There's a point of view that says that the only way that science can make progress is by showing that theories are wrong. The argument goes like this: It's impossible to prove that a theory is right no matter how many experiments agree with it, but if one single experiment disagrees with it, then the theory must be wrong. Well, that itself is a theory of knowledge, which is wrong! Because, there are theories in science, which are so well verified by experience that they become promoted to the status of fact. 
Dr. Walter M. Robinson, associate director of the division of medical ethics at Harvard Medical School, said, "A little knowledge of the history of science tells us that advances are generally not planned and can't be predicted, but come from accidents and serendipitous moments." (New York Times)
However, once the advance is made, the plodding scientific method is very useful for separating the scientific sheep from the fantasy goats.