Generally used by economists to refer to behaviour that is consistent with the axioms of utility or expected utility theory. This idea of rationality has been pretty well attacked. An early assault was from Keynes on the ground that its axioms were not obeyed by at least some financially successful people who ought not to be dubbed 'irrational' in light of the evidence. It has been attacked also because it is too narrow in excluding the emotional effects of not knowing things ahead of time, effects such as anxiety, disappointment and regret, because it is too demanding, and because its literal pursuit might actually reduce welfare (see Bounded Rationality). Moreover, identifying rationality as consistency with the axioms of a theory invites the unhelpful conclusion, every time any of these is actually violated (for example, in a controlled experiment), not that the theory is perhaps wrong but that the human subject in question is an irrational being.
It is useful to keep in mind the distinction between positive and nor- mative economics. In the former, there is no need to use the word 'rational' at all. The axioms of utility theory say it all and one seeks only to account for or predict what actually happens, not what ought to happen if people behaved consistently with any concept of rationality. To the extent that their behaviour flouts the axioms, then the theory is empirically the weaker for it, since any empirical refutation of an axiom is a refutation of the theory (whether one calls it 'rationality' or anything else). Frequent flouting is, of course, of more concern than occasional (and possibly predictable) flouting. Other criticisms of the economic concept of 'rationality' are simply untrue. For example, although it is a popular notion that 'rationality' entails selfishness, this is actually not usually a characteristic assumed in economics nor is it embodied in the axioms.
In the case of normative economics, the issues are more compre hensive and complex, for in normative analysis, one attaches welfare significance to 'rational' behaviour. If the basic axioms of the (positive) behavioural theory are wrong, then it seems hard to conclude that anything normative can be derived from them. But even if they are right, it may well be (as in extra-welfarism) that one needs more than mere rational choice making to identify what is in the best interest of society. Quite apart from the fact that subjective preferences hardly amount to a unique basis on which to evaluate social welfare, some individuals may be 'rationally' (i.e. consistently with the axioms) making perfectly ghastly, even evil, choices. If 'rationality' must be used at all, it is probably best, first, to be very clear about the meaning that attaches to it and, second, to use it in such a fashion as to include what the clients on behalf of whom the analysis is being done find ethically acceptable. There are numerous economic studies of behaviour that is apparently rational - but turns out to be inconsistent with the economist's model of rational behaviour, including the risk-laden consumption of hazardous goods that entail a raised probability of death.