Confirmatory factor analysis. Confirmatory factor analysis (CFA) was performed to
validate scale dimensionality of the remaining EFA items. Several goodness-of-fit
indices were used to report convergent and discriminant validity, as well as the
reliability properties of the measures to identify internal consistency and adequate fit of scale items. To assess the fit of the observed variables with the latent variables,
overall model fit indices were evaluated. CFA cannot be performed on constructs which contain fewer than three indicators. Therefore, the brand effect result is not presented. The 14 CSR items revealed four dimensions, according to the EFA results. The remaining 14 items of CSR did not fit well due to a large RMSEA value and x 2/df ratio. One item, csrexp 15, with the lowest squared multiple correlations (SMC) was excluded. The remaining ten-item battery measuring the PSQ construct yielded three dimensions based on the EFA results. The ten-item model indicated a poor fit with a RMSEA value marginally above the 0.08 threshold, and a large x 2/df ratio. The item “convenient location” (psqual8) with the lowest SMC was eliminated which resulted in improved goodness-of-fit indices. The two factors of the trust scale revealed a good fit. A
summary of the CFA results is presented in Table II.