Promoting computationally reproducible research in applied linguistics: Recommended practices and considerations

Abstract

Reproducible research is an important topic that has been discussed in many fields, including applied linguistics. It helps researchers to verify findings through attempts to re-create the numeric results reported in the original studies. Researchers can then evaluate the extent to which the analysis code is accurate, the statistical values are correctly reported, and errors in these aspects of the study influence the findings. Building on and extending recommended open science research practices, we describe and demonstrate how reproducible research can be further promoted and practiced from three methodological perspectives. First, we describe advantages of sharing supplementary information in online repositories, with a particular focus on IRIS (Instrument for Research Into Second Language Learning and Teaching) and OSF (Open Science Framework). Second, we consider ways to present such information in a reproducible manner by encouraging the use of R and R Markdown, a container, and a online platform. Finally, we focus on using simulated data created from the original data, which can be particularly useful when original data cannot be shared due to various reasons. Through these recommendations and considerations, we intend to enable readers to better understand how they can apply such practices to their own research context and eventually how these practices can advance the broader applied linguistics community.