Tips for Writing Scientific Papers
Kevin P. White, M.D., Ph.D.
Among editors, I have the distinct advantage of having had a long, highly successful career as a scientific researcher, and having served as an editor and reviewer for several, high-profile scientific journals. I can tell you that there is no way to ensure, with one hundred percent certainty, that your paper is going to be accepted by any given journal. Having said this, I also must tell you that, in my career, I only have had one scientific paper which I failed to have published in a peer-reviewed scientific journal. In other words, it is important for you to realize that scientific journals WANT to publish good research. If your study has been well designed and orchestrated, irrespective of a positive or negative result, it is publishable. It is critical, therefore, not to let improper editing prevent you from having your papers accepted. What follows are a few tips, which stem from what I look for when reviewing a paper. You never know, one day I may be reviewing yours.
Let me start with some very basic writing tips:
Now here are some specific tips for writing your Methods and Materials and your Results sections:
- While you, yourself, are editing your paper, read it out loud. This is something I always do and find extremely helpful.
- Never rush to submit a paper. Set yourself a pre-target date to have the paper written and edited, then let it sit a few days. Then pick it up and read it again. You will be surprised what changes you will think of.
- Never write and edit a paper without feedback from others. No matter how accomplished a writer you are, you are human. You cannot possibly think of everything that a group of reviewers will want to see. Nor can most writers pick up every potentially awkward phrase or sentence, every misspelling (even with spell checks, similar sounding words like there, their and theyíre invariably become confused. I suggest that you have at least one other scientist in your field read your paper and one other scientist outside your immediate field read it. Why someone outside your field? Just remember that, quite possibly, at least one person reviewing your paper for the journal will have an expertise somewhat distant from your own; if that reviewer cannot understand your writing, you may be sunk. Finally, I strongly suggest that you have a professional editor or another very meticulous writer read and edit it.
- Although you want all parts of your paper to be as strong as they can be, really, really make sure that its Methods and Materials section is meticulously written and edited, and that all data are presented clearly, both in the text and in easy-to-follow tables and figures. Rarely will a paper be rejected outright because of a weak Introduction or Discussion. The same is not true of weak Methods or Results.
A scientific paper that flows well and is easy to read is much more likely to be accepted for publication than one which is disjointed, confusing and error-filled, irrespective of the scientific merit of the study described. Do not let your paper be rejected because of how it has been written. You already have worked far too hard to design your study and collect your data.
- Make sure that you have a very clearly stated primary hypothesis that is well justified by your introduction. This hypothesis should immediately follow the paperís introduction and flow directly from it. Only after you have clearly stated your primary hypothesis or study objective, should you state or list your secondary hypotheses. Too often, I read papers in which every one of about ten different objectives is given equal importance. This is extremely confusing to reviewers and will lead them to believe that you never truly knew what you wanted to achieve. Also, the primary objective is the one which justifies your sample size.
- Make sure that it is very clear how you are going to analyze your data to meet each objective. When I am writing a research grant, the first thing I do is decide what my primary and secondary objectives are. Then I immediately sit down and plan my analysis scheme for each objective. This makes it easy later when I am doing the analysis on collected data and when I am writing a paper for submission. Most reviewers become extremely perplexed if they cannot draw a straight line between each objective and the analysis plan you have described. Make sure that none of your objectives is left without a clearly stated analysis plan. I find that using a matched numbering scheme for objectives and analyses can be very helpful. In other words, objective #1 will be met by using data analysis method #1, and so on.
- Donít forget to go back and fill in all the pertinent details that should fall between listing your objectives and describing your data analysis plan. The most frequently short-changed details in studies involving human subjects are those pertaining to subject recruitment. Describe in detail how subjects in each group (Treatment A, Treatment B, Placebo and so on) are recruited: from the clinic or from the community; randomly or consecutively; using a mailed questionnaire or by telephone using trained interviewers; and so on. Also make sure that you list all of your inclusion and exclusion criteria. These lists are very important to reviewers.
- If you randomize subjects into subject groups, describe how. If subjects are blinded to treatment, describe how.
- Justify your sample size. A recent study of scientific papers published in the last decade found that the majority never had enough subjects to adequately answer the primary question being addressed. Because of this, current reviewers generally are very sensitive to ensuring adequate sample size. I find that this is a detail that is neglected in the majority of submissions, and such an oversight significantly weakens any paper (or grant) in which sample size is a potential issue.
- If your study involves the use of questionnaires, describe them and justify why you are using them. Have they been used by your research group or other researchers before? Have the questionnaires been published? Have they been scientifically validated? All of these details will strengthen your submission.
- Finally, avoid Tables and Figures that contain too much data. You DO NOT have to have all data both in the text and in a table or figure. Tables and Figures should be reserved for your most important data and/or data that are much more easily presented pictorially than in text.