Travel Reference
In-Depth Information
such as German and also France, which constitute
about 35% of the total respondents.
matrix of latent variables was examined. The cor-
relations between each latent variables were lower
than 0.80, a cut-off point for discriminant valid-
ity, which in this measurement model it indicates
discriminant validity was established (Yanaman-
dram & White, 2006). In conclusion, it is reason-
able to claim that all the measures used in this
study possess adequate psychometric properties.
The measures of the proposed constructs have
achieved satisfactory reliability, convergent, and
discriminant validity.
4.2
Measurement model analysis
The first stage in the two-phase of data analysis
is the estimation of the measurement model. One
measurement model analysis was conducted to
assess the measurement properties of the latent
constructs. The measurement model consists of
two-factor structure of self-congruity, four-factor
structure of experiential value and two-factor
structure of behavioral intention. The measure-
ment model shows acceptable fit:
CFA: Chi-square
4.4
Structural model test
The structural model was tested to assess the
hypothesized structural relationships of the four
constructs (Refer to Figure 1). The results revealed
that the structural model has a significant
=
1093, df
=
426, (
χ
2/df
=
2.5),
p
.001, root mean square error of approximation
(RMSEA
<
=
.074), goodness of fit index (GFI
=
0.8), normed fit index (NFI
=
0.81), and compara-
χ
2 value
tive fit index (CFI
=
0.87).
(
.001) indicat-
ing inadequate fit of the data with the hypoth-
esized model. Based on the suggestion by Hair
et al. (1998), reliance on the chi-square test as the
sole measure of fit in not recommended due to
its sensitivity to sample size. Hence, alternative fit
indices were used as the test for model fit. Based
on the result of other fit indices (RMSEA
χ
2
=
558, df
=
181,
χ
2/df
=
3.08, p
<
4.3
Validity
The assessment of the Cronbach alpha and the
composite reliability has revealed good internal
consistency and reliability. All the constructs
have values exceeded the minimum benchmark
of 0.70. (Hair, et al., 2003). The average variance
extracted (AVE), which assesses the amount of
variance captured by the constructs' measures rel-
ative to measurement error and the correlations
among the latent constructs in the model has
values range from 0.67-0.86. The values exceeded
the minimum recommended level of 0.50, which
indicates the convergent validity is achieved.
Table 1 summarizes the result of the measure-
ment model. The reliability of the variables was
also achieved composite reliability value greater
than the threshold value of 0.60 (Bagozzi & Yi,
1988) and Cronbach alpha greater than the cut-
off value of 0.70 (Anderson & Gerbing, 1988).
To test for discriminant validity, the correlation
=
.06,
GFI
0.96), it was
shown that the model fits the data satisfactorily.
Hence, the study's attempt to establish a plausible
model that has statistical and explanatory power,
which could permit confident interpretation of
results, was thus fulfilled. Therefore, the model
was accepted as reasonable and the hypothesis
tests were interpreted. Table 1 illustrates in detail
the results of the hypothesized model.
=
0.91, NFI
=
0.94, and CFI
=
5
DISCUSSION AND CONCLUSION
The findings suggest that three of the hypoth-
eses in the study were supported by the data. The
hypothesis test results indicated that actual self-
congruity and ideal self-congruity has a significant
relationship with experiential value. The two con-
structs were found to have no effect on behavioral
intention. In addition, experiential value also has a
significant relationship with behavioral intention.
In addition, actual self-congruity has a stronger
effect on experiential value. Therefore, in the cur-
rent study, experiential value not only has a direct
effect on behavioral intention, but also mediates the
actual self-congruity as well as ideal self-congruity
and behavioral intention relationship. Hence, this
result supports previous studies, which found that
there is no direct relationship between self-con-
gruity and consumer behavior. For example, sev-
eral studied have found that such relationship was
mediated by satisfaction (Jamal & Goode, 2001;
Table 1.
Results of tested hypotheses.
Search WWH ::
Custom Search