The 146 dataset were coded and saved into SPSS version 16 and analyzed using AMOS version 7.0. During the process of data screening for outliers, three dataset were deleted due to Mahalanobis (D2) values more than the χ2 value (χ2=78.75; n=44, p<.001) leaving a final 143 dataset to be analyzed. We also conducted univariate normality computations using z-scores of skewness statistics and standard error of skewness as well as kurtosis statistics (see Table 3). Z-score skewness of more than 2 needs to be transformed since it is considered as non-normal data (Hair et al., 2006). Thus, the z-score skewness values highlighted in grey were transformed using reflect function for negative values followed by square root function and ending in reflect function. For positive z-values, reflect function was omitted (Tabachnik & Fidell, 2007). The transformed variables were then used in subsequent CFA and structural models
การแปล กรุณารอสักครู่..
