Table 1. Model estimates for different...

Table 1.

Model estimates for different configurations of the ECSI model

Path relationship	PLS-SEM with cusl2	Equal weights with cusl2	PLS-SEM without cusl2	Equal weights without cusl2	PLS-SEM with random indicators*	Equal weights with random indicators*
Complaints → loyalty	0.067	0.088	0.058	0.054	0.064	0.179
Expectation → quality	0.557	0.553	0.557	0.553	0.557	0.553
Expectation → satisfaction	0.063	0.076	0.063	0.076	0.061	0.078
Expectation → value	0.050	0.062	0.050	0.062	0.050	0.062
Image → expectation	0.505	0.508	0.505	0.508	0.505	0.508
Image → loyalty	0.196	0.189	0.195	0.206	0.194	0.354
Image → satisfaction	0.179	0.172	0.179	0.172	0.179	0.141
Quality → satisfaction	0.512	0.513	0.512	0.513	0.510	0.392
Quality → value	0.558	0.538	0.558	0.538	0.558	0.538
Satisfaction → complaints	0.528	0.519	0.528	0.519	0.530	0.318
Satisfaction → loyalty	0.485	0.406	0.489	0.464	0.489	0.133
Value → satisfaction	0.195	0.187	0.195	0.187	0.198	0.096
R² (loyalty)	0.457	0.365	0.454	0.427	0.458	0.299

Path relationship	PLS-SEM with cusl2	Equal weights with cusl2	PLS-SEM without cusl2	Equal weights without cusl2	PLS-SEM with random indicators*	Equal weights with random indicators*
Complaints → loyalty	0.067	0.088	0.058	0.054	0.064	0.179
Expectation → quality	0.557	0.553	0.557	0.553	0.557	0.553
Expectation → satisfaction	0.063	0.076	0.063	0.076	0.061	0.078
Expectation → value	0.050	0.062	0.050	0.062	0.050	0.062
Image → expectation	0.505	0.508	0.505	0.508	0.505	0.508
Image → loyalty	0.196	0.189	0.195	0.206	0.194	0.354
Image → satisfaction	0.179	0.172	0.179	0.172	0.179	0.141
Quality → satisfaction	0.512	0.513	0.512	0.513	0.510	0.392
Quality → value	0.558	0.538	0.558	0.538	0.558	0.538
Satisfaction → complaints	0.528	0.519	0.528	0.519	0.530	0.318
Satisfaction → loyalty	0.485	0.406	0.489	0.464	0.489	0.133
Value → satisfaction	0.195	0.187	0.195	0.187	0.198	0.096
R² (loyalty)	0.457	0.365	0.454	0.427	0.458	0.299

Notes:

The grey shaded rows highlight relationships with Loyalty that demonstrate particularly strong variations across different model estimations. *This demonstration assigns three additional randomly generated indicators to the Satisfaction construct

Source: Authors’ own work

[ViewLarge]