I need to review a research paper soon. How does one go about it?

Here's what I do:Read the paper over fairly quickly to get the main points of it.Read the materials and methods more carefully to make sure the approaches used in the investigation are defensible. If you think deficiencies in the science itself obviously warrant rejection, explain this to the editor and your job is basically done.At this point, if the paper still seems potentially publishable, you can go through the paper sentence by sentence and note any major points of concern. I try to limit my points to the most substantial concerns. For example, I will prepare a document with the following information:Give a brief summary of what the manuscript is about, with explicit focus on what objectives/hypotheses are being met/tested. You need to make sure the rest of the paper meets the objectives or tests the hypotheses that were listed in the abstract and/or introduction.Use headings that are the major sections of a manuscript: Abstract, Introduction, Materials & Methods, Results, Discussion, Tables & Figures, and Literature Cited. Under each heading, include line numbers of problematic phrases, sentences, or paragraphs, with a brief description of your concern.An example (with made up line numbers and details): Lines 55-56. The description of Method X is unclear. For example, what instrument was used to make measurement Y, and at what intervals?Another example: Lines 68-71. The description for Analysis A does not provided enough detail to understand the statistical procedures used in the study.Final tips:-Be most attentive to the methods, data analysis and presentation, and data interpretation.Keep descriptions of concerns as concise as possible.Focus on substantial concerns. Grammar and punctuation are not your concern, as copy editors will deal with that if (and when) the manuscript is accepted for publication.Don't be rude, but do be direct. Say exactly what you are thinking.Your review document, comprised primarily of line numbers followed by comments corresponding to those lines of the manuscript, should be no more than 3-5 pages long, double-spaced, size 12 font. Many reviewers keep it on the short end of this, or even shorter. I tend to be about 5 pages, but I'm pretty thorough compared with other reviewers.Decide, based on your review, whether the manuscript should be accepted as is (rare); accepted pending minor revision (occasional); accepted pending acceptable full revision (pretty common); rejected with an invitation to resubmit a fully revised version (most common); or rejected (occasional). Indicate your thoughts on this to the editor.