When we read some studies about consumer behavior, we will find some theory is not consistent; for example, ELM and Kang and Herr’s study about bias correction. ELM theory considers the favorability about the product attitude of high involvement people is only influence to the argument. That is, when argument is strong, people attitude toward the product is positive; on the contrary, if the argument is weak, the attitude of high involvement individuals is negative. Nonetheless, the outcome is overthrow by Kang and Herr’s study in 2006. Their research conclusion shows when the endorser is attractive but irrelevant to the target product, and the elaboration of participants are extremely high, there will lead to negative attitude. In Chiu’s study (2008), she considers the result of this conflict between ELM and Kang and Herr’s theory is short of central merit information of the target. In this condition, High involvement people will consider the attractive but irrelevant source as bias and correct will be triggered. Our study use the same structure of Chiu’ study and we adopt the argument strength as the mediator of our experiment.