Social sciences (like economics) use mathematics as a powerful tool to describe events in social reality; it became a modus operandi of a significant part of the modern mainstream economics. But is the use of those formal methods in various approaches of modern mainstream economics justified? It seems that despite the implications of the use of probability of economics, no one seems to be worried if the applications are correct. The use of formal methods (and also theory of probability) presupposes certain ontology. Econometrics and DSGE modeling is a part of modern economics that is utilizing random variables to model events in social reality. Thus, since they use theory of probability, all DSGE and econometric models also presuppose specific ontology. It is questionable whether the ontology that is presumed by every single use of random variables is the same ontology (or at least similar to some point) as the ontology that underlies the social reality. Since there is suspected discrepancy between ontology and epistemology of the theory of probability in economics, I think it is worth examining. I would like to discuss the limitations, implications and also conditions of the correct use of stochastic variables in economics. My aim is also to provide some comments on the use of theory of probability in economics.