Edward Harris
-
Opinion: Should journals encourage more registered report submissions across...
Paper Discussion
Registered reports separate hypothesis registration from result‑dependent acceptance. This reduces publication bias toward positive outcomes. What barriers prevent wider adoption of registered‑report format across more research fields?
James Jones, Jacob Lewis and 50 others4 CommentsView more comments-
Registered reports greatly mitigate publication bias and p‑hacking risks.
2 -
Some research fields struggle with pre‑registration for exploratory‑mode work.
2 - View more comments
-
-
How to build concise and actionable project risk registers for research prop...
Method Help
Project risk registers list potential risks, likelihood, impact and planned mitigation actions. They are often required for large grant proposals. Many researchers write very generic risk descriptions. What principles help you create a realistic, useful risk register document?
View more comments-
Each risk entry should name a concrete plausible threat specific to your project.
2 -
Write concrete mitigation steps instead of vague generic statements.
2 - View more comments
-
-
Thoughts on low‑resource multilingual research bottlenecks
Most cutting‑edge NLP models are built for high‑resource languages. Even with translation‑based approaches, subtle cultural and linguistic nuance gets lost. What promising directions do you see for advancing low‑resource language research without massive annotated corpora?
View more comments-
Weak supervision and multilingual transfer learning look promising, but performance variance across different language families remains huge.
1 -
High‑quality unlabeled raw corpus collection is actually one of the biggest bottlenecks, not model architecture in many cases.
1 - View more comments
-
