As a reviewer of over 400 scientific papers for conferences/journals, I love a well-written paper. The words sound right, the arguments feel logical, the sentences flow into each other at the right reading pace, the entire thing makes beautiful sense from start to finish. I feel a profound sense of satisfaction and gratitude when I finish reviewing such a paper because I feel like I had a sneak preview of a great work of art. The paper stays with me long after the fact.
Yes, research papers are accepted for publication because the underlying research is good. But, to be accepted, research papers must first be understood. This is where the quality of the writing makes a difference.
Great writing makes understanding feel effortless to the reviewer.
The effortlessness of reviewing evaporates when something I read jars on me. Vague claims, awkward phrasing, undefined acronyms, and convoluted sentences — all of these can strike a jarring note. Each of them is like a teeny speed bump in the road. A tiny distraction. As they pile up, these jarring notes are no longer a tiny distraction. They start to overshadow the research reported in the paper, the message the author is trying to convey. Jarring notes make my (reviewer’s) brain work harder. The research in the paper might be awesome, but my easily-distracted brain was lost in the sea of jarring notes. It’s like I’m driving on a highway with a series of speed bumps back to back, and I am unable to enjoy the scenery because I’m too busy slowing down and picking back up after each speed bump. My mind doesn’t recall the scenery, but can distinctly remember and dread the speed bumps.
“My most important piece of advice to all you would-be writers: When you write, try to leave out all the parts readers skip.”
— Elmore Leonard
As authors, we want the reader to drive through the paper, and fall in love with the scenery and not be distracted by any speed bumps. We want reading to feel effortless, and for our writing to bring joy.
Common Speed Bumps to Avoid
- Do not use passive voice. It’s better to say, “We ran the experiment 10 times,” instead of “The experiment was run 10 times.” My favorite way of identifying passive voice comes from Professor Rebecca Johnson at USMC. According to Professor Johnson, if you can add the phrase “by zombies” after the verb, your sentence has passive voice. The experiment was run 10 times “by zombies.” Indeed. Reviewers get distracted in trying to figure out who the subject is, when you use the passive voice in writing.
- Do not use hyperbole. Avoid the use of the following words in your scientific writing: very, extremely, highly, ultra, excessive, extreme, infinitely (unless it’s true), massively, simplistic, instantaneously (unless it’s true), rapidly, quickly, gargantuan, tons of (unless we’re talking about a real ton), lots of, humongous (please, not humongous). This is a scientific paper, not a phone conversation. Avoid using words that indicate quantity (lots, tons, rapidly, etc.) without specifying the actual quantity. It’s better to say “We ran 55 experiments” than to say “We ran a lot of experiments.” Vague descriptions of quantity do not have a place in a scientific paper. Reviewers get distracted by undefined quantities.
- Do not trash others’ work. When you describe others’ work, be objective, and describe both the pros and the cons of their work. Don’t disparage or belittle their work. Your own work should stand on, and be accepted for, its own merits. Your work is not going to be more loved or appreciated because someone else did worse, or because you trashed someone else’s work. Reviewers are mistrustful of authors who trash others’ work.
- Do not use pie-charts. Edward Tufte says, “The only worse design than a pie-chart is several of them.” The pie-chart is a common type of data visualization that forces the reviewer to compare slices of the pie visually, by comparing the angles of the slices. Pie-charts create more confusion and more work in the reviewer’s mind. 3D pie-charts are a horror, and have no place in a scientific paper. A bar chart or a simple table are always better than a pie-chart. Pie-charts create more work in the reviewer’s brain, which is a great reason to avoid them.
- Do not over-use acronyms. Every acronym is a point of possible distraction for the reader. Every acronym can make the reviewer attempt to recall what the acronym means (unless it’s a common acronym, such as HTML or TCP). If a reviewer has to go look up a word multiple times to keep up with your writing, you’re losing their attention. You’re making them work harder.
- Do not be overly broad in your claims. You don’t want to claim you’re solving every aspect of every problem in your field. There’s no way you can do that in a single research paper. You know it, and the reviewer knows it. Most research papers are about solving a narrow, but important, slice of a problem. Be precise in the specific problem. You lose credibility with the reviewer when you claim more than you ought to. It makes the reviewer skeptical of the rest of the paper.
- Do not name-drop the fashions of the day. Whenever there is a hot technology trend, there is a tendency for every paper to name-drop the hot technologies in the Introduction section of the paper. I see this in my own field, where the word “Cyber-Physical Systems” was the “it” word for a few years in every paper I that I reviewed, Now, it’s “IoT.” Following the fashionable word of the day does not set you apart — it makes your paper read just like everyone else’s. Reviewers get a sense of deja vu when they encounter the 100th name-drop of a technology trend.
- Do not turn your paper into a whodunnit. The reviewer needs to know, right up front, what’s new about your work, why it matters, and why they should continue reading. The reviewer is not a detective, and does not want to be one. The reviewer is not hanging on the edge of their seat waiting for the final shoe to drop. They need to know, from the beginning, that this paper is worth every minute of their time. Think of your paper as a magician’s act. The reviewer want to see the magic trick first. The trick pulls them in, makes them marvel, makes them want to know more. The reviewer is hooked. The rest of the paper then reveals how the magic trick worked.
- Do not extrapolate from a handful of data points. Unless you have enough data, and unless you’ve repeated the experiment with the right set of control variables and a statistically significant number of experimental runs, don’t attempt to infer linear or exponential (or any other type of ) behavior through wild curve-fitting. It’s always possible to draw a line through a bunch of selected points, and to call it linear. Whether the underlying phenomenon is linear in nature, is a different thing altogether. Reviewers are skeptical of claims of linear or exponential behavior without statistically-significant data to back it.
- Do not write long sentences. Long sentences are hard to write, and hard to read. Both the writer and the reader lose track of the subject, the verb, and the object. The reviewer scans long sentences, looking for natural breaks to pause the flow, seeking out commas, seeking out nouns and verbs, and trying to parse what you are saying. Long sentences are like driving down a mountain road with an endless set of hairpin bends — you can’t enjoy the scenery, you don’t know when the road ends, but you’re gripping the wheel for dear life at every turn, hoping it’s the last one.