significant but reduced in comparison with 2:1c, then we can conclude that partial
mediation of the ML→JP path by ME is supported (Hair et al. 2006 : 867).
Figure 7-9 Constrained model SEM 2:1c
The diagram in Figure 7 - 7 shows only the direct effects amongst the variables.
Because of the inclusion of ME as a mediating variable, a full understanding of the
inter-relationships in the model must also take account of the indirect effects –
i.e. those effects of ML on JP that are ‘transmitted’ through ME.
Calculating the magnitude of the indirect and total effects is a straightforward
task. The indirect effect of ML→JP via ME is calculated as a product of the
ML→ME and ME→JP coefficients:
0.558 * 0.258 = 0.144.
The total effect is simply the sum of the direct and indirect effects:
direct effect (0.271) + indirect effect (0.144) = 0.415.
The total (standardised) effects for Model 2 are illustrated in Table 7 - 13.
Independent
variables
Dependent variables
ME JP DSB
ML 0.558 0.415 0.400
ME - 0.258 -
Table 7-13 Total (standardised) effects for Model
