Adam Oliver, LSE Health and Social Care, London School of Economics and Political Science ISBN [07530 1932 9] Discussion Paper 2, February 2002. Health care resources are scarce, and there are competing moral claims on how the available resources ought to be distributed. Many of the claims focus upon the distribution of health outcomes, and thus assume that different health outcomes arising from disparate health care programmes can in some sense be compared. If cardinal values for health states could be elicited, they would help us to distribute resources more accurately towards our chosen health care objectives (whatever they might be). Techniques do exist with which to elicit health state values. However, all have empirical and/or conceptual limitations. Since health care decisions invariably involve a degree of risk, and risk can have an important effect on the valuation of outcome, a crucial requirement for an elicitation technique is that it internalises attitudes towards risk. Of the current commonly used elicitation techniques, only the standard gamble meets this important criterion. Unfortunately, systematic violations of the axioms from which the standard gamble is implied have been observed in many studies. Since the validity of the standard gamble relies on expressed preference, these observations offer a fundamental challenge to the usefulness of the method for eliciting cardinal health state values. It is important for the mainstream health economics community to fully recognise that, in its current state of development; the application of health state value elicitation could do more harm than good. Further methodological developments are required.