Factors Influencing Clothing Interest and Purchase Intention: a study of Consumers in Generation y from Malaysia and Singapore
tải về 475.41 Kb. Chế độ xem pdf
|
|
Display Table 1 here. Convergent and Discriminant Validity of the Constructs Confirmatory factor analysis was used to examine measurement model fit and the convergent and discriminant validity of the data (Hair et al. 2010). Criteria used to evaluate measurement model fit include the following indices: (1) Goodness of Fit (GFI), (2) Comparative Fit Index (CFI), (3) Root Mean Square Error of Approximation (RMSEA), (4) Tucker-Lewis index (TLI), (5) Parsimony Normed Fit Index (PNFI), and (6) Normed Chi- square (χ²/df). A model is considered fit when GFI exceeds 0.90, CFI exceeds 0.90, RMSEA is less than 0.08, TLI exceeds 0.90, PNFI is more than 0.50, and χ²/df is less than 3.0 (Hair et al., 2010). As for the present study, GFI was 0.908, CFI was 0.913, RMSEA was 0.049, TLI was 0.918, PNFI was 0.723, and χ²/df = 1.818 (χ2 = 760.044; df = 418), indicating that the measurement model was fit. Three conditions must be met in order to establish the convergent validity of the data (Hair et al. 2010). The three conditions are: (1) the factor loading each construct item must be at least 0.60, (2) the average variance extracted (AVE) of each construct should be larger than 12 0.50, and (3) the composite reliability of each construct should be greater than 0.70. As seen in Table 2, the factor loadings of the construct items range from 0.627 to 0.933, the AVE of each construct exceeds the cut-off point of 0.5, and each construct has a composite reliability of more than 0.70. Hence, the convergent validity of the data in the present study was established. Display Table 2 here. The discriminant validity of the data was assessed based on the guidelines suggested by Fornell and Larcker (1981). Discriminant validity is said to be established when the squared root of AVE for each of the constructs exceeds the variance shared between any two constructs. As shown in Table 3, all the shared variance between constructs (the off diagonal entries) are lesser than the squared root of AVE in the diagonal entries; hence, the discriminant validity was established. In sum, all constructs included in this study have demonstrated adequate reliability, convergent validity, and discriminant validity Display Table 3 here. The researchers also addressed the issue of common method variance (CMV) based on the suggestion by Malhotra et al. (2006). It is argued that CMV will generate a false internal consistency and correlation among the variables created by their common source (Chang et al., 2010; Malhotra et al., 2006). It is reported that common method bias may exist if the hypothesized model (with indicators for all the constructs to be modeled as a single factor) for CMV has a good model fit (Malhotra et al., 2006). As for the present study, the results showed that the hypothesized model based on the suggestion by Malhotra et al. (2006) is not fit; hence, common method bias is not a significant problem. tải về 475.41 Kb. Chia sẻ với bạn bè của bạn:
|