|
5.5.10 Preliminary Analysis – Perceived Consumer TCs
Seventeen items were used to measure the perceived consumer TCs construct. These items
were subjected to the two-step preliminary data analysis process shown in Figure 5.1, the
EUN1 EUN2 EUN3 EUN4 EUN5 EUN6 EUN7
EUN1
1.00
EUN2
.761
1.00
EUN3
.641 .775 1.00
EUN4
.559 .655 .776 1.00
EUN5
.403 .419 .444 .501 1.00
EUN6
.483 .527 .558 .636 .510 1.00
EUN7
.413 .432 .461 .519 .651 .619 1.00
234
results of which are displayed in Table 5.11. Evaluation of the correlation matrix through the
KMO and Bartlett’s Test results in a high KMO statistic (.937) and a significant probability
level (p< .001) for the Bartlett’s test. These results indicate that sufficient correlations were
found within the correlation matrix for factor analysis to proceed. In addition, bivariate
correlations were inspected and all coefficients fell within the acceptable range for factor
analysis of .30 and .90. EFA was then conducted and found that PRTC1, COTC4 and POTC1
loaded on two factors. After eliminating these three items, EFA then produced a three-factor
solution for the fourteen remaining items explaining 76.95% of the variance, with factor
loadings ranging between .63 and .87 and reliability of .93.
The first factor, composed of seven items, accounted for 35.04% of the variance explained.
The seven items loading on this factor reflected post-TCs with item loadings ranging
from .77 to .90. Cronbach’s alpha for Factor 1 was .94, indicating good reliability of the scale.
Accounting for 22.95% of the variance explained, Factor 2 included four items related to pre-
TCs. Factor loadings of these four items ranged from .84 to .91 and Cronbach’s alpha of .91
indicated an internal reliability. The three items loading on Factor 3 were related to
contemporaneous TCs. The variance explained by this factor was 18.96% with strong item
loading ranging from .86 to .93, and Cronbach’s alpha was .88, exceeding the threshold
of .70 (Nunnally 1978). At this point, as all the remaining fourteen items met the criteria of
Step One, they were retained for CFA analysis in Step Two.
In Step Two (refer Figure 5.1), CFA was conducted on the three-factor structure with
fourteen items retained from Step One. The model fit indices indicated that this measurement
model poorly fit the data (
χ
2
/df = 12.565 > 5, GFI = .873 < .90, RMSEA = .110 > .08, and
RMR =.155> .05).
It was apparent that some modification in specification was needed to
235
improve the model. By investigating modification indices, indicators which were related to
larger reductions of chi-square (Anderson and Gerbing 1988) were identified and eliminated
one by one. After removing four items (PRTC2, POTC2, POTC7 and POTC8), an acceptable
model was achieved. All overall goodness-of-fit statistics were within acceptable ranges. For
example,
χ
2
/df (3.73) was less than 5, GFI (.979), TLI (.984) and CFI (.990) were above .90,
and RMSEA (.053) was less than .08. RMR, however, was computed at .075 and, although it
was above the .05 cut-off, this statistic was considered acceptable given the large sample size
(962). By looking at the CFA fit indices results as a whole, it indicated a reasonable fit for the
modified measurement model. Table 5.11 shows the second-order measurement model in
which perceived consumer TCs was the second-order construct with three first-order
constructs (pre-TCs, contemporaneous TCs, post-TCs) as indicators. The standardized factor
loadings of the three first-order constructs used to measure the second-order construct
(perceived consumer TCs) ranged from .78 to .85, all exceeding the preferable criterion of .70
(Hair
et al.
2006). The AVE for the second-order construct was .66, exceeding the
recommended level of .50 (Hair
et al.
2006). The CR of the second-order construct was .854,
exceeding the threshold value of .70 (Nunnally 1978). More specifically, by inspecting each
first-order construct and its corresponding indicators, it can be found that the standardized
factor loadings of all items were relatively high and significant, ranging between .77 and .92,
and the AVE for the all first-order constructs exceeded the recommended level of .70,
ranging from .70 (pre-TCs) to .77 (contemporaneous TCs). CR of all first-order constructs
exceeded the recommended benchmark of .70, ranging from .874 (pre-TCs) to .925 (post-
TCs). Taken together, this three-factor model was considered to constitute a reliable and valid
measurement scale for perceived consumer TCs, and it was retained for the overall
measurement theory model.
236
Do'stlaringiz bilan baham: |
