If citizens of a democracy are expected to hold governments accountable for effective public service delivery, it is important to understand how citizens evaluate these services. Existing studies of citizen satisfaction have found mixed support for the now-widely used expectancy-confirmation model, which suggests that satisfaction (or dissatisfaction) comes from comparing perceived performance of service with prior expectations (Oliver 1977; 1980). In this paper, we provide theoretical analysis of the expectancy-confirmation model and then report the results of a survey experiment that tests a set of theoretically-motivated hypotheses.
We build on prior studies in three main ways. First, we more fully incorporate James’s (2011) distinction between positive (predictive) and normative expectations into the expectancy-confirmation model (a model originally developed in the private-sector marketing literature). We argue that this distinction is particularly important in the public sector, where citizens likely consume services even if their predictive expectation is that the service will be unsatisfactory. To date, most empirical studies in public administration have measured expectations with survey items that imply a predictive component. We argue that the cognitive logic of the expectancy-confirmation model is more compelling if expectations are defined (and measured) as normative expectation
Second, we theorize about the types of information that citizens use to form predictive and normative expectations. We argue that historical comparisons (past performance) will often form the basis of predictive expectations while social comparisons (other organizations’ performance) inform citizens’ normative expectations.
Finally, we introduce a methodological innovation from the marketing literature by attempting to experimentally manipulate expectations in order see how expectations causally affect satisfaction. We set up a (randomized) 2x2 survey experiment in which we manipulate information about historical and social comparisons, which we anticipate will affect respondents’ predictive and normative expectations. We then test how these experimentally-induced changes in expectations affect citizen satisfaction levels.
Our results provide insights into how ordinary citizens use information to form judgements about public service provision. The findings suggest that citizens may be more sophisticated and consistent in their evaluative processes than some prior work has suggested. At the same time, citizens may be somewhat vulnerable to manipulation of their expectations if service providers are able to present them with comparative performance information that will paint their own service provision in a more favorable light. Finally, this study fills a gap in the existing citizen satisfaction literature by providing a fairly strong causal test of the relationship between expectations and satisfaction.