Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
317
Structural equation modeling (SEM) is a technique for simultaneously estimating the relationships between 
observed and latent variables (the measurement model), and the relationships among latent variables (the structural 
model). 
Table 2 
The Result of a Confirmatory Factor Analysis 
Factor 
X2
df 
p-value 
GFI 
AGFI 
RMR 
t-value (C.R.) 
Schema plus tag 43.282
5 
0.000
0.924 
0.773 
0.035 
p<0.001 
Subtyping 
21.967
5 
0.001
0.962 
0.887 
0.055 
p<0.001 
Attitude
67.65 
5 
0.000 
0.890 0.671 0.105 p<0.001 
Image 
Differentiation 
Positioning 
Behavior 82.250
5 0.000 0.863 0.588 0.124 p<0.001 
 
SEM is a method that has gained popularity because it combines Confirmatory Factor Analysis (CFA) and 
regression analysis to model a variety of psychological, sociological, and other relationships (Hoyle & Smith, 1994; 
Jorskog & Sorbom, 1989). Table 6 showed that “positioning differentiating image” has x
2 
of 106.558, degrees of 
freedom of 65, p-value of 0.001, the goodness-of-fit index (GFI) of .962, the “adjusted” goodness-of-fit index 
(AGFI) of .876, the Normed Fit Index (NFI) of .977, and the Root Mean Square Residual (RMR) of .047. The result 
of t-test for Hypothesis 1 proved that image differentiation positioning of K airline, 
Cabin Service Facility
(p=.001), 
Cabin Entertainment System
(p=.000), and 
Crew Uniform
(p=.000) differently influence the schematic process; 
“
Schema plus tag
” and “
Subtyping
.” These three attributes more influenced on “
Subtyping
” than “
Schema plus tag
” 
(see Table 3). Table 4 shows the Path coefficient of Final model of the study through SEM.