Empirical evidence wins out in the end but… it’s not that simple. One name said a lot about this : Thomas Kuhn. Try giving The structure of scientific revolutions a read whenever you can. It’s old and there are more contemporary work, but Kuhn is still a reference in epistemology.