To continue the thread of why I took this job, another reason for me is to help shepherd the community to write great papers. This is mainly done by guiding referees to do a more thorough job at reviewing manuscripts, scrutinizing every aspect of it to make sure the study is conducted and presented as optimally as possible. This direction to the referees, however, can be treated as advice to authors.
The introduction should weave a story of previous studies on the chosen topic to build to the climax of the thesis statement: the unresolved question that still needs to be addressed. It should include citations and discussion of all relevant papers, including those very recently published to show the continued need for another investigation in this area. However, it should not drift beyond the those absolutely necessary to make the point, and should not include an exhaustive listing of papers in tangential or indirectly related fields. The introduction should be focused and make the case that the question is worthy of attention.
The methodology section needs to describe the experimental set up at a level so that others can repeat the analysis. By experiment, I mean the describing the observations and the sensors used to make those observations and/or the numerical model and the run configurations used for the simulations. It is perfectly acceptable to make this section very short, as long sufficient citations are included of previous papers that give the full details of the technique. Quotations of previous papers are also legitimate, especially in this section. If the set up is new or the analysis method has changed, then the section needs to be longer to fully describe what is new in this methodology. For observations, it should be demonstrated that the sensor is properly calibrated and the measurements are reliable. For simulations, it should be shown that the model has been verified against analytical solutions with grid convergence tests and conservation checks.
The results section should objectively present the data and describe the main features of interest in the plots or tables. For this section in particular, it is vital to judge the manuscript against the criterion of “only what’s needed to make the point.” It is easy for authors to include far more information than is necessary to address the problem. On the other hand, enough content should be given to convincingly support the eventual findings of the study.
Next comes the discussion section. Sometimes this is intermixed with the presentation of the results and other times it is intermixed with the conclusions and summary of the study; that’s the author’s choice in how they structure it. This section should bridge the gap from the objective presentation of the results to the conclusion addressing the question posed at the end of the introduction. It should make the case that the findings are an original and significant contribution to the field. It should mention and address any substantial caveats to the study and note any limitations to the applicability of the conclusions.
Some papers will have a final conclusions or summary section. I personally like this as a stand-alone section to the manuscript so that readers can quickly skip to the main findings of the study. It should repeat what has already been stated earlier in the paper and not include any new analysis points.
There are a few other considerations to ensure a great paper. One is that the title should be an appropriate distillation of the main focus of the study. It shouldn’t contain many, if any, acronyms, and should focus on the scientific discovery rather than the methodology used to make the advancement. The abstract should be a concise yet complete summary of the study that explains not only the finding but also the reasoning behind its significance. Tables and figures should be readable, understandable, and adequately and objectively described in the captions. The prose should be publication quality and not need extensive corrective proof-reading.
In short, a great paper contains, in an easily readable and understandable presentation, all that is necessary to convey the significance of the findings…and nothing else. I hope that referees heed this direction and write reviews with this in mind.