast won by a few hundred votes is statistically a tie. Particulary when minor differences in the election procedures could throw votes either way.
All true. Social science simply has more variables than the natural sciences. But science isn’t about experimentation per se; that’s merely one method of inquiry among many. Nor is it necessarily even about prediction.
For example, in biology, they have a pretty well developed body of literature on evolution. We know that species adapt to their environments over time, etc., etc. But nobody can predict what adaptations will occur and when.
Make a one-time donation