Based on Baron and Kenny (1986), four conditions for mediating effects were
investigated (Figure 2). Therefore, the structural equation model was re-estimated by
constraining the direct effect of customer satisfaction on behavioral intentions (b21 ¼ 0)
in each group. In the first condition, the independent variable (interaction orientation)
Hypothesized path
Unconstrained
model
x 2(df ¼ 64)
Constrained
model
x 2(df ¼ 65)
x 2
difference(edf ¼ 1)
H1: interaction orientation ! customer
satisfaction (g11) 288.34 303.17 14.83
H2: interaction orientation ! behavioral
intentions (g21) 288.34 385.47 97.13
H3: customer satisfaction ! behavioral
intentions (b21) 288.34 457.04 168.7
Table VI.
Moderating effects of FT
group vs FC group
Hypothesis Path
First-time customers
group
(standardized path
coefficients)
Frequent
customers group
(standardized
path coefficients)
H1 Interaction orientation ! customer
satisfaction
0.45 * * * 0.55 * * *
H2 Interaction orientation ! behavioral
intentions
0.02 0.14 * *
H3 Customer satisfaction ! behavioral
intentions
0.87 * * 0.78 * * *
Note: Significant at: *p , 0.05, * *p , 0.01 and * * *p , 0.001
Table V.
Path results of structural
equation modeling
Service
interaction
orientationaffects the mediator (customer satisfaction). In the second condition, the mediator
(customer satisfaction) affects the dependent variable (behavioral intentions). In the third
condition, the independent variable (interaction orientation) affects the dependent
variable (behavioral intentions). The first three conditions were met in the original
structural equation model, given the significant impact of interaction orientation on