學前兒童社會行為評量系統（教師評量表）之信度、效度及性別恆等性驗證

The Reliability, Validity and Measurement Invariance of the Social Behavior Assessment System for Preschool- Teacher Rating Scale

「學前兒童社會行為評量系統」(教師評量表)適用於4-6歲學前兒童之社會行為評估，本量表內容包括社會能力量表及問題行為量表兩部分，主要功用在及早發現情緒行為問題之學前兒童。本研究主要目的在檢驗教師評量表之信度、效度，以及是否符合性別測量恆等性。研究對象為台灣北、中、南、東四區共431名學前兒童，以結構方程模式進行教師評量表潛在構念之信、效度檢定，另以結構方程模式多群組分析法，進行性別恆等性分析，在測量恆等性分析的階層限制模式依序採用「測量加權」、「結構共變數」、「殘差變異數」。本研究結果發現：1.在信度方面，教師評量表之社會能力量表在個別指標信度介於.57~.89，問題行為量表則介於.38~.90；潛在變項組合信度在社會能力量表介於.80~.89，問題行為量表則介於.79~.87，故本評量表具有適切信度。2.在效度方面，教師評量表下社會能力量表及問題行為量表之測量模式大致符合基本適合度、整體模式適合度與模式內在結構適合度之標準，故本評量表具有適切效度。3.在性別測量恆等性方面，結果發現本評量表在評量不同性別學前兒童社會行為所得之分數具有相同的意義，顯示本評量表具有跨性別的可靠性及有效性。

The main purposes of this research were to evaluate the reliability, validity and measurement equivalence of the Social Behavior Assessment System for Preschool- Teacher Rating Scale (SBASP-TRS). Teachers of 431 preschool children ages 4 to 6 were included in this study. The data gathered were processed by Structural Equation Modeling. The results of this research are stated as follows: (a) By utilizing Structural Equation Modeling to examine the theoretical model, we found that the goodness of fit was good. (b) The SBASP-TRS provided good quality of reliability and validity. (c) The measurement equivalence of the SBASP-TRS was good by the test of multi-group analysis. The findings showed that the invariance had significant effects on interpretation based on gender latent mean difference as well as observed mean difference. According to the results, this study suggests the practical use of SBASP-TRS and recommends for future research.