Abstract
The author examines how rating a brand on specific attributes early in a survey affects responses to a later overall brand evaluation. Specifically, she investigates the conditions under which respondents’ answers to a brand evaluation are likely to be consistent with a previous attribute rating (carryover effect) versus inconsistent (backfire effect). In a laboratory experiment, the occurrence of carryover and backfire was related to the respondent's level of subjective product knowledge and the diagnosticity of the attribute rating for the brand evaluation. Implications of these findings for questionnaire design and measurement of consumers’ attitudes are discussed.