In Which Model Do We Trust and When? Comparing the Explanatory and Predictive Abilities of Models of E-Government Trust

With the growth in digital provision of government services (i.e., e-government), a substantial quantity of recent research has focused on models of satisfaction and trust with public sector services. Although few would deny the relevance of the satisfaction-trust relationship, there is little agreement about how to optimally model these relationships. In this paper, we compare an assortment of conceptual models of the e-government citizen satisfaction-trust relationship, drawn from service-quality and expectancydisconfirmation paradigms for their ability to explain trust, their parsimony, and their in-sample and outof-sample predictive abilities. We use survey data from the American Customer Satisfaction Index (ACSI) measuring citizen e-government experiences. Our findings suggest that while expectancy-disconfirmation model does well for explanation, service-quality model is better suited for prediction of citizen trust. Overall, the service-quality model also offers the best compromise between predictive accuracy and explanatory power. These findings offer new insights for researchers, government agencies, and practitioners.