本研究試圖結合以往學者所發現的規模效應、權益帳面值（即淨值）對市價比效 應、本益比較應、風險係數與前期報酬率等之變數，探討以上各變數對股票報酬率之影響程 度。在風險係數之估計上，本研究採用Scholes & Williams所採用的beta估計法，以求更能 代表真實市場之風險。此外，有鑑於Fama & French(1992)之迴歸方法，僅考慮了橫斷面資 訊，而忽略了縱斷面時間因素所產生序列相關之問題，本研究同時比較三種時間序列與橫斷 面混合迴歸分析(time-series/cross-section pooling regression)方法，解決了變異數齊 一性、時間序列獨立性之問題，希冀研究結果更能代表實際狀況。研究結果發現，臺灣股市 存在有規模效應、淨值對市價比效應，而風險係數為不重要的變數。月份效應方面，臺灣股 票市場月份間報酬率確實存在差異。進一步利用Duncan多平均數距檢定法，發現所有月份中 以二月份之報酬率最高。若以報酬率最高的二月份而言，不論規模效應、淨值對市價比效應 、本益比效應、風險係數與前期報酬率，皆能顯著影響股票報酬率。

　　　　　This paper analyzes the impacts of the size effect, BE/ME effect, P/E ratio effect, risk coefficient(beta) and prior returns on stock returns. Specifically, the beta coefficient is estimated using Scholes & Williams methodology. Moreover, in Fama & French (1992) paper, only cross-sectional information were taken into account, which ignored the serially correlated problem existed in the time series data. This paper utilizes three time-series/ cross-section pooling regression models so that both time-series and cross- section variations are considered. The results show that both size effect and BE/ME effect have significant impacts on the annual stock returns. The beta coefficient, however, has little influence on annual returns. For the monthly effect, the rates of returns among months are significantly different in Taiwan stock market. The rate of returns in February is the highest and the beta coefficient becomes the most influential factor in February. All of the variables, including size, BE/ME ratio, P/E ratio, risk coefficient and priors return, have significant impact on February stock returns.