The Null Hypothesis Statistical Testing Paradigm Undermines Knowledge Acquisition in Management Accounting Research: It Needs to Be Abandoned

During the last decade, several areas in the biomedical and social sciences experienced a reproducibility crisis, where mounting empirical evidence indicated that many published findings could not be successfully replicated. This crisis resulted in considerable introspection within the field of statistics because the null hypothesis statistical testing (NHST) paradigm is acknowledged as one of its root causes based on widespread agreement that it is deeply flawed. However, unlike in many other areas, there has yet to be a concerted effort within the discipline of accounting to acknowledge these developments, let alone steps taken to improve practice. This essay aims to spark discussion and debate on the validity of the NHST paradigm by presenting a comprehensive case, incorporating the latest arguments and findings, that demonstrates why the paradigm needs to be abandoned, especially in fields where statistical model misspecification looms large and statistical power is low, such as in management accounting. In so doing, the analysis exposes why obtaining robust knowledge in management accounting has proved elusive. Additionally, it offers a new perspective on the reproducibility crisis and critical insights for improving statistical practice.