The empirical sciences tend not to be in the business of proving. Rather, they participate in the corroboration and falsification of hypotheses by means of observation. This is common to both the social and the natural sciences. Of course, one might in the social sciences consider a collection of opinions to represent an observational sample, and then analyze this data. But the methods of analysis certainly do not revolve around further opinioning. Rather, its mostly really basic statistics, and in some cases you get social scientists doing something a bit more interesting, like quasi-experimental designs.