How to Review a Scientific Paper

The motivation of this short piece of text is to give a standard and easy guide to carry out a review of a scientific paper. Some good advice comes from people that have written many reviews [*] and others are mine. First, I start explaining a standard review structure and then give some tips about how to read a scientific paper to review it. 

A STANDARD REVIEW STRUCTURE 

  1. SUMMARY (length: 1-3 paragraphs). Here, you should convince authors you was able to understand the article. It is important that you don’t copy&paste pieces of the abstract; use your own words.
    – Practical advice: “write it how you want to explain the paper to your workmate”.
  2. CRITICAL EVALUATION (length: whichever is necessary). Your personal and scientific opinion of the paper should be in this section, from a technical to an opinion about intention of the paper. For instance,  “the problem is interesting, but the paper used an incorrect methodology”  or “the paper used a correct methodology to solve a many-times-solved problem”. Start with pros and then cons, both perspectives are based on what are (not) written … What is the paper hiding?
    – Practical advice 1: “write it how you want to teach the paper pros and cons to your workmate/student”. Remember authors have to learn about your review.
    – Practical advice 2: “The paper may be bad, but not authors;  your anonymity does not give the right to be cruel — avoid ‘your work’ or ‘you should’ “.
    – Practical advice 3: “first look for arguments to accept the paper and then arguments to reject it “.
  3. MINOR COMMENTS (length: an “infinite” list of minor observations). Here, you list all typos, unclear sentences, uncited figures, etc. If these mistakes are many, so the article is not understandable and should be rejected because it is not clear what you are reading. In addition, here you can list questions that are not important for the paper goal.

HOW TO READ A SCIENTIFIC PAPER TO REVIEW IT 

While you are reading the paper, write down notes that include questions. For example, How will authors analyze data? Why do authors use the methodology A and not B? Why aren’t authors are addressing theses “classic” issues of the problem that they want to solve? At the same time, highlight every promise that the paper does. Then, during the reading of the paper, check it out whether or not questions are answered and promises are fulfilled.

At the end, while you are reading the article, you can highlight all minor comments of the paper, but don’t forget that the most important is the critical evaluation.

[*] Thanks:
Mark Hill, University of Wisconsin
Madison Kathryn S McKinley, The University of Texas at Austin
Shriram Krishnamurthi, Brown University
Johan Fabry, University of Chile