Reproducibility in NLP: What Have We Learned from the Checklist?

TL;DR

This study analyzes the impact of the NLP Reproducibility Checklist on research reporting and acceptance rates, revealing improvements in transparency and identifying areas needing further enhancement for reproducibility.

Contribution

First comprehensive analysis of the NLP Reproducibility Checklist responses, showing its effects on reporting practices and acceptance, and offering recommendations for future conference policies.

Findings

Increased reporting of efficiency, validation, and hyperparameters after Checklist implementation.

Higher acceptance rates for submissions with more 'Yes' responses on the Checklist.

Submissions with new data are less likely to be accepted and have lower reproducibility scores.

Abstract

Scientific progress in NLP rests on the reproducibility of researchers' claims. The *CL conferences created the NLP Reproducibility Checklist in 2020 to be completed by authors at submission to remind them of key information to include. We provide the first analysis of the Checklist by examining 10,405 anonymous responses to it. First, we find evidence of an increase in reporting of information on efficiency, validation performance, summary statistics, and hyperparameters after the Checklist's introduction. Further, we show acceptance rate grows for submissions with more Yes responses. We find that the 44% of submissions that gather new data are 5% less likely to be accepted than those that did not; the average reviewer-rated reproducibility of these submissions is also 2% lower relative to the rest. We find that only 46% of submissions claim to open-source their code, though submissions that do have 8% higher reproducibility score relative to those that do not, the most for any item. We discuss what can be inferred about the state of reproducibility in NLP, and provide a set of recommendations for future conferences, including: a) allowing submitting code and appendices one week after the deadline, and b) measuring dataset reproducibility by a checklist of data collection practices.