Since the mid-1970s, most social scientists treat the covering law model
and the methodological prescriptions of testing hypotheses by predictions
as metaphors only and not as rigorous requirements. Social scientists deal
with systems that contain a lot of ‘disorder’ (also called ‘chaos’ and ‘noise’).
As a consequence, the best they can do is to search for ‘underlying mechanisms’
capable of explaining the emergence of phenomena. Any predictions
they make will not provide decisive tests of their theories.