本論文分別以報酬波動性和財務比率兩種方式來判別個股的投機性，並依此標準建構投資組合進行研究。除了使用代表三個投資族群之情緒指標外，並以各情緒指標之第一主成份的因素負荷量建構一個複合情緒指標，探討情緒對股票當期報酬影響與未來報酬預測。最後，參考Baker and Wurgler(2006)所建構之「高－低特徵投資組合之迴歸模型」討論情緒對股市橫斷面的預測效果。實證結果顯示：使用複合情緒指標可觀察到情緒與股票當期報酬呈正向關係，且與下期報酬呈負向關係，即報酬產生反轉。亦印證了投資人情緒對投機性愈大之股票之報酬的影響愈大。但在三因子為控制項之模型下，未能提供顯著且一致的結論證實當情緒高（低）時，具有投機特徵的股票會有較低（高）的未來報酬。

This study measures stocks' speculative level according to the volatility of stock returns and specific financial ratios, and then designs portfolios for both measures. Not only does this study adopt several sentiment indexes representing three different investor groups, but it also constructs a complex sentiment index using loadings of first principal component analysis for all the sentiment indexes in this study. The goal of this study is to discuss the impact of investor sentiment on current and future stock returns. Finally, we used the "Predictive Regression for Long-Short Portfolios" model (Baker and Wurgler, 2006) to examine how future returns are relatively low (high) when investor sentiments are high (low) for stocks with speculative characteristics. The empirical results of this study show the following: (1) When using the complex sentiment index, the deviations in stock prices caused by investor sentiment are corrected in the next month. (2) Overall, the returns on speculative stocks are most affected by investor sentiments. (3) If the Predictive Regression Model is controlled with the variables of market risk premium (RMRF) and Fama-French factors (SMB and HML), there is no strong evidence that future returns are relatively low (high) when investor sentiments are high (low) for stocks with speculative characteristics.