「臺灣地區中老年身心社會生活狀況長期追蹤調查」短版CES-D量表之心理計量特性

The Psychometric Properties of a Short Form of the CES-D used in the Taiwan Longitudinal Study on Aging

研究目的：本研究檢測「臺灣地區中老年身心社會與生活狀況長期追蹤調查」（TLSA）使用之短版CES-D量表分數的信度、效標關聯效度、以及建構效度，以瞭解此短版CES-D量表測量臺灣老人憂鬱症狀嚴重程度之適用性。研究方法：本研究分析TLSA自1989至2003年間五波調查均採行之10題CES-D量表分數的內部一致性信度、相鄰兩波間之相關、及與人口學變項之關係，並以日常生活活動、工具性日常生活活動、自評健康與生活滿意度作為效標，探討量表分數之效標關聯效度，另亦以探索性因素分析檢視此短版CES-D量表之因素結構。研究結果：TLSA短版CES-D量表分數具良好內部一致性信度，女性與不識字老人之憂鬱症狀嚴重程度較高，高憂鬱症狀程度亦伴隨較差之日常生活活動與自評健康，以及較低之生活滿意度。此短版CES-D量表內容呈現「身體症狀」、「憂鬱情感」與「正向情感」三因素。研究結論：TLSA之10題短版CES-D量表具良好的心理計量特性，應可作為測量臺灣老人憂鬱症狀嚴重程度之工具。

Purpose: The 20-item Center for Epidemiological Studies Depression Scale (CES-D), developed by Radloff (1977), has been widely used in survey research to assess depressive symptoms among older adults. Of the original 20 CES-D items, 10 have been adopted repeatedly in the Taiwan Longitudinal Study on Aging (TLSA) in 5 waves of investigation, as well as in other studies. The psychometric properties of this short form of the CES-D therefore deserve a thorough examination. We examined the reliability, criterion-related validity, and construct validity of scores from the 10-item short form of the CES-D used in the TLSA to evaluate their appropriateness for assessing depressive symptoms of older adults in Taiwan. Methods: We analyzed data collected with the 10-item short form of CES-D in the 5 waves of the TLSA conducted between 1989 and 2003. Internal consistency reliability of scale scores, stability of the scores over time, and the relationships between measured scores and demographic variables, activities of daily living (ADL), instrumental activities of daily living (IADL), life satisfaction, and self-rated health were examined. Exploratory factor analysis was performed to assess construct validity. Results: The short form showed satisfactory internal consistency reliability and stability over time. Females and less-educated older adults tended to have more depressive symptoms. Participants who indicated poorer ADL/IADL, lower self-rated health, and less satisfaction with life were likely to report more depressive symptoms. Three factors were found underlying the short form, namely Somatic Symptoms, Depressed Affect, and Positive Affect. In addition, the results also supported the invariance of the 3-factor structure of the short form across 5 time points. Conclusions: The 10-item short form of CES-D in the TLSA demonstrated satisfactory psychometric properties and can be considered a reliable and valid measure of geriatric depressive symptoms among older adults in Taiwan.