Application of Many-Facet Rasch Model to Analyze the Skills of Elite Athletes in Double Trap

本研究的目的是應用多層面Rasch模式分析雙不定向飛靶的射擊技術。研究對象是國家射擊雙不定向飛靶代表隊現役最佳選手，包含男子五名與女子三名；分析資料包括國內2005~2006年正式比賽及國家隊選拔測驗，比賽形式皆為奧運賽制，每人各射擊三回合、每回合150發，決賽加射一回合。所有紀錄的數據，透過Facets測驗軟體分析。射擊技術分析八個重點（層面）包括選手、賽制、日期、拋靶類型、射擊回合、開始靶位、射擊靶位及目標。利用適合度考驗(INFIT & OUTFIT)、難度(measure)、估計誤差(Model S. E.)和非期望反應(unexpected responses)建立各層面技術的量尺，然後根據非期望反應和標準殘差(standardized residual)進一步探討選手個人的技術類型和優缺點。結果能夠具體、精確的分析選手在雙不定向飛靶的射擊技術，提供教練在訓練上有一個明確的指標，排除主觀的觀察的弱點，並提升選手的成績表現。研究結論是利用測量的非期望反應可以探查雙不定向飛靶運動員的射擊技術。

關鍵字

多層面Rasch模式；非期望反應；射擊技術；雙不定向飛靶

並列摘要

The purpose of the study was to analyze the shooting skills of elite athletes by using many-facet Rasch model in the Double Trap. The subjects were five male athletes and three female athletes who were in the best ranking of national Double Trap teams. The data of Double Trap included the performance of national game 2005-2006 and experiment shoot 150 by 3 rounds in Olympic discipline. The data were analyzed by Facets Rasch software. The shooting skill of Double Trap included eight key elements (facets): Shooter, Program, Date, Trap Setting, Round, Starting station, Firing station and Target. The scale of skill was created, using Infit (the information-weighted mean-square fit statistic), Outfit (the outlier sensitive, mean-square fit statistics), Measure (logit estimate of the calibration/measure in log odds units), Model S. E. (standard error), and unexpected responses in all facets. A person's skill type and character were based on unexpected responses and standardized residual to analyze further. The results could objectively and precisely explain the athlete's shooting skills in the Double Trap. It had an index clearly for coaches in the sports training. It eliminated the subjectivity and promoted athletes' performance. Therefore a conclusion was that using the unexpected responses of Rasch measurement might detect the shooting skills of shooters in the Double Trap.