The construct validity of the research instrument was assessed
via confirmatory factor analysis (CFA). To perform a CFA, all of the
constructs and reflective indicators were depicted and composed
as a measurement model in which all constructs were allowed to
correlate with each other.
Generally speaking, the process of validation comprises the
steps as described below: the measurement model should demonstrate
robustness for the empirical data and meet the requirements
of the certain indexes. For example, chi-square normalized by
degrees of freedom (k/df), should be less than five (Bentler, 1989)
and it is 780/296 = 2.64 here; and comparative fit index (CFI)
should all exceed 0.9, and root mean square error (RMSEA) should
be less than 0.10 (Henry & Stone, 1994) and it was 0.92 and .065
here respectively. It suggests an adequate model fit for the empirical
data. In sum, the indices indicate that the model, which is a
representation of the HBM, can be accepted from an empirical
point of view and its robustness (for more details see: Hu and
Bentler (1999) and Kaiser and Scheuthle (2003)).
Furthermore, it is important to say that the v2-statistic generally
was significant in our models. Note that the v2-statistic is
affected by sample size, which is quite large in this study
(N = 389). Given the large sample size, however, significant v2
(Kaiser, 2006; Kaiser & Scheuthle, 2003) is consistent with previous
findings. In the next step, HBM was tested consecutively. The
results of SEM revealed that the (standardized) path coefficients
indicate the strengths of relationships between the variables
The construct validity of the research instrument was assessedvia confirmatory factor analysis (CFA). To perform a CFA, all of theconstructs and reflective indicators were depicted and composedas a measurement model in which all constructs were allowed tocorrelate with each other.Generally speaking, the process of validation comprises thesteps as described below: the measurement model should demonstraterobustness for the empirical data and meet the requirementsof the certain indexes. For example, chi-square normalized bydegrees of freedom (k/df), should be less than five (Bentler, 1989)and it is 780/296 = 2.64 here; and comparative fit index (CFI)should all exceed 0.9, and root mean square error (RMSEA) shouldbe less than 0.10 (Henry & Stone, 1994) and it was 0.92 and .065here respectively. It suggests an adequate model fit for the empiricaldata. In sum, the indices indicate that the model, which is arepresentation of the HBM, can be accepted from an empiricalpoint of view and its robustness (for more details see: Hu andBentler (1999) and Kaiser and Scheuthle (2003)).Furthermore, it is important to say that the v2-statistic generallywas significant in our models. Note that the v2-statistic isaffected by sample size, which is quite large in this study(N = 389). Given the large sample size, however, significant v2(Kaiser, 2006; Kaiser & Scheuthle, 2003) is consistent with previousfindings. In the next step, HBM was tested consecutively. Theresults of SEM revealed that the (standardized) path coefficientsindicate the strengths of relationships between the variables
การแปล กรุณารอสักครู่..
