摘要
本研究的目的是：探討運動技術測驗設定試做的適宜次數。從信度和效度的理論來分析，針對運動技術測驗中常用的設定試做十次為對象，引用林德隆先生的四種打擊測驗原始資料，利用目前非常方便的SAS統計套裝軟體，由變異數分析法計算組内相關係數Intraclass Correlation Coefficient估計信度，採用相關分析法Correlation Analysis和内部一致性效標Criterion of Internal Consistancy分析法，同時考驗效度。結果得到下列結論： 一、信度係數，發現右內打法在試做四次時，左內打法在試做五次時，其餘的在試做六次時，達到試做十次的決定係數之80％；因此，試做六次可視為接近試做十次的信度水準。 二、相關法分析結果，試做七次時四種打擊測驗都達到α=0.0001顯著水準，而且決定係數(r平方）也達到試做十次的0.80以上；所以，試做七次時就具有相當於試做十次的效度。 三、内部一致性效標方面，僅右外打法在試做六次時，達到α=0.0000顯著水準，其餘的都在試做五次時就達顯著水準；因此，試做六次就有試做十次的内部一致性效標水準。
關鍵字
無資料並列摘要
The purpose of this study was to find a good fitness trials for repeated measurment. The approach was referred to reliability & validity theory-based. The subjects took ten trials of the fixed-length test. The raw data was four stroke test and from to LIN. Analysis was utilized for SAS packages of computer programs. The reliability were estimated by Intraclass Correlation Coefficient, and validity analysis of two types of methods were used Correlation Analysis and Criterion of Internal Consistancy. The results were concluded as the following: 1. Four trials of right-to-in stroke, five trials of lift-to-in stroke, and the other at six trials were reached to ten trials 80% coefficient of determination in reliability. Therefor, six trials approximate to ten trials on the level in reliability. 2. The results of Correlation Analysis reached the level of significance (α=0.0001) in the four types of strokes at seven trials, and the coefficient of determination could be more than 0.80 often trials. Therefor, seven trials heve equaled ten trials on the level in validity. 3. On the Criterion of Internal Consistancy, only six trials of right-to-out stroke reached the level of significance (α=0.0000), and the other at five trials had reached the level of singificance too. Therefor, six trials heve had ten trials on the level in Criterion of Internal Consistancy.