Because you play another round until your number wins!
Manuscript writing seems like it should be a straightforward ordeal. You explain the current body of research on the subject and identify the knowledge gap that your hypothesis fills, explain the rationale and objectives for the study, describe all the methods you used, present the data results, and then interpret them in the discussion. Oh and don’t forget the bibliography. Simple!
Oh contraire. Many manuscripts grow and then end up splitting into two or more, or you add a collaborative project on after the fact using the samples you’ve already collected. Sometimes you just say “let’s test these and see what happens”, and you don’t have a specific hypothesis except for “it could be cool”. Moreover, when you work in a very novel, difficult, unpopular, or boring field, there often isn’t a lot of previous research for you to read up on. It makes it more challenging to write what should be the easiest section, the Introduction, because you don’t have much background to introduce. While this does justify your work and the need for more research, it also makes it difficult to plan an experiment because you don’t know what outcomes or problems will crop up, and it can make your interpretation of the data problematic.
Methods: probably the worst section.
Sometimes you end up with more or less data than you planned. And most often, you didn’t just use commercial kit instructions, you probably had to piece together methods from two to ten different journal articles, many of which were not verbosely described to maintain a sort of proprietary hold on the procedures, until you end up with a heavily-citationed Frankenstein’s monster of a Methods section. Not to mention that you probably had to mess around with procedures to find just the right settings on your equipment, so you have to go back through your lab notebook and try to tease apart what you did months or years ago. My suggestion: write the Methods while you are running the experiment. Whenever you finish a procedure that worked, type it up, especially if you are stuck waiting for something to process or grow anyway.
In 2015, I worked on the DNA sequencing section of a project that had begun four years earlier when the original animal feeding trials were run, and which had been sequenced nearly a year prior to my taking over the data. Not only did the original Principal Investigator (PI) have trouble digging up the project files from four years ago, but the technician who had sequenced the data was no longer a member of the lab. Between the two, it was very difficult to track down what had been done, and which sequencing file name corresponded to which sheep sample. Even if you think the project will never be published, TAKE GOOD NOTES. Really specific, legible ones, trust me- you’ll thank me later.
Results and Discussion: Let’s be honest, the only two sections anyone actually reads.
Results is the easiest section to write, but possibly the most difficult to make appealing and understandable to a general scientific audience. Naturally, you need to know how to properly summarize your data and how to graph it. Seems easy: something about means and standard deviations, liberally sprinkle in some p-values… But in reality, there are lots of ways to statistically validate or measure something, and most of these are minor variations on each other to accommodate slightly different data or situations. Maybe your data has a bell-curve normal distribution like people’s height in North America; maybe it’s heavily skewed to one side, like my preference for maple-frosted donuts over celery. Or you need an ordination plot that takes non-Euclidean distance samples and graphs their relationship to each other by plotting one point, then rotating the axis and plotting another until you’ve plotted all your points. No matter how sophisticated your presentation techniques, if someone can’t look at your graph and the graph summary out of context and understand what you are measuring, you haven’t done your job well. I’ve heard many scientific authors complain that a reviewer demanded changes to the manuscript because they did not under the results or statistical analysis. That can be frustrating, and sometimes it feels like the reviewer is just being obtuse, but as scientific authors it’s our job to properly explain what we did.
The Discussion section is always my favorite, because now you interpret your results into the context of other findings and speculations- in short, you finally get to tell the story of what is happening and why in a more interesting way.
The rest is just details. The Conflict of Interest section is always very interesting. Here you must disclose any conflicts you have, anything from a funding source that paid for your work and may or may not have had input in the experimental design (sometimes commercial companies will contract researchers to do a specific experiment that they more-or-less designed), or that the commercial lab you sent your samples to be tested at has you on the payroll. The Conflict of Interest is usually blank for studies coming out of academic universities, but it’s a good way to track down researchers who might be biased towards or against something.
You have an acknowledgements section where you can thank personnel that may have assisted you in some small way, someone who you bounced ideas off of in planning and interpretation, someone who gave you samples to work with free of charge. In my case, I most often thanked the hunters who had dutifully collected a jar of rumen (stomach) contents, and sometimes colon contents, from moose while they were field dressing. Or the numerous undergrads that helped feed my newborn lambs five times a day until they were weaned.
Last but not least, the Bibliography or References Cited. Sounds easy enough. But you’d be surprised how pesky it can be. Different journals often want different formatting for your submission, some want authors lists to look like “Last, First; Last, First”, or maybe “Last, F., Last, F.”, or even “Last F, Last F”. Some want years in parentheses, others don’t. Some want issue number, or the journal name to be abbreviated, or a certain part of the reference to be bolded. Trying to reformat 50-100 references for submission to a different journal can be a nightmare. Luckily, there are plenty of citation managers that will create a digital library for your references, and allow you to search for citations while you are writing. Then, you hit “Insert Bibliography” and it numbers or alphabetizes it, and puts them into the desired format. That is, assuming you had put all the correct bibliographic information in. I like Mendeley because I can import references from my web browser; however, on older PDFs sometimes it can’t pick up the info it needs and you have to do it manually. I’ve gotten some interesting inputs for authors’ names when it gets confused.
Manuscript writing can take months, especially with complicated projects or those with many co-authors, as all co-authors need to approve the final version before it can be submitted. Once submitted, a Journal Editor will send the manuscript out to two or three Journal Reviewers, who are researchers in academia or industry that are in that field of expertise and can opt to volunteer to read and review the article. Nearly always, the authors do not know who the reviewers are, and in many cases the reviewers do not know who the authors are, although it is helpful for reviewers to see the authors’ names. If they have a conflict of interest with the author, such as they don’t get along personally, they might be married, or they are currently working on another project together (anything that might bias them for or against), the reviewers are supposed to decline to review. Reviewers have two to four weeks, depending on the journal, but some will submit their reviews late. The Editor considers all the reviews and makes their final decision to accept as is, accept with minor revisions, accept with major revisions, decline with major revisions (authors may edit and submit a new manuscript for consideration), or decline. It takes a few weeks to find reviewers, several more to get the revisions in, and another one or two for the editor to make a decision, so this can take anywhere from six weeks to four months.
Often journals will decline without reviewing if they are not interested in the subject material or feel it is outside the scope of the journal. If you have revisions, some journals request that you submit two new versions of the manuscript- one with the changes highlighted. Additionally, you need to address each reviewer comment by explaining what you did. For spelling mistakes, this is as simple as writing “corrected” after the comment. For more complex things, you need to explain the change along with quoting the new text, or explain why you aren’t changing things. If the Editor and Reviewers do not feel that you made all the changes, they may reject the re-submission or send you more edits. Usually they send you more edits that they didn’t notice the first time.
Eventually, a journal might accept your manuscript, and then you only have to approve the author proofs – unless your figures don’t have a high enough resolution, and then you need to remake them or figure out how to increase your dpi. Typically it takes between six months to a year to complete the whole peer-review process, depending on the study results and the journal’s internal process.
While tedious and arduous, the manuscript peer-reviewing procedure works very well. Experts in your field can assess the validity of your work, and experts in related fields can give you an outside perspective, especially when you have gotten used to using a very specific jargon or not completely explaining things. Most importantly, it improves the quality of the writing and presentation, and it maintains a standard of integrity and excellence. By the end of the submission process, you are dizzy and you want to get off the ride. But by the time you get through the next project, or eat a soft pretzel, you’ll be ready to climb back on that carousel horse.