Overall measurement quality was assessed using confirmatory factor analysis (Anderson and Gerbing, 1991). Although measurement quality is sometimes assessed factor, each multiple-item indicator was considered simultaneously to provide for fullest test of convergent and discriminate validity. A number of indices were used to determine if the fit the data to the model is adequate. In addition, it was desired that all of the indicator variables for each factor in the measurement model have a t-statistic of 2.0 or greater. Also it was important that no standard errors associated with the t-statistics be near zero (such as 0.0003).
As can be seen from Tabel 2, all loadings exceed 0.6 and each indicator’s t-value exceeds 10.0 (p<0.001) Standard errors were 0.05 and above. Co-efficient alpha is above 0.8 for each scale, exceeding the recommended cut-off point of 0.7 suggested by Nunnally (1998). The measurement model. The x2 statistic is 344.21 with 234 degrees of freedom (p<0.01),something which suggest discrepancies between the data and the proposed overall model. However, this may be attributed to sample size and the highly restrictive assumption of perfect model fit in the population. The x2 to degree of freedom statistic is 1.47. A number of goodness-of-fit indices were used to assess practical significance. The LISREL 8.52 output showed an adequate model fit as
การแปล กรุณารอสักครู่..