A large-scale study on research code quality and execution

Ana Trisovic,Mercè Crosas,Thomas Pasquier,Matthew K Lau

doi:10.1038/s41597-022-01143-6

Abstract

This article presents a study on the quality and execution of research code from publicly-available replication datasets at the Harvard Dataverse repository. Research code is typically created by a group of scientists and published together with academic papers to facilitate research transparency and reproducibility. For this study, we define ten questions to address aspects impacting research reproducibility and reuse. First, we retrieve and analyze more than 2000 replication datasets with over 9000 unique R files published from 2010 to 2020. Second, we execute the code in a clean runtime environment to assess its ease of reuse. Common coding errors were identified, and some of them were solved with automatic code cleaning to aid code execution. We find that 74% of R files failed to complete without error in the initial execution, while 56% failed when code cleaning was applied, showing that many errors can be prevented with good coding practices. We also analyze the replication datasets from journals’ collections and discuss the impact of the journal policy strictness on the code re-execution rate. Finally, based on our results, we propose a set of recommendations for code dissemination aimed at researchers, journals, and repositories.

Highlights

Researchers increasingly publish their data and code to enable scientific transparency, reproducibility, reuse, or compliance with funding bodies, journals, and academic institutions[1]
We find that 74% of R files failed to complete without error in the initial execution, while 56% failed when code cleaning was applied, showing that many errors can be prevented with good coding practices
This paper presents a study that provides an insight into the programming literacy and reproducibility aspects of shared research code

Summary

Introduction

Researchers increasingly publish their data and code to enable scientific transparency, reproducibility, reuse, or compliance with funding bodies, journals, and academic institutions[1]. Studies have reported a lack of research reproducibility[2,3] often caused by inadequate documentation, errors in the code, or missing files. Paradigms such as literate programming could help in making the shared research code more understandable, reusable, and reproducible. Dataverse repositories allow researchers to deposit and share all research objects, including data, code, documentation, or any combination of these files. A bundle of these files associated with a published scientific result is called a replication package (or “replication data” or dataset in Dataverse repositories). Researchers’ code from replication packages usually operates on data to obtain the published result. For the Harvard Dataverse repository, replication packages are typically prepared and deposited by researchers themselves in an unmediated fashion (self-curated)

Objectives

Methods

Results