雙重處理模型和修正基底模型的整合性框架研究：論證診斷性的調節作用

Evan though the dual process models and correction-based models have been largely employed in a variety of studies examining how attitudes are formed or changed, the current study proposes that the interpretation of how target evaluations are made in high involvement conditions fails to converge. A moderating role of argument diagnosticity is suggested in the current study to consolidate the discrepant theoretical perspectives, and an integration framework of the dual process model and the correction-based model in attitude formation and change is derived accordingly The persuasive message is regarded as high in argument diagnosticity when a large number of cognitive responses relevant to target’s central merits can be engendered in the process of advocating information. On the other hand, the persuasive message is regarded as low in argument diagnosticity when few cognitive responses relevant to target’s central merits can be engendered in the process of advocating information. Argument diagnosticity is orthogonal to argument quality and target favorability. That is, when consumers or social perceivers are high in cognitive resources, the persuasive message high in argument diagnosticity largely facilitates the impact of argument quality on target assessments. The proposed judgmental process is consistent with the dual process model. On the other hand, the persuasive message low in argument diagnosticity enhances the scrutiny of undue impacts made by peripheral cues or contextual stimuli to high involvement consumers or social perceivers, and further give rise to bias correction in an attempt to accurately make target assessments. In addition, the current study proposes an integrated framework to seamlessly accommodate the dual process model and correction-based model in the theorization of attitude formation and change. By virtue of a clear identification and clarification of the moderating role of argument diagnosticity, the distinctive and shared antecedents and consequences of the dual process model and the correction-based model can then be well comprehended and consolidated in a comprehensive framework. The current study adopts a 2 (involvement: high vs. low) ? 2 (cue: attractive vs. average) ? 4 (argument: diagnostically strong vs. non-diagnostic vs. none vs. diagnostically weak) between participant factorial design. Except the typical findings of dual process model and correction based model, more importantly, the current study provides unexplored theoretical resolutions on 1) high involvement does not necessarily result in the information process via central route or systematic process, 2) the main effect of argument quality is not always available when involvement is relatively high, 3) high sensitivity on bias influence or bias knowledge does not always engender bias correction when elaboration likelihood is at a relatively high degree, and 4) bias reminders are not a necessary condition for the activation of bias correction.