Robin Neill asks: 
 
> 
>     What, in the History of Economics, corresponds to proof in an 
>empirical social science?  
> 
>     The simple question remains. 
 
The question is not so simple. I would like to turn the question around. 
What, in empirical social science, constitutes a proof? 
        Consider econometrics. By the end of the nineteenth century, 
statistics became identified with inferring conclusions from historical 
data; in the twentieth century, inferential statistics became widely used 
and soon institutionalized as the single method of inference from data to 
hypothesis. With the rise of econometrics, economists became confident that 
the standards of scientific demonstration had now been objectively and 
universally defined. In econometrics, experiment is comparatively rare and 
the standard statistical tool is regression analysis, which is often 
applied without extensive rationalization. For example, the interpretation 
of probability is hardly ever mentioned in econometrics and, if it is 
discussed at all, the relative frequency theory is favored by alluding to 
Fisherian or Neyman-Pearson statistical theory, which is based on the 
notion of repeated sampling or experimentation.  Furthermore, with respect 
to the error term justification, the presence of disturbances is commonly 
rationalized by some mixture of random sampling from a population and an 
epistemic or ontic interpretation of the influence of omitted variables, 
without an analysis of the implications of these different renditions. 
        This is only the start of the trouble. 
 
                                                --Esther-Mirjam Sent 
 
___________________________________________________________________________ 
___ 
Department of Economics         426 Decio Hall          131 S. William 
Street 
University of Notre Dame        (219)631-6979 (O)       South Bend, IN 
46601 
Notre Dame, IN 46556            (219)631-8809 (F)       (219)289-4844