心理幸福感量表簡式中文版信效度及測量不變性：以大學生為樣本並兼論測量不變性議題

Reliability Validity and Measurement Invariance of the Brief Chinese Version of Psychological Well-Being Scale among College Students

本研究目的有二， 一是獲得更多有關Li（2014）18 題簡式中文版心理幸福感量表的信、效度訊息， 二是進行跨樣本交叉效度及跨時間測量不變性的檢定。研究樣本有三套，以大學生為對象，分別命 名為樣本一、樣本二及共同樣本，人數分別為442、349、394 位參與者。以結構方程式模型技術來 處理資料。研究發現，符合理論架構的六因素斜交模型獲得良好的適配；在信度方面，分量表內部 一致性α 係數在 .73 ~ .87，總量表高達 .91；分量表間隔2 ~ 3 週的再測信度介於 .59 ~ .65，總量表 則達到 .70。在效度方面，因素負荷量在 .65 ~ .92，具有聚斂效度，而因素間的相關性在 .30 ~ .72， 通過區辨效度的檢驗。另外，跨樣本的交叉效度亦獲得一定程度的支持；而測量不變性檢驗亦獲得 因素負荷量及截距在跨時間上符合量尺不變性及純量不變性的要求，有利於構念在跨時間上的因素 間關係比較及因素平均數比較。

The two aims of this research were to obtain reliability and validity information of Li’s (2014) 18-item brief Chinese version of the Psychological Well-Being Scale across samples， and to test its measurement invariance across time. Three samples of college students were recruited in the study. They were named as sample I， sample II， and common sample ， in which 442， 349， and 394 students， respectively. Data were analyzed using structural equation modeling techniques. Results showed that theoretical framework of six-factor oblique model had good model-fits. Cronbach’s α coefficients ranged from .73 to .87 for subscale scores， and was .91 for the total score; 2 ~ 3 weeks test-retest reliability coefficients of subscale scores ranged .59 ~ .65， and was .70 for the total score. Based on the 6-factor model， factor loadings were among .65 ~ .92 providing convergent validity ， and interfactor correlations were ranging .30 ~ .72 offering discriminant validity . Besides， cross validation was supported. Furthermore， metric invariance and scalar invariance existed across groups in the measurement equivalence tests; therefore it ensured that the scale can be compared on its constructs across multiple measurement times ， including comparisons in interfactor correlations and latent means.