Testing contingency hypotheses in budgetary research using moderated regression analysis: a second look

This paper addresses the use of moderated regression analysis (mra) in contemporary management accounting and control research. It follows up on a discussion started by us [hartmann & moers, accounting, organizations and society 24 (1999) 291–315] in this journal and provides a reaction to arguments put forward by dunk (this issue). In doing so, this paper addresses the relationship between substantive theory and statistical test, emphasizes the need to distinguish between confirmatory and exploratory uses of mra and argues that the importance of moderated hypotheses and tests may have been overstated in the management accounting literature under review.