We propose that consumers’ preferences for brands are systematically affected by whether they make direct choices between the brands or rate them separately in terms of purchase likelihood. Specifically, preference reversals are expected such that low quality, low price brands and low quality brands that have unique product features are preferred more in choice whereas high quality, high price brands are preferred more in ratings. These predictions are derived from an analysis of the differences between choice and ratings and the compatability between these taks and the attribues of the considered options. Binary choice often involves within-attitude comparisons and accross-attribute comparisons of intervals (e.g., whether the difference between a Sony TV and a Magnavox TV is worth ($100), which is expected to enhance the importance of attributes that produce precise and unambiguous differences (e.g., price) relative to more complex, qualitative attributes that are more difficult to compare (e.g., price) relative to more complex, qualitative attribute that are more difficult to compare (e.g, brand names). In a rating task, on the other hand, brand names may often serve as powerful cues, and the cost of a better brand name in terms of price and features is less salient and more difficult to detect. The predicted preference reversals were supported in six studies, which also examined several rival explanations. We discuss the theoretical implications of this research and explore its consequences for the measurement of buyers’ preferences and for marketers’ pricing, merchandising, distribution, and communications strategies.