Civil Engineering Reference
In-Depth Information
Tabl e 6 Summary of
simultaneous multiple
regression analyses for two
factors of KS predicting IM
( N
R 2
Va r i a b l e
B
S E
β
t
change
Constant
1.532
0.116
13.156***
SB
0.340
0.029
0.407
11.775***
0.367
SW
0.268
0.033
0.285
8.241***
0.042
=
967)
*** p
<
0
.
001
PA
0.42*
SW
0.31*
OA
0.49*
0.76*
0.83+
0.30*
SB
0.84*
0.34*
0.71*
EI
KS
0.69*
0.81*
AL
0.45*
0.74*
0.35*
KOC
0.51*
IT
0.75*
0.60*
ES
0.43*
0.70+
0.83*
IM
0.78*
0.81*
IL
0.35*
TW
0.31*
0.79*
0.83*
0.68*
HR
IPR
0.39*
0.38*
TA
0.31*
0.53*
ITI
Fig. 2
The path diagram and relationship of KOC, KS, and IM
factors of KS predicting IM. The outcome shows that both SW and SB are proved to
have significant influences. And SB of KS affects IM more than SW (see Table 6 ).
4. SEM of KOC, KS, and IM is proved to have a good model data-fit.
Figure 2 shows the path diagram of KOC, KS, and IM. All the factor loading
of 14 measurement indicators are larger than 0.68, which shows they can explain
the three latent variables adequately. The next question is whether the model fits the
data or not. We found that all the data-fit criterions except for chi-square(
2 )were
χ
2 ) criterion for model data-fit itself
has a few flaws and limitations, one of which is that the more the samples are, the
satisfied (see Table 7 ). However chi-square(
χ
Search WWH ::




Custom Search