Three endogenous variables (i.e. EI, AT, and AU) were tested in the model II. Similar to themodel I, more than half the variance in EI (53.7%) is explained by PU, PEOU, and SN. Thismodel also explains a substantial proportion of the variance in AT and AU (40% and 25.4%,respectively). The second model exhibited a good fit with the data; x2 (450) 1⁄4 1145.112, p 1⁄4.000, x2/df 1⁄4 2.545, RMSEA 1⁄4 0.062, CFI 1⁄4 0.917, IFI 1⁄4 0.918, and TLI 1⁄4 0. 909. As maybe seen in Figure 4, the six hypotheses were supported in the model II. With respect to parameterestimates shown in Figure 4, there were six hypotheses (i.e. H1b, H3b, H7b, H8b, H10b, H11b)that reached significance. Specifically, the parameter between SN and AU was significant (b 1⁄40.219, CR 1⁄4 2.467, p 1⁄4 .014). In addition, the parameter between PU and AT was significant (b1⁄4 0.240, CR 1⁄4 3.635, p 1⁄4 .000) as was the parameter between SN and AT (b 1⁄4 0.480, CR 1⁄48.145, p 1⁄4 .000). The parameters between PU and EI (b 1⁄4 0.167, CR 1⁄4 2.951, p 1⁄4 .000),between SN and EI (b 1⁄4 0.650, CR 1⁄4 11.154, p 1⁄4 .000), and between AU and AT (b 1⁄40.329, CR 1⁄4 4.580, p 1⁄4 .000) were significant.