ABSTRACT

The measurement model constructed for CFA had a relative chi-square value of 2.325 < 3, an incremental fit index (IFI) of 0.946 > 0.9 and a comparative fit index (CFI) of 0.916 > 0.9. The standardized loadings (λ) for all constructs were high (λ > 0.5) and the corresponding t-values were statistically significant. These results indicated uni-dimensionality among the research constructs. A check of the modification indices for the measurement model conducted during the CFA process revealed no significant cross loadings among the variables (λ > 0.85), which indicated good discriminant validity (Kline 1998). The scores for valid variable items in each construct were then averaged as a single score to be used in the model analysis.