RAP-group / guide_to_open_science

A guide to open science and reproducibility for students, advisors, and early career researchers
https://doi.org/10.31234/osf.io/spz4w
0 stars 0 forks source link

R2: preregistration, be more inclusive (again) #54

Open jvcasillas opened 2 weeks ago

jvcasillas commented 2 weeks ago
  • “Linguistic research is multifaceted and spans diverse areas such as corpus analysis, conversation/discourse analysis, experimental research, and more.” --> This is a very reduced list of linguistic research areas, please add more areas and discuss in more detail for which areas preregistration makes sense. For instance, can it be used in theoretical linguistics? It is mentioned in the beginning that “we focus on who might want to consider preregistrations”, but this does not become very clear. Please add at least a paragraph about what types of linguistic studies can benefit from preregistrations.

TODO

GaboSJW commented 2 days ago

Linguistic research involving hypothesis testing, experimental designs, or statistical analyses can all benefit from preregistration. For instance, by adopting the practice of preregistration, psycholinguistic research on language processing, production, and comprehension can reduce bias by specifying hypotheses and analysis plans in advance. Sociolinguistic studies examining language variation across different social groups and corpus-based studies involving large datasets and multiple statistical comparisons can also benefit by enhancing the transparency of data handling procedures. Additionally, studies in phonetics or syntax that explore subtle effects or relationships are prone to researcher degrees of freedom, making preregistration particularly valuable for mitigating flexibility in data analysis and improving the robustness of findings.

Even theoretical linguistics can benefit from preregistration, though its application may look different compared to experimental or quantitative studies. In theoretical work, preregistration can help clarify the scope of the analysis, specify the theoretical predictions being tested, and outline criteria for evaluating competing theories. This reduces the risk of post hoc reasoning and helps ensure that interpretations are consistent with the original theoretical framework.