成人依戀量表台灣修訂版的心理計量特性與預測心理適應之探討

Psychometric Properties of the Chinese Version of Revised Adult Attachment Scale and Its Prediction to Psychological Adjustment

本研究的目的乃在編修適用於台灣的成人依戀量表，並檢驗建構於向度測量概念之中文版成人依戀量表修訂版（RAAS-C）在台灣樣本的心理計量特性。本研究採用兩個樣本，樣本一有177位參與者，用於進行探索性因素分析，樣本二有240位參與者，用於進行驗證性因素分析與信效度檢驗。研究結果如下：首先，探索性與驗證性因素分析結果顯示，焦慮依戀與逃避依戀兩因素結構是較佳的模式，此結果與現行成人依戀理論主要的兩因素結構一致。其次，焦慮依戀因素、逃避依戀因素及全量表的內部一致性係數分別為 .85、 .79、 .84，間隔兩週的再測信度分別為 .85、 .79、 .85，具有良好的內部一致性與時間的穩定性。此外，焦慮依戀與逃避依戀分別跟自尊、憂鬱及焦慮有顯著相關，其中，焦慮依戀可顯著預測兩週後的憂鬱與焦慮分數，具有良好的建構效度。兩因素也分別跟中文版人際關係問卷的兩種人際依戀向度與四種人際依戀型態有顯著相關，具有良好的同時性效標關聯效度。最後，本文討論採用焦慮依戀與逃避依戀兩種人際依戀向度在探索相關心理適應研究上的重要性。

The present study aimed to suitably refine the broadly used, dimension-based measurement of adult attachment in Taiwan, i.e., the revised adult attachment scale (RAAS). In order to examine the psychometric properties and applicability of its Chinese translated version (RAAS-C), two samples of college students were recruited. Sample 1 (n = 177) was used to explore the potential factor structure via exploratory factor analyses. Sample 2 (n = 240) was used to verify its factor structure via confirmatory factor analyses and to examine the reliability and validity of the scale. Exploratory and confirmatory factor analyses demonstrated that, consistent with the theory and existent adult attachment research, a two-factor structure, i.e., anxiety attachment and avoidance attachment, can better fit for the model of the RAAS-C. The RAAS-C displays good internal consistency and test-retest stability, with Cronbach's alpha coefficients .85, .79, and .84 as well as 2-week test-retest reliabilities .85, .79, and .85 for anxiety attachment, avoidance attachment, and total scale, respectively. It also demonstrates adequate construct validity with correlations to self-esteem, depression, and anxiety, and with anxiety attachment dimension predicting consequent depression and anxiety assessed over 2 weeks. Its criterion-related validity is supported by associations with the widely used relationship questionnaire (RQ), in that anxiety and avoidance attachment dimensions separately and adequately associated with two dimensions (i.e., model of self and other) and four categories (i.e., secure, fearful, anxious, and dismissive) of attachment on the RQ. Finally, the importance and applicability of adopting anxiety and avoidance attachment dimensions to explore emotional distress is discussed.