Abstract:This paper contributes to the debate about the role of realism and instrumentalism in social science. It asserts that the debate should distinguish between ex-ante and ex-post perspectives. Ex-post, i.e. once the research is done, instrumentalism might be given priority because in this situation, signifi cant predictive power is the key. On the other hand, ex-ante, i.e. before any model revealed its predictive power, only realism may give a hint of which direction is promising to pay off the effort. Thus both realism and instrumentalism play their role in the scientifi c activity, depending on the stage of the research. In this paper I argue that in order to develop economic realism, we should have recourse to phenomenology. Phenomenology analyses the thought constructs of ordinary people, classifi es them and seeks the underlying invariant structure. Realistic social science should build on these invariant structures; only then the link to social reality is not lost.