Expectation Confirmation in Technology Adoption: An Examination of Six Competing Theoretical Models

We discuss six different models of expectation confirmation: assimilation, contrast, generalized negativity, assimilation-contrast, experiences only, and expectations only. We identified four shortcomings of prior empirical research on expectation confirmation in IS and other fields: reliance on a single model, assumption of linearity, use of difference scores, and direct measures. Using constructs from the widely studied technology acceptance model (TAM), we test each of these six models, which suggest different roles for expectations and experiences, using expectations and experiences about ease of use and usefulness as predictors of behavioral intention. Data were collected in a field study from 1,113 participants at two points in time. Using polynomial modeling and response surface analysis, we demonstrated that expectations and experiences have differential impacts on intention depending on whether we are examining ease of use or usefulness. Specifically, the results did not support any of the existing models of expectation confirmation but rather, we present a modified generalized negativity and contrast model. INTRODUCTION Competing models of expectation confirmation, with different underlying theoretical explanations, have been put forth in organizational behavior, marketing, and psychology. Among the various models, four overarching models examining expectations, experiences, and consequent outcomes have received a great deal of attention: assimilation, contrast, generalized negativity, and assimilation-contrast (Anderson 1973; Klein 1999; Oliver 1977, 1980; Yi 1999). In addition, models hypothesizing no role for expectations or experiences—i.e., experiences only and expectations only—have emerged (e.g., Hom et al. 1999; Irving and Meyer 1994; Pulakos and Schmitt 1983). Each model varies in terms of its prescriptions on how expectations should be set in order to produce desirable outcomes. Thus, the goal of expectation confirmation research is to understand how missing, meeting, or exceeding expectations influences the acceptance of an information system (or the ultimate outcome of interest). Prior organizational behavior research has examined these models in a variety of contexts, such as organizational commitment (Wanous et al. 1992), job satisfaction (Greenhaus et al. 1983), and intention to turnover (Wanous et al., 1992). There is also a long history of research on expectations and its ties to IS outcomes (e.g., Bhattacherjee 2001; Ginzberg 1981; Staples et al. 2002; Szajna and Scamell 1993). IS research draws from organizational behavior and psychology (e.g., Churchill and Suprenant 1982; Festinger 1962; Oliver 1980), although none of the four models that we mentioned initially have been explicitly examined.