目標：主要說明SF-36台灣版的發展過程，同時針對SF-36台灣版量表進行心理計量特質分析。方法：本研究以2001年國民健康調查12歲（含）以上的受訪者為分析樣本（N=18,142），其中共有17,515人進行SF-36台灣版的心理計量特質分析，分析內容涵蓋資料品質、量表結構與信度檢測。結果：在資料品質方面，本研究發現，SF-36台灣版的量表平均漏答率皆低於4%，同時男性的漏答率皆高於女性，而老人高於年輕的受訪者；量表結構方面，各子題與所屬量表的相關係數在0.40到0.83範圍內，同時除心理健康量表外，各量表皆通過區分效度的檢測。在信度檢測方面，則發現內部一致性，除社會功能量表外均達到可接受的信度水準（α＞0.70）。結論：分析結果顯示，SF-36台灣版具有良好的心理計量特質。根據IQOLA研究準則，卻增進SF-36台灣版的解釋力，仍有待常模的建立與後續效度的研究。

Objectives: This paper aims to document the development of the SF-36 Taiwan version for assessment of health related quality of life and perform psychometric testing of the scales. Methods: The study sample was composed of 18,142 subjects aged 12 and above from the 2001 Health Interview Survey. The psychometric testing, which included tests of data quality, scaling assumptions and reliability, was performed on a valid sample of 17,515 subjects. Results: Regarding data quality, the average scale missing rate of all items was less than 4% with male respondents tending o have higher missing rate than female ones, this trend held true for elderly respondents relative to the younger counterparts. In terms of tests of scale assumptions, item-scale correlation coefficients range from 0.40 to 0.83 Except for MH, the rest of the scales have passed the tests of item discriminant validity. Also, internal reliability has reached acceptable level for all scales (α＞0.70), except for SF. Conclusions: In Summary, the SF-36 Taiwan version was shown to possess good psychometric properties. In order to enhance the explanatory power of SF-36 Taiwan version, further research is required to establish the norm and validate the instrument.

SF-36台灣版；心理計量特質；信度；健康生活品質

SF-36 Taiwan version; psychometric property; reliability; health-related quality of life