Blind peer review: how to set up single-blind and double-blind review

Blind peer review is a method in which the identities of authors and reviewers are hidden from each other: in the single-blind model the author does not know the reviewer, and in the double-blind model the reviewer does not know the author either. A sound process needs four things working together: files that are genuinely anonymised, conflicts of interest filtered out, the right reviewer assigned and a decision passed to the author with reasoned feedback. This guide explains how to design each of them for academic journals and congresses.

What is the difference between single-blind and double-blind review?

In single-blind review the reviewer knows who the author is, but the author never learns the reviewer's identity. It is easy to run because files need no anonymisation, yet what the reviewer knows about the author's institution, seniority or earlier work can colour the assessment without anyone intending it.

In double-blind review neither side knows the other. The aim is an assessment focused on the work rather than the person, and it is the method many academic journals and congresses prefer. The price is rigorous anonymisation: a single file name or acknowledgements paragraph can reveal who the author is.

There are also models such as open review, where identities are shared, but this article focuses on blind models. Whichever you choose, announce it to authors and reviewers in advance and include it in the written policy of the journal or congress.

Journal and congress review rest on the same principles but differ in scale and pace. A single journal article may be reviewed over several rounds for months, while a congress decides on hundreds of abstracts in one round within a few weeks. That is why automated assignment and reminders matter most for congresses, and a clean record of revision rounds matters most for journals.

Blind review is not an intention but a process design; identity stays hidden only when it is hidden on every channel.

How do you anonymise a submission?

Anonymisation is more than deleting author names from the title page. Identity can also show up in the file name, the author field of document properties, the acknowledgements, funding details, ethics approval and the author's citations of their own earlier work.

• Author names, affiliations and contact details should go on a title page kept separate from the main text.
• Uploaded files should be renamed by the system, and the original file name should never reach the reviewer.
• Author and organisation fields in document properties should be cleared, and tracked changes and comments removed.
• Acknowledgements, funding and ethics approval details should be left out of the review version and added after acceptance.
• Self-citations should be written in the third person or masked where necessary.
• Emails sent to reviewers should never contain the author's name, institution or email address.

Some of these steps are the author's responsibility, but the system has to do its part too. A reviewer screen that never shows author identity, files delivered under anonymous names and automatic emails checked for identifying details are the safeguards that make up for human error.

The editor needs a checkpoint as well. Checking anonymity during the initial screening, and sending the file back to the author for correction when needed, costs far less than discovering an identity leak once review is under way.

How do you check for conflicts of interest?

A conflict of interest is any relationship that makes impartial review harder: working at the same institution, recent co-authorship, a supervisor and student relationship, running a competing project or a personal connection. Blind review does not remove these relationships; it only makes them invisible.

The first line of defence is an automatic check. Comparing the affiliations and email domains of reviewers and authors can block assignments within the same institution. A reviewer assessing their own paper, or an article they co-wrote, should be impossible at system level.

The second line is the reviewer's own declaration. The invitation should include the title and abstract, and reviewers should be asked to declare any conflict before accepting. There should also be an easy way to tell the editor about a relationship noticed during review.

Editors are bound by the same rule. When an editor is an author of a submitted article or paper, handing decision authority to another editor is the minimum that publication ethics expects.

How should reviewers be assigned?

The right reviewer knows the subject, has the time and has no conflict of interest. Meeting all three at once requires a structured reviewer pool that records each reviewer's areas of expertise, keywords, languages and how many reviews they currently hold.

Load balancing is often overlooked. Sending everything to the same dependable reviewers looks easy in the short term but leads to delays and careless reviews. Seeing each reviewer's open workload during assignment, or letting the system spread the load, prevents the problem from the start.

Congresses review hundreds of papers in a short window, so automatic assignment by sub-theme, combined with conflict checks and load balancing, helps enormously. Journals handle fewer manuscripts but assign each one with more care: the editor looks at suggested reviewers, sends invitations and tracks whether they are accepted or declined.

