本研究目的在於修訂中文版寂寞量表(Loneliness Scale)，並檢驗其效度與信度。研究首先將De Jong Gierveld 與 Kamphais 所發展之寂寞量表翻譯及修訂為中文版，並以108位65歲以上之老人大學學員為研究對象，予以施測此中文版量表。本研究效度方面以建構效度、內部一致性、分殊效度、幅合效度和區辨效度檢驗之，信度方面則是以內部一致性信度(KR-20)檢驗之。研究結果顯示：一、中文版寂寞量表可區分為「社會性寂寞」與「情緒性寂寞」兩個分量表，總累積解釋變異量為50.98%；二、兩個分量表分數與總量表分數間呈現中至高度的顯著相關，但兩分量表分數之相關低且未達顯著；三、量表與兩種人格特質(神經質和外向性)之相關結果顯示中文版量表總分與「神經質」分數呈現顯著正相關，而與「外向性」分數呈現顯著負相關；社會性寂寞分量表分數與「外向性」分數呈現顯著負相關，而與「神經質」分數呈現顯著正相關；情緒性寂寞分量表分數與「神經質」分數呈現顯著正相關，但與「外向性」分數相關極低且未達顯著；四、在信度部份，整體內部一致性信度KR-20為.67，兩個分量表的內部一致性信度KR-20則分別為.79與.56。整體而言，修訂之寂寞量表具有可接受的效度與信度，未來可運用中文版寂寞量表於測量老年人之寂寞程度與兩種寂寞類型。

The purpose of this study was to revise Loneliness Scale (LS) into Chinese version and examined its validity and reliability. We first translated the LS into a Chinese version. The validity and reliability of the Chinese version LS was then tested on the elderly of 108 senior college's members. The dimensionality of LS was examined by factory analysis, and other examinations for the validity included internal consistency、convergent validity and discriminative validity. Reliability was shown by coefficient of internal consistency. The results were as follows: (1) The Chinese version LS was divided into two subscales of 「emotional loneliness」 and 「social loneliness」, and the total explained variance was 50.98%; (2) Moderate to high correlations were found between two subscales scores and the whole scale scores, and low correlations was found between two subscales scores; (3) We examined the relationship of the whole scale score and two subscale scores to neuroticis and extraversion scores, results indicated that the whole scale score was significantly positively correlated with neuroticis scores and significantly negatively correlated with extraversion scores; besides, the subscale score of social loneliness was significantly negatively correlated with extraversion scores and significantly positively correlated with neuroticis scores; the subscale scores of emotional loneliness was significantly positively correlated with neuroticis scores; (4) As to reliability, the internal consistency reliability was assessed using Kuder Richardson 20 (KR-20), and KR-20 of the total scale and two subscales were .67, .79 and .56. In sum, this study found that Chinese Version LS with appropriate reliability and validity and suggested that future study may apply the scale to study elderly people's loneliness and two types of loneliness.