I don't think there is another way of proceeding other than what is being called 'empiricism' here.Outside of math and logic, you can only look at the facts or evidence and try and draw inductive inferences from them building up to explanation of causal patterns. You need a certain amount of empirical evidence before you can even formulate a hypothesis let alone test it. All theories to some extent must be based on empirical observation. Even if one believes in knowledge a priori, such knowledge would only account for a miniscule amount of what we do or can know. It is a mistake to counterpose theory and facts since all observation of the facts depends on already assimilated theory. The famous example by N.R. Hanson was an x-ray. When I look at an x-ray I see gray and white bloches, a doctor looks at an x-ray and sees a fractured tibula. Same with data and causal patterns in the world. Much of "theory" in the social sciences is not theory in the same sense that evolution by natural selection is a theory because you cannot predict anything from "theories" in the social sciences-- its just too complicated with too many variables. The best one can do is ex post causal explanation. Further, theories in social science will always be underdetermined i.e. multiple explanations are true of the same hypothesis. Much of social sciences consists of starting with your conclusion and working backwards trying to get the 'facts' to fit into your theory. Especially in economics, theorists start with what they are trying to prove and then go to work. The political conclusions are drawn at the outset. Sam Pawlett