目的：探討現場記錄與錄影帶記錄兩種方式是否會影響計時起走測驗之施測者內或施測者間信度。方法，共徵召36位平均年齡44.4歲（範圍20~76歲）之13位男性和23位女性志願者（18位正常人，18位神經或肌肉骨骼系統疾病患者）參與本研究。由三位施測者於現場以碼錶記錄受試者完成計時起走測驗所需的時間（秒），共施測兩次，同時錄影受試者兩次的表現。三位施測者分別再於一週後由錄影帶觀察此兩次重複測驗之成績，再於間隔兩個月後，再次觀察受試者錄影帶中第一次的表現。結果：現場記錄與錄影帶記錄兩種方式之施測者內信度與施測者間信度之組內相關係數(ICC)皆高達0.99 (p<0.0001)。然而無論於現場記錄或錄影帶記錄，三位施測者所得之受試者第一次表現結果（平均16.8秒）均顯著大於第二次表現結果（平均為16.2秒）。由現場記錄與由錄影帶記錄受試者同一次表現結果，則約相差0.3秒，但無統計上顯著差異(p>0.05)。由錄影帶兩次記錄受試者同一次表現結果，則約僅相差0.1秒(p>0.05)。結論：現場記錄和錄影帶記錄計時起走測驗均可獲得極佳之施測者內及施測者間信度。然而受試者兩次測試之表現不同，第二次明顯地於較短時間內完成測試動作。

Background and Purposes: The purpose of this study was to test whether rating the performance of the subjects using the "on-site" versus "off-line" analyses would affect the intrarater and interrater reliabilities of the timed "Up and Go" Test (TUGT). Methods: Sixteen healthy adults and sixteen subjects with neurological or orthopedic diseases (mean age=44.4 years; 13 males, 23 females) participated in this study. The subjects completed two trials of the TUGT consecutively on the same day. The time to complete the test was recorded on-site by 3 experienced physical therapists. The performance of the two trials was recorded on videotapes for analysis on a later day. One week later, the 3 therapists analyzed the performance of the recorded trials from videotapes (off-line 1).Two months later, the 3 therapists analyzed the first trial of each subject from the videotapes again (off-line 2). Results: Excellent intrarater and interrater reliabilities were found for the TUGT using both the "on-site" and "off-line" analyses (ICC≧0.99). The performance of the subjects in the second trial (mean=16.8 seconds) was significantly faster than that in the first trial (mean=16.2 seconds, p<0.0001). There was only approximately 0.3 seconds difference between the same trial analyzed with the "on-site" and "off-line" methods (p>0.05). The repeated analysis of the same trial from the videotape (off-line 1 vs. off-line 2) showed only approximately 0.1 seconds difference (p>0.05). Conclusions: Intrarater and interrater reliabilities for the TUGT were excellent for both "on-site" and "off-line" analyses. However, the performance of the subjects in the second trials was significantly faster than that in the first trial.