Building awareness of the replication crisis in linguistics and how to improve it. 

If a research finding is true, then it should replicate. 

In other words, if we maintain all facets of an original study and attempt to re-create its results, we should be successful. Replication is a vital part of the scientific process, and allows us to determine whether the work we are doing is cumulative and self-correcting. The term cumulative refers to the idea that as time goes by, earlier work should serve as a foundation to be built upon and allow us to construct theory. Self-correction in science refers to the long-run convergence toward the true nature of our object of study. That is, if a published finding is a false positive result based on random chance, then contradictory results obtained from replication studies over time should “correct” the significant finding of the original work that was, in reality, a byproduct of our choice of statistical analysis. 

The replication crisis suggests that a significant proportion of the body of work in many fields does not replicate. 

In 2015, the Open Science Collaboration carried out an ambitious project: to determine what proportion of the most influential studies in psychology would replicate in (close to) exact replication studies. To do so, they recruited a team of 270 scientists. The results showed that only 36% of studies replicated.

There are many factors that have been proposed to underlie the replication crisis. 

These include, but are not limited to: cognitive bias (Nickerson, 1988), multiple analytical pipelines in the research process (Coretta et al. 2023), statistical power (Button et al., 2013), and the incentive structure of the academy (Munafò et al., 2017). 

The goal of this workshop is to equip researchers with practical tools and theoretical awareness with the ultimate goal of increasing replicability in future linguistic work.

Organizing and scientific committee

(in alphabetical order)