Do statistical myths and urban legends matter for the sustainability of strategy research? : the case of moderation analysis

Bergh, Donald; Li, Ming; Sharp, Barton; Vandenberg, Robert

Abstract

Statistical and methodological myths and urban legends (MULs) are perpetuated truisms which in reality are not true. It is widely believed that MULs can perpetuate incorrect methodological decisions, lead to misapplications of analyses, produce inaccurate inferences and provide errant guidelines for reviewers and editors who decide on the merit of manuscripts. However, other than identifying the practices that can lead to MULs and documenting their likely existence, few if any studies have considered their implications for theory development. In the present study, we test whether an MUL associated with using hierarchical linear regression analysis to test moderation can lead to differences in findings and influence conclusions: we address two questions in particular: (1) Did strategy scholars employing hierarchical linear regression interpret and make conclusions regarding main effect coefficients separately from significant interaction terms and (2) would the findings change if such an interpretation would have been made jointly with significant interactions? Based on a content analysis of SMJ articles, we find that MULs are present with respect to testing moderation and that they could adversely influence results and proposed implications for theory and knowledge development. We find that MULs matter in strategic management research and could shape the sustainability of its knowledge base.

Citation

Bergh, D., Li, M., Sharp, B., & Vandenberg, R. Do statistical myths and urban legends matter for the sustainability of strategy research? : the case of moderation analysis

• Book Type: Authored Book
• Deposit Date: Sep 21, 2015
• Journal: Research memorandum (University of Hull. Business School) ; 91
• Peer Reviewed: Not Peer Reviewed
• Book Title: Do statistical myths and urban legends matter for the sustainability of strategy research? : the case of moderation analysis
• ISBN: 9781906422271
• Keywords: Statistics, Research methods, Linear regression, Strategic management research
• Public URL: https://hull-repository.worktribe.com/output/379152