在社會行為科學領域中,有兩個很基本的問題需要去面對。第一個問題是有關測量部份;第二個問題是有關雙數間的因果關系以及他們能解釋的範疇。結構方程模式(Structural Equation Modeling , SEM)正可以有效地解決這些簡題。結構方程模式主在測拭兩類模式。其一為測量模式(measurement model),此模式的目的在於說明問觀察變數(observed variables)是如何地測量潛在變數(latent variables)。其二爲結構模式(structural model),此模式的目的則在於辨認潛在變數之間的因果關係。而線性結構關係(LInear Structural RELations , USREL)屬一個被用來測試結構方程模式的統計程式軟體,亦爲社會行爲科學領域常用之軟體。此文的目的乃是對結構方程式模式及LISREL作一個簡單初略的介紹。
There are two basic problems which need to be solved in the social and behavioral sciences. The first relates to measurement. The second concerns how far causal relationships are accounted for by related variables. The method of structural equation modeling (SEM) can be used to solve these problems. Structural equation modeling includes two models. One is a measurement model that is used to explain how observed variables measure latent variables. The other is a structural model that specifies the causal relationships among latent variables. The statistical program LISREL (LInear Structural RELations) is the most common way for testing structural equation modeling. The purpose of this paper is to introduce structural equation modeling and LISREL.