Reproducible research in statistics: A review and guidelines for the Biometrical Journal

Abstract

Reproducible research (RR) constitutes the idea that a publication should be accompanied by all relevant material to reproduce the results and findings of a scientific work. Hence, results can be verified and researchers are able to build upon these. Efforts of the Biometrical Journal over the last five years have increased the number of manuscripts which are reproducible by a factor of 4 to almost 50%. Yet, more than half of the code submission could not be executed in the initial review due to missing code, missing data or errors in the code. Careful checks of the submitted code as part of the reviewing process are essential to eliminate these issues and to foster RR. In this article, we reviewed n=56 recent submissions of code and data to identify common reproducibility issues. Based on these findings, guidelines for structuring code submission to the Biometrical Journal have been established to help authors. These guidelines should help researchers to implement RR in general. Together with the code reviews, this supports the mission of the Biometrical Journal in publishing highest quality, novel and relevant papers on statistical methods and their applications in life sciences. Source code and data to reproduce the presented data analyses are available as Supplementary Material on the journal's web page.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 35 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.