This is an excellent point. I think we're in a bind here, because I believe that the scientific method is still appropriate to the "social sciences", but I also believe that human nature and the media will ensure that all carefully caveat'd theories will find their way into the public discourse sans caveats.
Why is the scientific method appropriate to the "social sciences"? Because it's way better than the alternative (which I'm unclear on, but I assume is just a more descriptive role rather than an explanatory one). In the social realms, there are causes and effects which are important for us to try to understand. However they are difficult and often impossible to understand perfectly. The scientific approach at least constrains somewhat those who study these phenomena. Sometimes the scientific method is even useful! The uncontrolled experiments that are the meat of the social disciplines are increasingly being complemented with controlled experiments modeling components of the system. That way lies insight if not absolute truth.
(I come from the business world, and my analogy is marketing. Marketing is messy; there are few controlled experiments; there is much room for attractive BS to gain sway in a discussion. However, a scientific approach yields great results, if one appreciates the limits and the need to complement it with plenty of creativity. However, no one ever seems interested in the marketing analogy :-P)
I think we have to accept that with the murkiness of the questions and the data in the social sciences, attractive but incorrect theories will simply take longer to debunk than would be the case in hard science. And remember, even there it took decades for plate tectonics to win out.
