From the Executive Editor: Reviewing scientific papers: Results, discussion, implications, and references
Cate Dewey, DVM MSc, PhD
In the last issue of the Journal of Swine Health and Production, I described an approach to reviewing the introduction and methods sections of a scientific paper.1 To summarize, the introduction must provide the background literature illustrating why the study is important and a clear description of the purpose of the research, including its breadth and depth. The methods must be described completely so that the reader could reproduce the study exactly. The following is a description of the general approach to reviewing the rest of the manuscript.
The results section includes the written portion, tables, and figures. I read the tables first because they provide the overview of the results and show me the key findings. Each table must stand alone. If you need to read the text to understand the table, it does not stand alone. Tell the author what is missing. Typical examples are a description of the statistical analysis, a clear understanding of the headings used for the columns and rows, and the number of observations used for each analysis. Does the title describe the table fully by providing the overall purpose of the table but also the source of data? Is there sufficient information in the footnotes to understand the acronyms, statistical analyses, and the levels of significance found in the analysis? As with tables, figures must stand alone. They must include the statistical test used for comparisons and a note of the associations that were statistically significant. Figures are meant to provide a visual summary of the results that is more informative than what would have occurred with either a table or the text. Authors may present the same results in both table and figure form. The reviewer must determine which is best to illustrate the data. Figures that are confusing or misleading are not helpful. Those that exaggerate the results are inappropriate. An example of this would be a bar graph with a narrow range of y-axis values that then depicts a difference between treatments when there was not a statistical difference. Tables and figures often provide a large amount of information in a relatively small space. However, they are expensive to publish. We ask that reviewers indicate when a table or figure is not necessary.
The text of the results usually begins with an explanation of the sample used for the analysis. If a study begins with 2000 nursery pigs but ends with 1900 pigs because of mortalities and culls, the author must explain the details of the pigs not remaining in the study to the end. This information enables the reader to understand the analysis and also to determine whether the pigs lost to follow-up were a result of the treatment or condition in the study. The reviewer must fully understand how and why numbers of observations change during the analyses, all throughout the paper. As an example, specific parts of the data may be missing for some pigs (observations). Perhaps parity is missing for some sows.
The text of the results highlights the findings in the tables, but should not repeat everything from the tables. The reviewer determines whether the text clearly refers to the tables to enhance the reader’s understanding. Finally, do the results match the first part of the paper? Do the results reflect the purpose? Are there results that were never mentioned in the methods, leaving you wondering where that came from? Do the results mention all analysis described in the methods?
The discussion is the author’s interpretation of the results and also the place where the author describes how the findings agree or disagree with the work of other researchers. Does the discussion flow in a logical manner? Did the author include the relevant and recent research work of others? There should be no results in the discussion that are not first mentioned in the results section. Over-interpretation of results is a common problem. Are you comfortable that the conclusions reached are valid on the basis of this study and previous literature? Even if the study is valid, well designed, and appropriately carried out, the discussion and implications must stay within the boundaries of the study. Did the authors state the strengths and weaknesses of the study? Did they recommend the next logical research project?
When examining the reference section, first determine if each reference has a complete citation. Is the list sufficiently extensive and up to date? Are there key references or scientists who work in this area who have been left off the list? Did the author use conference proceedings or the relevant peer-reviewed literature?
After I have reviewed the paper once, I like to set it aside for a day or two. I take that time to think about what I learned and what questions I still have. I think about whether the discussion and implications stay within appropriate bounds. Then I re-read the paper to see if there are new issues that I missed. This is when I write my overall impression of the paper. I finish by providing a list of questions or concerns that are referenced to the manuscript’s line numbers. I try to provide positive and constructive feedback.
Reference
1. Dewey CE. Reviewing scientific papers: Introduction and materials and methods. Editorial. J Swine Health Prod. 2010;18:67.
— Cate Dewey, DVM, MSc, PhD