A twitter comment from César complaining about references in supposedly-anonymous papers that clearly identified the authors, summarizes my main criticism to double-blind reviews (not awfully common but, unfortunately, still requested by some journals/conferences): they just don’t work!

In theory, double-blind reviews seem a good idea (anything that, in theory, helps to provide more objective reviews should be welcomed) but we already know that “In theory, theory and practice are the same; in practice, they are not” and double-blind reviews are a clear example of that. Maybe it’s just the size of my research community, but I’d say that in a 90% of the cases I’m able to identify at least the research group where the paper comes from, even if the author name is hidden.

Then, you could say, well, double-blind reviews are not so bad because in the worst-case scenario (when the reviewer can correctly guess the name of the authors) they just work as good as single-blind reviews, so why you don’t like them?

In fact, I don’t like them as an author myself!!! Since I don’t believe they represent a significant improvement in the peer-review process I hate the extra-time I’ve to put in a paper to prepare it for a double-blind review process. It’s not as simple as not writing my name on it, guidelines for double-blind reviews also request that you “hide” self-references and make difficult to add links to tools to be downloaded and tested (hiding the contributors of a tool is almost impossible!, if the tool has been properly done of course, for instance, it’s absolutely impossible when the tool is an official Eclipse project just because of the rules imposed by the Eclipse community).