📝 SummarXiv

Abstract

In recent years, Bayesian statistics has gained traction across a wide range of scientific disciplines. This paper explores the growing application of Bayesian methods within the field of linguistics and considers their future potential. A survey of articles from different linguistics journals indicates that Bayesian regression has transitioned from fringe to fairly mainstream over the past five years. This paper discusses the main drivers of this shift, including the increased availability of user-friendly software and the replicability crisis in adjacent disciplines, which exposed the shortcomings of the traditional statistical paradigm. It outlines the fundamental conceptual distinctions between frequentist and Bayesian approaches, emphasizing how Bayesian methods can help address the problems. Additionally, the paper highlights the methodological benefits of Bayesian regression for a diverse array of research questions and data types. It also identifies key theoretical and practical challenges associated with Bayesian analysis and offers a set of good practices and recommendations for researchers considering the adoption of Bayesian methods.