The majority of subjective data collected by survey researchers uses closed-ended questions with rating scales. Questions assess how much one agrees or disagrees with some statement, how important something is, or how frequently one engages in some behavior. The success of measuring this type of subjective phenomena depends on the characteristics of the scale used. However, the literature addressing the appropriateness of various types of rating scales is not always consistent in its conclusions or recommendations for use in applied settings. Using a U.S. Department of Agriculture diet and health survey instrument, we designed a study to investigate some of the inconsistencies reported in the literature. We conducted 300 telephone interviews with adults recruited from the D.C. metropolitan area. Three scale characteristics were manipulated: 1) the length of the scale; 2) the extent of verbal labeling, and; 3) whether to use an unfolding scale. This paper reviews some of the literature on rating scales, and presents preliminary results from the study.