Discrete choice experiments (DCEs) are being used increasingly in health economics to elicit preferences for products and programs. The results of such experiments have been used to calculate measures of welfare or more specifically, respondents' ‘willingness to pay’ (WTP) for products and programs and their ‘marginal willingness to pay’ (MWTP) for the attributes that make up such products and programs. In this note we show that the methods currently used to derive measures of welfare from DCEs in the health economics literature are not consistent with random utility theory (RUT), or with microeconomic welfare theory more generally. The inconsistency with welfare theory is an important limitation on the use of such WTP estimates in cost–benefit analyses. We describe an alternative method of deriving measures of welfare (compensating variation) from DCEs that is consistent with RUT and is derived using welfare theory. We demonstrate its use in an empirical application to derive the WTP for asthma medication and compare it to the results elicited from the method currently used in the health economics literature. Copyright © 2004 John Wiley & Sons, Ltd.