Restricting reviewers to the work assigned to them is part of assignment too. The reviewer screen should not show other submissions, other reviewers' comments or the details of the decision process.

How is the editorial decision made?

Reviewers recommend; the editor or scientific committee decides. The distinction matters: when two reviewers disagree, the editor weighs the reasoning behind each report, assigns a third reviewer if needed or decides on the basis of their own expertise.

The decision options should be defined in advance. Journals usually work with accept, minor revision, major revision and reject; for congresses, accept, accept after revision and reject are often enough. Every decision should be followed by a letter to the author giving the reasons and the reviewers' comments.

Revision rounds should continue on the same record. The revised version and the author's response to reviewers should be visible alongside the previous round, so reviewers can judge in the second round how their own comments were addressed.

There should also be a defined route for authors to appeal a decision. When the policy states to whom, within what period and on what grounds an appeal can be made, authors can trust the process and editors do not have to negotiate each case from scratch. Having the appeal reviewed by someone other than the editor who made the original decision strengthens impartiality.

How can the quality of reviewer feedback be improved?

A good review report tells the author how to improve the work. Reports that only say acceptable or inadequate give the editor too little to decide on and give the author no direction. Report quality depends to a large extent on how the review form is designed.

• The review form should have separate criteria for originality, method, findings and conclusions.
• Comments for the author and notes for the editor only should be collected in separate fields.
• Reviewers should be asked to explain their recommendation in a few sentences.
• Comments should address the work rather than the person, in respectful and specific language.
• Before forwarding a report, the editor should check it for identifying details or inappropriate wording.

Making reviewers' work visible also affects quality. Publishing an annual list of reviewers or giving them a certificate of review shows that this largely voluntary work is valued.

A short, clear reviewer guide also raises report quality. It should cover the criteria, the expected length of a report, the deadline, the duty of confidentiality and what to do in case of a conflict of interest. For early-career researchers reviewing for the first time, a sample report is the easiest way to make expectations concrete.

Editorial turnaround statistics deserve regular attention too. The time from submission to first decision, and from acceptance to publication, shows where the process stalls and lets you give authors a realistic timeline.

Frequently asked questions

Is double-blind review more reliable than single-blind review?

The double-blind model aims to reduce the chance that the author's identity influences the reviewer, which is why many journals and congresses prefer it. Its reliability depends on how carefully anonymisation is done, though; a file whose author is easy to identify makes the double-blind label meaningless.

What should a reviewer do if they can guess who the author is?

In small research fields a reviewer may well guess the author. The reviewer should then tell the editor and withdraw if they feel they cannot stay impartial, and the editor assigns someone else if necessary.

How many reviewers does a manuscript need?

Common practice in academic journals is at least two independent reviewers, with a third added when opinions diverge. The exact number should be stated in the journal's policy and be consistent with what the indexes you plan to apply to expect.

Should reviewer comments be forwarded to the author unchanged?

The substance of the comments should be passed on unchanged, but the editor should remove anything that identifies the reviewer and any personal or hurtful language. Notes written for the editor alone should never be sent to the author.

What if a reviewer does not respond to an invitation?

Set a reasonable response period for invitations and send a reminder before it runs out. If there is still no answer, withdraw the invitation and invite another reviewer from the pool. Keeping a record of unanswered invitations helps you make more realistic choices in future assignments.

Letting the system enforce the rules

The quality of blind review depends on the written policy actually being applied at every step. Write the policy down, add the anonymisation and conflict-of-interest rules to your author and reviewer guidelines, and then entrust those rules to the system itself rather than to your team's vigilance. Reviewing the process once a year, looking at anonymisation slips, late reviews and appeals together, shows that the policy is more than words on paper.

In a well-built process, editors spend their time on scientific decisions rather than on cleaning files. At Bir Web Ajans we build these rules into the software itself, in Nova JournalPort for journals and Nova CongressPort for congresses: in double-blind mode reviewers never see author identity, and each reviewer can reach only the work assigned to them.