Objective: To evaluate the functional methodology of Norman H. Anderson in eliciting utilities for health outcomes. Methods: Lay people in Tours, France, rated the undesirability of 40 scenarios of possible outcomes of cardiopulmonary resuscitation on individual linear scales (Study 1) or on a single long scale (Study 2). The outcomes were either 1 of 8 undesirable outcomes, combined with 1 of 5 likelihoods, or else complete recovery, combined with the complementary likelihood. Results: The mean utilities were consistent with previous studies. On the individual level, the internal coherence of most participants’ ratings – defined as the consistency and regularity of the graphic representation of their ratings – improved in Study 1 from their 1st to their 2nd rating. The single scale took less time, but allowed participants to disregard the information about likelihood. Conclusions: Functional methodology provides a powerful means of checking on the understanding and consistency of each person whose utilities are elicited.