Several valuation techniques are in use for quality adjusting life years in cost utility analysis. The paper gives an overview of the variability in results. A close inspection of a number of instruments with respect to their theme, instructions, decision framing and the phrasing of questions make many of the observed differences in results understandable. When judging the validity of the different techniques, three points should be kept in mind. One is that statements about validity should be made with respect to concrete versions rather than broad categories like 'the rating scale', 'time trade-off' etc. Another point is that a valuation technique that is valid in clinical decision analysis may not be valid in health program evaluation, and vice versa. The third point is that quality weights for life years are empirically more meaningful, in the sense that they are more amenable to empirical testing, if they are interpreted simply as preference weights rather than measures of amounts of well life in the utilitarian tradition. Time trade-off with a moderate time horizon is recommended in clinical decision analysis, while a combination of time trade-off and a variant of person trade-off is recommended in health program evaluation.