Research suggests that individuals do not respond to survey questions on the basis of a single, fixed set of psychological considerations. To the contrary, they respond on the basis of whatever material happens to come to mind at the moment of answering. Furthermore, the particular material that comes to mind often depends upon the nature of the question and the manner in which it is posed. This has important implications when attempting to assess the independence of beliefs and emotions as predictors of global attitude judgments. When addressing this question, it is important to consider the extent to which any given measurement procedure specifically targets ‘emotionally relevant beliefs’ in memory. Past research has failed to target this specific subset of beliefs. As a consequence, beliefs have failed to fully account for the relation between emotions and global attitude judgments. Completely different findings emerge when the researcher employs ‘decomposition’ and ‘emotionally cued recall’ to enable respondents to generate ‘emotionally relevant beliefs’. Under these conditions, beliefs fully account for the relation between emotions and global attitude judgments. However, this only occurs when one defines and operationalizes ‘emotionally relevant beliefs’ at the individual level. The theoretical and methodological implications of these results are discussed. © 1997 by John Wiley & Sons, Ltd.