近年，投資人對資訊反應不足與反應過度，是行為財務學側重的一塊。許多文獻與實證研究皆探討投資人的投資決策與其對資訊反應不足與反應過度之間的關連性，以及究竟投資人是否有設定投資定錨的傾向。而本研究為了實證台灣股市存在之反應不足或過度反應的現象，且投資人具有限注意力並有設定投資定錨的傾向，在博覽眾多相關文獻與研究後，決定依循Li and Yu（2012）的理論基礎，使用週價比與歷價作為反應不足與過度反應的代理變數，並將模型改建成適用於台灣股市的迴歸模型。 本研究以台灣股價加權指數作為主要迴歸樣本，將模型分為週資料迴歸模型與月資料迴歸模型，檢視兩個代理變數在控制景氣循環以及沒有台灣經濟泡沫樣本下可否用於台灣股市反映投資行為，並作為預測未來市場報酬的指標。最後，為實證投資人有限注意力與心理定錨設置之理論成立，本研究將台灣50指數與摩根台股指數作為樣本分別進行迴歸，以探討是否有比台灣加權股價指數更具顯著預測能力的指數存在。 研究結果顯示，假說一「週/歷價比與未來市場報酬成正/負相關」與假說二「選用能見度高的指數作為樣本應使週價比與歷價比更具顯著的預測能力」皆成立。即台灣股市確實存在反應不足或過度反應的現象，且投資人具有限注意力並有設定投資定錨的傾向，而台灣市場中最具顯著預測能力的指數為摩根台股指數。此外，週資料迴歸模型比月資料迴歸模型更適用於台灣股票市場。 / Currently, much academic research is concerned about investor underreaction & overreaction in behavioral finance. It mainly discusses the relationship between investment strategies and underreaction & overreaction and whether investors have tendency to utilize investment anchor. This study determined to follow basis of Li and Yu（2012） in order to examine the existence of underreaction & overreaction, investor limited attention and the tendency to set anchor in Taiwan stock market. We use nearness to the 52-week high and nearness to the historical high as proxies capturing the degree of investor underreaction and overreaction to news. In this study, we adopt TAIEX as main data and run two different types of regression model based on weekly and monthly data. Except under normal condition, we further examine these two proxies with controlling business cycle and without Taiwan economic bubble. Finally, we compare the predictive ability to forecast future aggregate market returns among TAIEX, TW 50 and MSCI Taiwan index. Our empirical results support the hypothesis 1, “nearness to the 52-week high positively predicts future market returns and nearness to the historical high negatively predicts future market returns” and hypothesis 2, “using index with higher visibility results in significantly predictive ability for nearness to the 52-week high and nearness to the historical high,” while MSCI Taiwan Index is the most significant. Besides, weekly regression is more suited to Taiwan stock market than monthly regression. These findings are consistent with the limited investor attention and anchoring research.