Convergent validity was established from a review of the t-test for the factor loading. All the t-test are significant, effectively indicating that all indicators measure the same construct. Table 2 show that all 0f the t-values are greater than 2.0 and significant at p<0.01. Moreover, each item’s co-efficient is greater than twice is standard error indicating convergent validity.
In addition to assessing the reliability of the individual indicators, the composite reliability value for each latent variable was calculate using Fornell and Larcker’s measure of composite reliability. Values greater than 0.6 are desirable. The values composite reliability index are shoe in Table 2 and support construct reliability for the variables.
A complementary measure to composite reliability is the average variance extracted, the average variance extracted show directly the amount of variance that is captured by the construct in relation to the amount of variance due to measurement error. Values of 0.50 and above are desirable (Fornell and Larcker, 1981). The results indicated in Table 2 provide additional confidence in the operationalisation of the variables.
Discriminant validity was assessed by determining the confidence interval around the correlation for each pair of factors. The confidence interval is equal to plus or minus two standard errors of the respective correlation co-effcient. If the confidence interval does not include 1.0 then discriminant validity is demonstrated (Anderson and Gerbing, 1998). The test indicated a high level of discriminant validity for the measure. Another test for discriminant validity is to examine if the correlations between measures of different factors using the same method of measurement are lower than the reliability coefficients (Crocker and Algina, 1986). The correlation co-efficients for all measures (except INN3), suggesting that the measures have discriminant validity
การแปล กรุณารอสักครู่..