139 JOURNAL OF EMERGING TECHNOLOGIES IN ACCOUNTING Vol. 4 2007 pp. 139–159 Confidence in the Implementation Process of a New Information System Kevin E. Dow Kent State University Robert A. Leitch University of South Carolina, Columbia ABSTRACT: The perceived likelihood that a new system will be able to support the transaction processing and decision-making needs of an organization can enhance the confidence in an information system. The objective of this paper is to examine antecedent constructs (training, communication, top management support, and change management) that can enhance or diminish a user’s confidence in a new information system for a large financial services organization, as it is phased in over time. Using structural equation modeling techniques, the relationships of these antecedents to con- fidence are analyzed. The structural model fits the data well and the model explains approximately 80 percent of the variance of confidence. In general, these antecedents affect confidence in the new system. We further find that change management, during the implementation process, partially mediates the effect of the other three antecedents and that this mediating effect changes significantly between groups of implementers. Keywords: information systems; implementation; confidence; structural equation mod- eling; multigroup analysis. INTRODUCTION W e investigate the effect of key antecedents of user confidence in a new information system that will lead to outcomes that support the needs of a large financial services organization. We do this for a phased-in implementation, where one group of branches starts the implementation process prior to other groups. We contribute to the accounting and Information systems (IS) implementation research because we con- sider user confidence rather than behavioral intention or use as others have (Anderson and Young 1999; Davis et al. 1989; Hartwick and Barki 1994; Jackson et al. 1997) and we do this over time. The results relating to change management and the phased-in implementation are important to the implementation of new and emerging technologies because they often represent big changes in an organization’s IS. The results of this stream of research are extremely important given the sizeable investments organizations make in their information systems. The authors thank Peter Bentler, Rex Kline, and Steven Burnkrant for their helpful assistance with the multigroup structural equation modeling techniques. The authors also thank Vairam Arunachalam, Ralph Greenberg, Penelope Greenberg, and Jeff Wong for their helpful comments on early drafts of this study. Corresponding author: Kevin E. Dow Email: kdow@kent.edu