I mean, clearly they [political science, sociology, economics, psychology] can be studied in more systematic ways than art or literature, but something tells me they fall short of the label of science. — rickyk95
The behavioral sciences have taken some deserved hits to their reputations lately. Valid results should be duplicable, for instance. A survey found that many psychology results could not be duplicated. As it happens, a lot of research in the biological sciences (which are "harder" than behavioral science) hasn't been duplicated, either. Good science is hard whether it's physics or economics.
Human beings are one of the main problems in behavioral science. Studies of behavior in rats and pigeons have produced solid, replicable findings about how rat and pigeon' brains work. Many of those findings are applicable to human beings. If we conducted research on human beings in exactly the same way we conduct research on rats and pigeons, we might nail down some answers to difficult questions. That kind of research has (quite properly) been ruled out of bounds.
Even a relatively small ability to think; even limited freedom of will; and that softest and most elusive topic of research -- mind -- all get in the way.
Take for instance research into cooperative behavior in dogs and monkeys. Researchers discovered that both dogs and monkeys pay attention to the rewards their cooperating partners are getting. If some dogs see that they are getting no rewards (while other dogs are), they stop cooperating with the researcher. If monkeys see that their cooperating partners are getting better rewards than they are getting (pieces of apple instead of pieces of cabbage) they stop cooperating. Unexpected end of experiment. What the researchers discovered is that their captive subjects were capable of feeling cheated, and would then not play any more.
It is worth noting that economists generally can not predict economic crashes. Or, at least, that is my understanding. They can't, or they haven't. Lots of people -- economists and kibitzers -- thought that housing prices were absurdly high in 2006-2007, but nobody predicted near economic collapse and a credit freeze in 2008.
Geologists can't really "experiment" with plate tectonics and earthquakes because the subjects -- these big continental and oceanic plates -- are way too big. They can observe them, however, and make predictions. The world economy is a similar problem: it's too big, too many moving parts, and all those parts are always moving.
A lot of psychology research looks like slop to start with, but even if it is very, very good, there is still the problem that human beings can, and regularly do deliberately misrepresent themselves, can fake cooperation, lie, dissemble, refuse to cooperate, and so on -- all of which undermines the validity of psychological research. Any study which involves self-reporting is practically doomed.
Studies of sexual behavior (like the use of condoms, frequency of sexual encounters, preferred sexual activities, etc.) that are based on self-reporting are notoriously unreliable. As it happens, it is quite difficult to observe all of these behaviors. Most people don't like having a note-taker in their bedroom while they hare having sex. Where observation has occurred, the findings are often quite different than self-reporting.