Unexpected or Non-Significant Results in Experimental Research

Éditeurs invités

• Lori Bhaskar, Indiana University
• Tamara Lambert, The University of Manchester

Synthèse

There is a growing concern among academics that publication bias negatively affects our understanding of the phenomena studied. In auditing research, there is also concern that scarce experimental participant resources may be consumed inefficiently by continuing to test theories or research questions that have failed to be supported by data because other researchers are unaware of the respective designs and outcomes.

The Call seeks to counteract distorted inferences that might occur as a result of publication bias. Submissions could demonstrate that a potentially important result in the experimental literature is not reliably replicated or document and explore conditions under which tests that are well-supported by theory deliver results that are unexpected or not statistically significant. Publishing these results will inform the research community about the conditions under which these theories have been tested so that future research can take these designs and findings into account.

Extrait. Lire l'appel complet sur le site de l'éditeur.

Thèmes proposés

• Publication bias and experimental design in auditing research
• Replication of experimental findings and documentation of non-significant results
• Conditions under which theoretically-supported hypotheses yield unexpected or non-statistically significant outcomes
• Methods for ensuring adequate statistical power, construct validity, and internal validity in experiments
• Guidelines for inference and interpretation when results are not statistically significant
• Efficient use of experimental participant resources and research design
• Development of best practices for designing, analyzing, and reporting experiments with non-significant results