This shows you the differences between two versions of the page.

Link to this comparison view

Next revision
Previous revision
structural_equation_modeling [2011/04/12 07:07]
Charles Chen created
structural_equation_modeling [2016/10/22 20:12] (current)
Line 1: Line 1:
 == Structural Equation Modeling [SEM, 結構方程模式]== == Structural Equation Modeling [SEM, 結構方程模式]==
-Structural Equation Modeling (SEM) 結構方程模式是一種統計技術 
-is statistical technique for testing ​and estimating causal relations ​using a combination of statistical ​data and qualitative causal assumptionsThis definition ​of SEM was articulated by the geneticist Sewall Wright (1921)[1], the economist Trygve Haavelmo (1943and the cognitive scientist Herbert Simon (1953)[2], and formally defined by Judea Pearl (2000) using a calculus ​of counterfactuals.[3]+**結構方程模式(Structural Equation Modeling, SEM)**是一種結合了質性因果假設與統計資料,進行測試與估計因果關係的統計方法。 
 +結構方程模式也能根據給定模型,就現有資料歸納出理論模型間的參數評估值。通常,初期假設模型需要透過證據作修改。當**結構方程模式**用於純粹的探索研究時,通常在探索心理評量因素分析的脈絡下。 \\ (With an initial theory SEM can be used inductively by specifying ​corresponding model and using data to estimate the values of free parametersOften the initial hypothesis requires adjustment in light of model evidence. When SEM is used purely for explorationthis is usually in the context of exploratory factor analysis as in psychometric design.) 
 +This allows ​the modeler to explicitly capture the unreliability of measurement in the modelwhich in theory allows the structural relations between latent variables to be accurately estimated. Factor analysis, path analysis ​and regression all represent special cases of SEM. 
 +[Translate from [[wp>​Structural Equation Modeling]](2011-Apr-12)]