There is growing interest in discrete-choice experiment (DCE) as a method to elicit consumers' preferences in the health care sector. Increasingly this method is used to determine willingness to pay (WTP) for health-related goods. However, its external validity in the health care domain has not been investigated until now. This paper examines the external validity of DCE concerning the reduction of a health risk. Convergent validity is examined by comparing the value of a statistical life with other preference elicitation techniques, such as revealed preference. Criterion validity is shown by comparing WTP values derived from stated choices in the experiment with those derived from actual choices made by the same individuals. Both tests provide strong evidence in favour of external validity of the DCE method.