Research Note—Using Expectation Disconfirmation Theory and Polynomial Modeling to Understand Trust in Technology
作者研究的是用户对技术的信任，具体关注当技术表现不符合用户预期时，信任会如何变化。以往研究多假设预期与信任是线性关系，但作者认为这种关系可能更复杂，且受情境影响。作者引入“预期成熟度”这一概念，指用户对技术的熟悉程度，通过产品介绍期的长短来测量。作者收集了三种不同预期成熟度的技术使用场景的数据，采用多项式回归和响应面分析（一种统计方法）发现：在预期成熟度高（介绍期长）的情境中，预期不符与信任意图呈非线性关系；而在预期成熟度低（介绍期短）的情境中，二者呈线性关系。这表明，研究技术信任时，需根据情境选择线性或非线性模型。
Trust in technology is an emerging research domain that examines trust in the technology artifact instead of trust in people. Although previous research finds that trust in technology can predict important outcomes, little research has examined the effect of unmet trust in technology expectations on trusting intentions. Furthermore, both trust and expectation disconfirmation theories suggest that trust disconfirmation effects may be more complex than the linear expectation disconfirmation model depicts. However, this complexity may only exist under certain contextual conditions. The current study contributes to this literature by introducing a nonlinear expectation disconfirmation theory model that extends understanding of trust-in-technology expectations and disconfirmation. Not only does the model include both technology trust expectations and technology trusting intention, it also introduces the concept of expectation maturity as a contextual factor. We collected data from three technology usage contexts that differ in expectation maturity, which we operationalize as length of the introductory period. We find that the situation, in terms of expectation maturity, consistently matters. Using polynomial regression and response surface analyses, we find that in contexts with a longer introductory period (i.e., higher expectation maturity), disconfirmation has a nonlinear relationship with trusting intention. When the introductory period is shorter (i.e., expectation maturity is lower), disconfirmation has a linear relationship with trusting intention. This unique set of empirical findings shows when it is valuable to use nonlinear modeling for understanding technology trust disconfirmation. We conclude with implications for future research.