20 A Dynamic Examination of Online Customer Gratitude and Trust Second, we test Hypotheses 1 to 7 via the conditional latent growth curve model. We find that relationship benefits have the positive impact on the level and velocity of gratitude (β = 0.11 & 0.14), and that relationship benefits also have the positive impact on the level and velocity of trust (β = 0.11 & 0.12). Gratitude velocity is positively and significantly related to trust velocity (β = 0.13). The initial levels of gratitude and trust have the positive and significant effect on the initial level of customer loyalty (both attitudinally and behaviorally). Both gratitude velocity and trust velocity positively and significantly affect customer loyalty (in terms of attitude and behavior). Lastly, we compar how the level and velocity of gratitude and trust affect attitudinal loyalty and behavioral loyalty via the chi-square difference between the free model and the constrained model. The results are significant (Δχ2(1) = 12.25 & 7.26, Δχ2(1) = 5.69 & 16.4). We find that the mean dynamic effect of gratitude and trust on attitudinal loyalty and behavioral loyalty is greater than that of the static. 4. Research Limitations/Implications The results of this study suggest that gratitude can have a positive effect on trust and customer loyalty over the long term. Gratitude is a key factor that can help enterprises develop and maintain online customer relationships more effectively. However, our sample is merely derived from a business-to-consumer (B2C) platform. To enhance the external validity of our conceptual framework, future studies can further investigate business-tobusiness (B2B) and consumer-to-consumer (C2C) platforms. When future studies employ relationship benefits as a potential variable, the impact of different relationship benefits such as trust benefits, social benefits and special benefits can also be measured. Besides, we collect our data at three points in time, and while it allows us to analyze changing trends among the research constructs, we still recommend that future studies collect more data points over time in order to plot a more precise change trajectory. 5. Contribution Although previous research has pointed out that the relationship between sellers and buyers would change over time, most scholars still test the relational constructs in their theoretical model by using data collected at a single point in time. In response to recent calls in the literature, future studies should explore the relationship change rate correlation
RkJQdWJsaXNoZXIy ODg3MDU=