The purpose of my article (Yi, 1989) was to explore what might be the benefits of employing a structural representation of expectancyvalue (EV) attitude, compared with the traditional representation (i.e., a summary value). In doing so, the implications of the structural representation were investigated. Specifically, the question addressed was: Does the structural representation of EV attitude provide any advantage for understanding (a) the relationship of EV attitude to A-act and BI, and (b) the dynamics of belief change? Vanden Abeele (1989) made some critical comments about this article and questioned the validity of the results. He concluded that “many of the claims made in the paper are debatable; more importantly, such debate is the result of a more general disagreement with the strand of EV research exemplified by this paper” (p. 85). His comments can be summarized as the following three key issues. (1) The EV framework is a structural model, not a measurement model; joint testing of structure and measurement in a model -. should be avoided. (2) The hypotheses and results in the paper are ambiguous. (3) The generalizability of the results in the paper can be questioned. These are important and interesting issues in the investigation of EV attitude. However, I have found that many of the comments are difficult to accept and stem from a misunderstanding of hy claims. In this paper, I will examine the validity of these comments and clarify the confusion surrounding my original arguments. To facilitate the presentation, each of the three issues will be discussed in sequence.