PEMODELAN MULTIGROUP STRUCTURAL EQUATION MENGGUNAKAN ESTIMASI MAKSIMASI LIKELIHOOD DENGAN ITERASI FISHER SCORING; MULTIGROUP STRUCTURAL EQUATION MODELLING USING MAXIMUM LIKELIHOOD ESTIMATION WITH FISHER SCORING ITERATION

Saputra, Trisna H, Dedi Rosadi

2015 | Skripsi | FMIPA

Traditionally the analysis of Structural Equation Model (SEM) is often treats all variables as continuous variables. Some researchers say that the ordinal variables should be treated as ordinal variables, not as a continuous variable because it will be obtained unappropriate results. Multigroup SEM is an analytical method that uses the application of the SEM sample groups, where each group will build a path to the beginning of the same diagram. Maximum Likelihood choosen because the samples used in the study is quite small campared with other estimates. To maximize the function of parameters themselves need to use numerical method with Fisher Scoring iteration. Object of the research is Student Faculty of FMIPA UNY that used IM3 operator. Researcher suspects that there is difference in customer satisfaction with group who use companion card other IM3 than group who doesn?t use a companion card. It is very interesting to be reviewed because of the companion card usage can be an effect for dissatisfaction with the facilities provided by IM3 operator

Kata Kunci : SEM; Multigroup SEM; Path Analysis; Maximum Likelihood; Fisher Scoring; Customer Satisfaction Index.