目的：利用Rasch 測量分析桌球女子單打五段技術。方法：由四位評分者針對我國特優選手鄭的三場國際正式比賽，使用桌球比賽五段技術評分表執行標記分析，以Facets 軟體的多層面Rasch 模式估計選手、評分者與技術三個層面。結果：一、選手和技術層面的信度優良，且具鑑別度和樣本的代表性；適合度考驗結果資料適合模式，即具有測量效度。二、評分者層面估計結果評分者一致、評分沒有高低差別，且評分者嚴謹度沒有差異。三、技術層面：次數分布大致平均，有利於分散相持段技術訊息被隱藏的部分，即五段比賽技術分析是較佳的桌球比賽技術分析法。四、鄭選手表現較好是發球前段、接發球前段、發球後段，而接發球後段、相持段明顯較差，表示前七拍得分能力較好，七拍之後到相持段是需要加強的部分。結論：利用Rasch 測量分析五段比賽技術分析可以量化桌球單打比賽表現，且獲得各層面的測量訊息。

The purpose of the study is using Rasch measurement to analyze the competition techniques of five stages for women's singles in Table Tennis. Method is Four raters aim at three competition of player Cheng, using the five stages of Techniques to analyze. Use the many-facet Rasch model in Facets software to estimate the three facets of players, raters and techniques. The results are as the following. The homogeneity tests reject raters' measurement was homogeneous which indicated the measured players have discrimination. Norm test was fit to Normal distribution which revealed the samples of players themselves were representative. The data fitted for the mode which had measurement validity. Second, the Rasch reliability of rater facet estimated the result that the rater was consistent and there were no differences at rating score and rigor. The third one is the technique facet. It accepts null hypothesis which meant the null hypothesis of techniques conform to normal distribution and the techniques are representative. The data distribution is approximately equal; it is beneficial to scatter the hidden part of rally-stage skill message. Fourth, Cheng player the best performing is before five strike, poor performance is after sixth strike, need to strengthen. The conclusion is that five-stage competition analysis is a better way to analyze table tennis competition skills. And use the Rasch measurement analysis can quantify the performance of singles table tennis game and obtain the measurement information on each facet.