本篇論文係以台灣新上市股票作為研究對象，欲檢驗「投資者意見差異變數」 對於股票長短期績效之解釋能力，使用相似之「投資者意見差異變數」，是否同樣可以合理解釋台灣新上市股票市場的現況，台灣的實證是否同樣支持Miller所提出的假設。並進而探討在不同承銷制度下，是否會影響「投資者意見差異變數」對於上市績效之解釋能力與影響方向。 研究使用「蜜月期天數」、「買賣價差百分比」與「新股蜜月期間周轉率」三者作為「投資者意見差異程度」之代理變數，研究結果發現：此三變數在排除公司特性、發行市場現況等控制變數的影響部分之後，其餘無法被控制變數所解釋的部分，確實對於新上市股票之長短期報酬具有解釋能力，且各意見差異變數與新股短期報酬率之間存有正向關係，而與一年至三年之長期持有報酬率間存有反向關係，此研究結論相當支持Miller假設，除此之外，研究並發現：不同新股承銷制度不會顯著地影響投資者意見差異程度，且若使用公開申購與競價拍賣兩種不同的新股配售方式，僅會影響「投資者意見差異變數」與「新股長期報酬」的影響程度與方向。 / This paper focus on the Taiwan’s IPOs and tries to test the direct relationship between “ divergence of opinions ” and “ the short- run and long- run performance of the initial public offerings (IPOs). ” We use several similar variables to explain the situation of Taiwan’s stock market and find out if data can support the Miller’s hypothesis. Besides, we want to discuss the relationship between “ IPO method ” and “ divergence of opinions ”. We check that different IPO method will have influence on the explanatory power of the “ divergence of opinions”. We use three different variables：“ the number of the honey moon date”, “ the percentage of bid-ask spread ” and “ the turnover rate during the honey moon” as the proxies of the divergence of opinions. The research result finds out that three proxies can truly provide significant explanatory power of the short-run and long-run performance of the IPOs after controlling for issue quality. The results support Miller ( 1977 ), who suggests that greater divergence of opinions or uncertainty about an IPO can generate short-run overvaluation and long-run underperformance. In other words, the divergence of opinions have positive relationship with the initial return and negative relationship with the long run holding period return of IPOs. Besides, the two IPO method ( fixed-price offerings and discriminatory auction ) will only change the long run relationship.