Publishing and Peer Review › Module 3: Peer review from the inside
Single-blind, double-blind, open
Lesson 3 of 3 · 2 min
Venues hide or reveal identities in different ways. Each model tries to balance fairness, honesty and accountability.
The main models
| Model | Reviewers know authors? | Authors know reviewers? |
|---|---|---|
| Single-blind | Yes | No |
| Double-blind | No | No |
| Open review | Yes | Yes, and reviews may be published |
- Single-blind
- Reviewers know authors?: Yes
- Authors know reviewers?: No
- Double-blind
- Reviewers know authors?: No
- Authors know reviewers?: No
- Open review
- Reviewers know authors?: Yes
- Authors know reviewers?: Yes, and reviews may be published
Why double-blind?
Knowing who the authors are can bias reviewers toward famous names and institutions. A 2017 study by Tomkins and colleagues at the WSDM conference compared single-blind and double-blind reviewers for the same papers: single-blind reviewers were significantly more likely to favour papers from famous authors and top institutions.
Why open review?
Open review makes reviewers accountable and lets readers see the debate. OpenReview, used by ICLR and other AI venues, posts reviews and author responses publicly (some only for accepted papers). Journals like eLife and F1000Research publish reviews alongside papers.
What you should do
- Follow the venue's model exactly: anonymise properly for double-blind.
- Write every review as if it might be published with your name on it. That's good practice under any model.
RememberSingle-blind hides reviewers, double-blind hides both sides, open review shows everything. Know which one your venue uses.