| dc.description.abstract | Stmctural pquation modeling (SEM) is one of mulliYariatc tcclmiqucs that can eslinmtcs a series of interrelated dependence relationships from a number of endogenous ..tind cxogcnou~__,.,xariablcs. as "·ell as latent (unobserved) ,·ariables sintullancously. To cstintatcf their panunctcrs. SEI\tt based on stn1cturc co,·ariancc nultrix. there arc scYcrriis 111cthods can be used as csti1nation n1cthods. na1ncly lllit'\i111u111 lih~liliuuJ. (i\lL). \\L'ighll .. :d l1.;C1~l :::..quan .. :~ (\\'LS). _gl:1n:1ali1:~J h.:a:::.l :::..quan.::::.. (GLS) and unweighted least squares (ULS). The pul)lOSe~ of this paper yto learn l~ ·tncthody in csthnating SEM panunctcrs and to cotnparc their consistency. accuracy and scnsith·ity based on saniplc si1.c and niultinonnality assun1ptio11 of obscrccd yariablcs. Using a full\ crossed design. data \\Crc genernted lor 2 conditions of nonmlity and 5 different sample sit.cs. The result showed that when data arc nonmlly distributed. lvlL and GLS more consistent and accurate then the other methods. | en |
