Microsoft Word 2007 ichrie conference Proceedings Final-Final 06-06-07. doc

 
2007 Annual International CHRIE Conference & Exposition 
221
Figure 2 
Mean Differences in Service Quality Evaluation Between Avery’s and Devin’s 
Test of Hypothesis for Perceived Value 
To test Hypothesis 3, a series of regression analyses was conducted. Mediation can be explained by 
examining the direct effect of a predictor (service quality) on a mediator (perceived value), the direct effect of the 
mediator on a criterion (customer loyalty), and the direct effect of the predictor and the criterion. The total effect of 
the predictor on the criterion is equal to the indirect effect of the predictor on the criterion (the effect of the predictor 
on the criterion via the mediator) and the direct effect of the predictor and the criterion. Complete mediation occurs 
if there is no direct effect of a predictor on a criterion, or the total effect of a predictor on a criterion is completely 
accounted for by the indirect effect of the predictor on the criterion through a mediator. Partial mediation occurs if 
the total effect of a predictor on a criterion is due to both a direct and indirect effect of the predictor on the criterion.
The regression coefficient is an estimate of the effects. 
The regression analysis showed partial mediation in the customized service setting and the non-customized 
customized service setting (Figure 3). The total effect of service quality on customer loyalty (
β
total
= 0.29306 for 
the customized service setting; 
β
total
= 0.24633 for the non-customized service setting) was larger than the direct 
effect of service quality on customer loyalty (
β
sq
→
loy
= 0.16181 for the customized service setting; 
β
sq
→
loy
= 
0.16651 for the non-customized service setting). When customer loyalty was partitioned into behavioral loyalty and 
affective loyalty, the identical partial mediation was found in both settings.