運動技術測驗通常採用主客觀的方法設定給分類別，但應該根據何種機率理論最適當？因此，本研究的目的是模擬研究運動技術測驗給分類別的最適合機率分配。研究對象是：常態分配、羅吉斯分配、二項分配等三種機率分配，研究方法先利用SAS和Minitab隨機產生模擬資料，然後用Rasch評分量尺模式估計資料，實驗設計為樣本數和測驗長度二因子，各模擬24次。結果：第一樣本大小在3,000個反應資料以上時，常態分配是理想的機率分配。第二樣本大小在1,500個反應資料以下時，羅吉斯分配是較好的機率分配。結論：Rasch模式評分量尺類別設定宜依據適當的機率分配，且選擇的機率分配和樣本大小有關。

The testing of sports skills usually was used to set the score categories by subjective method, but what should it assign the most appropriate probability theoretical basis? Therefore, the purpose of the study was to optimize distributions of probability the score categories of sports skills testing in simulation research. The objects of the study were normal distribution, logistic distribution, and binomial distribution. The method was to use SAS and Minitab to produce random data, and then use Rasch rating scale model to estimate data. Experiment design was two-way design (sample × test length), and we simulated them 24 runs for each cell. The results were: 1. Normal distribution was the ideal distribution of probability when sample size was over 3,000 in response data. 2. Logistic distribution was the better distribution of probability when sample size was less than 1,500 in response data. The conclusion of the study was that Rasch rating scale model wish set the appropriate category based on probability distribution, and the probability distribution of selection was related to the sample size.