Enhancing the Reproducibility of Research Findings

There has been increasing attention to the reproducibility of research findings in the social sciences. These discussions have focused on the extent to which published studies can be replicated or reproduced, and the kinds of practices that can contribute to the reporting of spurious results. Of particular concern is “p-hacking” (Open Science Collaboration, 2012, 2015), which can include practices such as collecting additional data when interim analyses suggest effects are approaching conventional statistical significance thresholds, omitting results, observations or conditions that did not reach statistical significance, omitting outliers to influence statistical results, and conducting many analyses but only reporting those that are favorable to the study hypotheses (Head, Holman, Lanfen, Kahn, & Jennions, 2015).

Separately, there have been concerns expressed about the potential impact of conflicts of interest in what is submitted and published in peer reviewed journals. This is particularly notable in biomedical research where corporate funding may be associated with an increase in p-hacking practices (e.g., Kjaergard & Als-Nielsen, 2012). Statements that increase transparency in manuscripts, such as the 21 Word solution proposed by Simmons, Nelson, and Simonsohn (2012) ¹ , have been proposed to reduce spuriously significant findings.

In response to these concerns, and to maximize the reproducibility and transparency of research reported in this journal, Sexual Abuse: A Journal of Research and Treatment (SAJRT) is implementing several changes in its editorial and peer review process. This includes requirements that authors consider and complete statements regarding statistical significance testing, research funding, and disclosures of financial interest. These guidelines can be found on the journal website, with examples of how the guidelines could be addressed, as best as possible, in different study designs. We encourage the reporting of effect sizes and confidence intervals in addition to, or instead of, traditional Null Hypothesis Significance Testing (Wasserstein & Lazar, 2016). We also recognize that replications are necessary and should be valued, especially for controversial and/or unexpected findings. Submissions that attempt to replicate or reproduce previous findings are welcomed.

These revised guidelines are not meant to discourage submissions, or to suggest that submissions that cannot meet one or more of the suggestions will not be considered or accepted for publication. Our hope is that the submission guidelines inspire authors to follow practices that can enhance the robustness and reproducibility of research findings. As Editor-in-Chief, I hope that implementing these policies will increase the confidence of readers in the results that are reported on these pages.