由於網際網路缺乏適當的過慮機制，使得謠言充斥於網路之中，而不實的網路謠言可能會對企業造成極大的影響，本研究即針對闢謠之相關議題進行探討，有助於廠商在受到謠言影響時思考其因應的策略。影響闢謠效果的因素衆多，闢謠者的來源可信度知覺與情感認同，雖無關於訊息內容，但卻可能影響收訊者對訊息的相信程度，而收訊者的涉入程度，則牽涉到訊息對其的重要性，因此可能影響到收訊者是否會仔細思慮闢謠訊息，或是僅以闢謠者的身分等相關週邊線索，來判斷是否相信闢謠的澄清聲明。本研究利用過去學者所提出的社會判斷理論、認知失調理論、平衡理論、推敲可能模式等理論，針對闢謠者可信度、情感認同、涉入程度等變數，與闢謠效果間的關係進行探討。研究採取問卷調查的方式，將所獲得的資料利用描述性統計、t-test、集群分析、Pearson相關分析、線性結構關係分析、X^2、MANOVA等統計方法進行分析，研究結果發現，闢謠的確能夠造成訊息接收者對謠言相信程度上的改變，且收訊者的涉入程度與其對闢謠者的可信度知覺都會造成人們對謠言相信程度之改變。

Lacking of appropriate screening mechanism, the Internet becomes a channel for transmitting rumors. However, negative rumors may influence the sales, profits, reputation, and even stock price of the firms affected by rumors. Previous researchers have discussed the contents, formation, and transmission of rumors. Few studies focus on the denial effect or had not empirical evidences to support their argument. This study tries to provide insights into the belief reduction of the rumors that can help the firms adopt proper refutation strategies. Enormous variables might be expected to affect receivers' belief including the personal characteristics, the tone of denial source, the source of rumors, and etc. In this study, we examined that f source credibility and receivers' affection toward the source would influence their belief in that rumor. Moreover, due to the difference of receivers' involvement, they may either elaborate the refutation message carefully or adopt the peripheral cues such as sources' characteristics to make the judgment about that rumor. This study adopt social judgment theory, cognitive dissonance theory, balance theory, and elaboration likelihood model to explore the relationships among source credibility, affection, involvement, and the effects on belief reduction. 304 usable questionnaires were collected from undergraduate students at three universities at Taiwan. In this study, we chose a rumor concerning about Microsoft auto-detect technique. It may help us to identify the involvement of the receivers. In the beginning, the respondents were asked to measure their affection and credibility either toward Microsoft or toward a professor teaching in the Department of Information Management. Then, they were asked to read the information (rumor) concerning Microsoft c auto-detect technique and measure their belief and involvement in that information. After that, we provide them with information released either from Microsoft or form the professor claiming that the information concerning Microsoft's auto-detect technique is a rumor. The subjects were asked to re-measure their affection and credibility either toward Microsoft or toward the professor. Finally, we show them the information again and asked them f they believe in the information. After completing the questionnaire, subjects were debriefed using the statement regarding the truth of the rumor. Several statistics technique including descriptive statistics, t-test, cluster analysis, Pearson correlation, linear structural relation, X^2, and MANOVA were used for analysis. According to the results, 93 (30.6%) respondents did reduce their belief in that rumor, 164 (53.9%) respondents still remain the same, and 47 (15.5%) respondents even put more belief in the rumor. From the results we obtained, the denial efforts may be effective in reducing the belief in rumors. However the causes of negative effect may need further discussion. To facilitate further discussion of the impact of personal involvement, respondents were classified into high or low involvement groups. Results show that both source credibility and involvement have significant influences on belief reduction. And further, source credibility and affection have more influence on low involvement group than on high involvement group. In order to further examine the effects of personal involvement and affection on denial effects, this study also classified the respondents into four groups, namely high involvement and high affection group, high involvement and low affection group, low involvement and high affection group, low involvement and low affection group. Through multiple comparisons, we find that respondents reduced their belief in that rumor not only relied on their initial involvement or affection status. Those who are in the high involvement and high affection group are easily changing their on the rumor or shifting their affection toward the denial source.