The Relationship between SCI and BI as Mediated by TOG and PC

The conceptual model of this study hypothesized the indirect relationships between SI and BI. One of the indirect paths between them was discussed above, and the other path will be discussed in this section.

This section discusses the relationship between SCI and BI with the mediated effect of two mediating variables that are TOG and PC. Table 4.76 shows the results of the mediation analysis using PROCESS macro tool. This time Model 6 by Hayes (2012) was adopted to run the analysis. This model studies the effect of two mediating variables on the relationship between the independent and independent variable.

The tested model in this analysis shows the relationships between the variables following this path:

SCI (independent variable) → TOG (first mediator) → PC (second mediator) → BI (dependent variable) The model in Table 4.76 shows SCI predicting TOG. It explains the relationship between SCI (independent variable) and TOG (the first mediating variable). In this table, R square shows that 2.63%

variation in TOG is explained by SCI. The F-test is highly significant, F (1, 225) = 6.0321, p=0.0148. This means that the independent variables SCI is statistically significant in predicting TOG. The regression

PC

BI TOG

0.6486* 0.4632

*

0.4632 (with PC) 0.4233* (without PC)

125 analysis also confirms significant effect of SCI on TOG at a significance level of p=0.0148 and the estimated coefficient of this model is 0.1619.

Table 4.76: The Relationship between SCI and BI as Mediated by TOG and PC (Model 1) Outcome: TOG

R R-Square MSE F df1 df2 sig

0.1623 0.0263 0.9780 6.0321 1 225 0.0148

Model 1 B SE t Sig. LLCI ULCI

Constant 0.0003 0.0659 0.0039 0.9969 -0.1297 0.1302

SCI 0.1619 0.0659 2.4560 0.0148 0.0320 0.2919

Table 4.77 shows the results of model 2, which explains the relationship between SCI and PC (the second mediating variable with the mediated effect of TOG. R square shows that 42.44% variation in PC is explained by SCI and TOG. The findings of this table also show that the F-test is highly significant, F(2,224)

= 81.8551, p<0.001. This indicates that the independent variable SCI is statistically significant in predicting PC with the mediated effect of TOG. The results of regression analysis in model 2 show that TOG has a significant influence on PC at a significance level of p<0.001 and an estimated coefficient of 0.6381, while SCI with the mediated effect of TOG has an insignificant influence on PC.

Table 4.77: The Relationship between SCI and BI as Mediated by TOG and PC (Model 2) Outcome: PC

R R-Square MSE F df1 df2 sig

0.6515 0.4244 0.5811 81.8551 2 224 0.000

Model 2 B SE t Sig. LLCI ULCI

Constant -0.0072 0.0508 -0.1419 0.8873 -0.1074 0.0929

TOG 0.6381 0.0516 12.3625 0.0000 0.5364 0.7398

SCI 0.0643 0.0515 1.2483 0.2132 -0.0372 0.1658

126 Table 4.78 shows the results of model 3, the model of the indirect effect. This model explains the relationship between SCI and BI with the mediated effect of TOG and PC. R square shows that 31.63%

variation in BI is explained by SCI, TOG, and PC. The findings of this table also show that the F-test is highly significant, F (3, 221) = 34.0792, p<0.001. This indicates that the independent variable SCI is statistically significant in predicting BI with the mediated effect of TOG and PC. The results of the regression analysis in model 3 show that TOG has an insignificant influence on BI, p=0.1259. On the other hand, PC has a significant influence on BI at level p<0.001 and the estimated coefficient is 0.4509.

SCI also has a significant influence on BI at a significance level p=0.1133, the estimated coefficient is 0.1133. These results mean that TOG does not predict BI, but PC predict BI.

Table 4.78: The Relationship between SCI and BI as Mediated by TOG and PC (Model 3) Outcome: BI

R R-Square MSE F df1 df2 sig

0.5624 0.3163 0.6925 34.0792 3 221 0.000

Model 2 B SE t Sig. LLCI ULCI

Constant 0.0045 0.0555 0.0815 0.9351 -0.1048 0.1139

TOG 0.1125 0.0732 1.5361 0.1259 -0.0318 0.2568

PC 0.4509 0.0733 6.1538 0.0000 0.3065 0.5953

SCI 0.1133 0.0564 2.0077 0.0459 0.0021 0.2245

Table 4.79 shows the regression analysis of the full model. This model explains the direct relationship between SCI and BI without the influence of the mediating variables. In this table, R square shows that 4.31% variation in BI is explained by SCI. The F-test is highly significant, F(1, 225)=10.0434, p=0.0017.

The result of regression analysis in Table 4.79 shows that SCI significantly influence BI at significant level of p =0.0017 and the estimated coefficient of this model is 0.2071.

127 Table 4.79: The Relationship between SCI and BI as Mediated by TOG and PC (Total Effect

Model) Outcome: BI

Model Summary R R-Square MSE F df1 df2 sig

0.2076 0.0431 0.9605 10.0434 1 225 0.0017

Model B SE t Sig. LLCI ULCI

Constant 0.0014 0.0653 0.0211 0.9832 -0.1274 0.1301

SCI 0.2071 0.0653 3.1691 0.0017 0.0783 0.3358

Figure 4.26 illustrates the standardized regression coefficients between SCI and TOG, TOG and PC, PC and BI, as well as SCI and BI. These relationships were statistically significant. Table 4.80 shows the indirect effect between SCI and BI. Model 1 in this table explains the relationship between SCI and BI with single mediation that is TOG. The result of model 1 shows that zero lies within the bootstrapped confidence intervals range from -0.0017 to 0.0658. This means that the indirect relationship of this model does not exist. In other words, there is no mediation affect the relationship between SCI and BI. Thus, we can conclude that the indirect effect of TOG is insignificant. As for model 2 in Table 4.80, it explains the relationship between SCI and BI with double mediators that are TOG and PC. The results of the bootstrapped confidence intervals show that zero does not occur between the LL and UL of the confidence interval since the confidence intervals range from 0.0078 to 0.1106. This means that the indirect effect of TOG and PC is significant. Model 3 shows the relationship between SCI and BI with the mediated effect of the second mediating variable (PC). This path has not been proposed in the conceptual model of this study; however, including this path was necessary in this test for running this analysis. The result of model 3 shows that bootstrapped confidence intervals range from -0.0017 to 0.0658, which means that that zero lies within this range. This means that the indirect relationship of this model does not exist. Thus, we can conclude that the mediated effect of PC is insignificant.

Table 4.80: The Indirect Effect of SCI on BI Effect Boot SE BootLLCI BootULCI

Total: 0.0938 0.0472 0.0137 0.2016

Model 1 : 0.0182 0.0161 -0.0017 0.0658

Model 2 : 0.0466 0.0251 0.0078 0.1106

Model 3 : 0.0290 0.0259 -0.0124 0.0917

128 Since the indirect effect of SCI on BI as mediated by TOG and PC is statistically significant, but insignificant with single mediation. We can conclude that the mediation has occurred. This means that the path of the indirect relation between SCI and BI with the mediated effect of TOG and PC is stronger than the indirect relation with a single mediator. This result indicates that TOG and PC together are strong mediators of the relationship between SCI and BI.

Figure 4.26: The Coefficient for the Relationship between SCI and BI as Mediated by TOG and PC and PC