CFA
1. CFA only without deleting item, using AMOS version 20
2. CFA only with deleting item 10, 9, 15 based on high modification indices value,
using AMOS version 20
(some researchers use this method but I suppose it’s not recommended because it could
have good value in one type of index (eg. CFI) but not in others (eg. RMSEA).
3. EFA to CFA (CFA run based on EFA result)
EFA
Exploratory factor analysis was conducted to select the appropriate number of extracted factor. The
principal component analysis (PCA) was performed to explore the variance of all components or factors. the eigen
value was set to be greater than 1. According to theKaiser’s stopping rule, only the number of factors with
eigenvalues over 1.00 should beconsidered in the analysis (Brown, 2009). The Varimax rotation was operated with
maximum iteration of convergence of 25.
RESULT
The KMO value of .7 (sig. < .05) showed that the number of sample was adequate.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. .746
Bartlett's Test of Sphericity Approx. Chi-Square 943.359
df 105
Sig. .000
The table of total variance showed that four factors had an eigen value of 3.920, 1.726, 1.409 and 1.170for factor
one, factor two, factor three and factor four respectively. The four factors explained 54.832% of the total variance.
Total Variance Explained
Extraction Sums of Squared Rotation Sums of Squared
Initial Eigenvalues Loadings Loadings
Compon % of Cumulativ % of Cumulativ % of Cumulativ
ent Total Variance e% Total Variance e% Total Variance e%
1 3.920 26.132 26.132 3.920 26.132 26.132 2.760 18.399 18.399
2 1.726 11.506 37.638 1.726 11.506 37.638 2.126 14.176 32.575
3 1.409 9.394 47.031 1.409 9.394 47.031 1.675 11.169 43.744
4 1.170 7.801 54.832 1.170 7.801 54.832 1.663 11.088 54.832
5 .993 6.617 61.449
6 .950 6.333 67.782
7 .788 5.253 73.036
8 .775 5.165 78.200
9 .644 4.290 82.491
10 .553 3.686 86.177
11 .536 3.576 89.753
12 .470 3.133 92.886
13 .411 2.742 95.628
14 .364 2.427 98.056
15 .292 1.944 100.000
Extraction Method: Principal Component Analysis.
According to the rotated component matrix, item 8 and 11 have low factor loading (<.5), whereas
item 4 and 14 have cross loading. Thus, the factor with either low factor loading or cross loading
was considered to delete.
Rotated Component Matrixa
Component
1 2 3 4
Item3 .755
Item15 .721
Item14 .696 .338
Item6 .667
Item2 .533
Item13 .763
Item1 .713
Item5 .703
Item4 .331 .509
Item11 .304 .350 .344
Item10 .837
Item7 .786
Item9 .779
Item12 .758
Item8 .453
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 5 iterations.
CFA based on EFA with no deleting item
CFA based on EFA with deleting item 8 and 11 because of low factor loading
CFA based on EFA with deleting item 8 and 11 because of low factor loading and item 4
and 14 because of cross loading
Table Result of models’ fitness tests
Goodness of fit indices
Factor
RMSEA NFI CFI
CFA, no EFA & no deleting item .1 .6 .7
CFA, deleting item 5 factors 10, 9, 15
.1 1.0 1.0
based on high modification indices value
CFA based on EFA with no deleting item .07 .7 .8
CFA based on EFA with deleting item 8
.08 .8 .9
and 11 because of low factor loading
CFA based on EFA with deleting item 8
and 11 because of low factor loading and .07 .8 .9
item 4 and 14 because of cross loading
Reference < .08 > .9 ≥ .9
Note. NFI = Normed fit index; χ2:df = ratio of chi square to degree of freedom; RMSEA = root mean
square error of approximation; CFI = comparative fit index;
