The constrained model (SEM2:1b, Figure 7 - 8 ) yielded a satisfactory model fit with
χ^2 = 53.494, d.f. = 41, sig = 0.091; RMSEA = 0.038; CFI = 0.991; and SRMR =
0.0746 and all structural coefficients were statistically significant.
Figure 7-8 Constrained model SEM 2:1b
The chi square estimates for the constrained model were then compared with
those from the hypothesised model (SEM 2:1, Figure 7 - 7 ). The chi square
difference (Δχ^2 ) between these two models is 9.479 (53.494 - 44.015) and with 1
d.f. (degree of freedom) - this yields a p value of 0.0021. It can be concluded
from this finding that, because the global model fit for SEM 2:1 (the hypothesised
model containing the direct ML→JP path) is statistically significantly better
(because the p value for the χ^2 difference test is <0.01) than the constrained
model (SEM 2:1b), that full mediation of the ML→JP path by ME is not supported.
To assess whether or not Work Meaning (ME) has a partial mediation effect on the
ML→JP relationship it is necessary to estimate a third model in which there is no
indirect effect (ME→JP) on Job Performance (JP). This model is illustrated in
Figure 7 - 9 along with the estimates for the structural coefficients and the global
fit measures.
The structural coefficient values for the ML→JP path are now compared between
model SEM 2:1c and the hypothesised model SEM 2:1. In the constrained model
(SEM 2c) the value (effect size) for the ML→JP path is 0.422 while in the
hypothesised model (SEM 2) this value is 0.271. When comparing the originally
specified model (2:1) with Model 2:1c, because the value of ML→JP in 2:1 is