8 
purpose. For some purposes, the whole apparatus of a large 
government machine must he moved, which is understandably 
difficult and time using. Such efforts, moreover, are not the 
easiest way to reputation. Most economists avoid it and work 
, with deductive methods or live on such data as come their way. 
Much of modern econometrics is an attempt to compensate for 
the lack of data by increased refinement of the methods of pro 
cessing them. The question is hardly ever asked whether this is 
possible: The data - in most cases short aggregated time series 
which are highly autocorrelated - cannot yield more information 
than they contain and that is usually very little.lt can be 
said - and I do not think I am too prejudiced here - that the 
dissatisfaction with formalism, deductive methods and threshing 
of empty straw (poor information) has become fairly wide spread. 
Sociology is suffering from a division between empirical methods 
and speculative or philosophical methods. I take it that each of 
t 
these positions is untenable . In practice, sociology is suf 
fering from the lack of recognition of its usefulness, which is 
not helped by the above division. In fact, here as in the case 
of economics, one of the keys to the development of the field 
is in the kind of education given by the universities. 
had 
Operational Research is the discipline which has A the greatest 
practical impact, which has the closest contact with scientists 
and engineers, and which has in some cases a good empirical