In addition to the scaling method, there are many other aspects of the measurement process that may affect rater judgments of the relative desirability of health states. Although we find little compelling evidence of population differences in preferences due to demographic characteristics, there is some evidence suggesting that medical knowledge and/or experience with illness may influence raters' valuations of health states. Other aspects of the rating process that affect rater judgments can be classified as one of two types: inconsistencies due to limitations in human judgment, and inconsistencies due to situation-specific variables. When inconsistencies are due to limitations in human judgment, such as framing effects, a reasonable solution is to help the rater to see and correct the inconsistency. When inconsistencies are due to situation-specific variables, such as the way the health state is defined and presented, investigators should attempt to standardize conditions across studies.