本研究主要探討於高低涉入情況下，接受不同共識性訊息的受試，在閱讀不同品 質之言論後，在態度上所產生的差異效果。本研究採２（高╱低涉入）×２（高╱低共識性 ）×２（強╱弱論點）之多因子設計（ factorial design ）。以受試者對論點的同意程度 、及語意區別測量做為受試者態度之指標。本研究預測會得到顯著的三因子交互作用。亦即 ，在高涉入的情況下，強論點的說服效果會大於弱論點，而共識性高低則不會有顯著的影響 。在低涉入情況下，論點品質與共識性會產生交互作用；亦即，論點品質在說服上的不同效 果，只有在高共識性時才會出現，在低共識時則無此效果。大致而言，研究結果與此預測十 分一致。 然而，認知反應與態度指標之迴歸分析顯示，造成此結果的內在歷程與 Petty 及 Cacioppo （ 1986 ）的思考可能性模式，及 Chaiken （ 1987 ）的啟發性處理的推測並不 完全一致。本研究就此討論情感反應（ affective response ）在說服歷程中，可能扮演的 功能，以說明本研究所發現的矛盾現象。

　　　　　An experiment was conducted to examine the effects of involvement, consensus information and argument quality on subjects' attitudes. A 2(high vs. low involvement) × 2 (high vs. low consensus information) × 2 (strong vs. weak argument) factorial design was employed. It was hypothesized that under the high involvement condition the strong-argument message would be more persuasive than weak-argument one, and the consensus information would have no impact on subjects' attitudes. However, under the low involvement condition an interaction effect between argument quality and degree of consensus would be expected. These predictions were fully supported by various attitude measures. These findings suggest that the high consensus information (majority influence) may enhance subjects' motivation to process the message, but this effect dissipates when subjects were highly motivated to process the message by the high involvement manipulation. However, the relationships between their attitudes and cognitive responses within each experimental conditions have revealed that under the low consensus information condition (minority influence) subjects' attitude changes may be mediated by their elaboration of the messages, but this is not the case under the high involvement condition. This is inconsistent with the prediction derived from Petty and Cacioppo's (1986) ELM. The authors suggest that to reconcile this conflict findings the role of subject's affective responses while processing the message, besides their cognitive responses, should be included in future research.