Table 1 Descriptive statistics Notes: **p < 0.01 level. Two-tailed test. The study followed the two step approach of structural equation modelling (SEM) to analyse the data. The second order factor model for the export performance scale was adopted. In this pace, the proposed CFA achieved good fit (y’/df = 1.988; RMSEA = 0.043; GFI = 0.93; AGFI = 0.92; CFI = 0.92; NNFI = 0.93; RMR = 0.08). The study further tested the convergent and discriminant validity of all measures with the CFA results. Convergent validity occurs when the factor loadings and average variance explained (AVE) in items by their respective constructs are greater than 0.50 and factor composite reliability is equal to or greater than 0.60. Factor loadings of all items to their latent variables are over 0.50 thresholds; while the internal consistence reliability estimates (Cronbach’s alphas) are over the threshold of 0.70. Table 1 presents the descriptive statistics of study. All AVE’s are higher than the squared correlations between the focal construct and other constructs, which provides evidence of discriminant validity of the measures. Overall, the fit statistics indicate that the model provided an adequate fit to data for the sampling. The RMSEA equalled 0.042 which is lower than 0.1. In the meantime, the two GFI and AGFI were, respectively 0.953 and 0.934, which show good fitness of the Notes: ***p < 0.001 level. **p <0.01 level. Two-tailed test. Table 2 Results of model testing Table 3 Direct and indirect effect of model
