Writing quality
Why journal editors think that the quality of scientific writing is generally bad
Leslie Sage, a senior editor at a Nature journal, once wrote in a book chapter “… most papers are badly written … Some scientists who are appallingly bad writers even believe that they are good.” A didactic editorial in a Nature journal has the title “Scientific writing 101,” implying that many researchers have yet to learn the basics of good scientific writing. The displeasure of journal editors was summarized in that editorial in the following words: “editors and reviewers … must slog through papers that seem to go on forever and, more dishearteningly, have the main points and interesting bits inexplicably hidden.”
The purpose of scientific writing is not to impress or confuse with words. It is to convey the information and the message to interested scientists worldwide simply, clearly and concisely. The poor quality of scientific writing has several aspects, including chronological presentation rather than provision of a clear and logical account, abstracts that emphasize details rather than presenting the overall picture, discussions that are a rehash of the results, and lack of clarity.
Lack of clarity is not necessarily caused by inadequate knowledge of the English language. Sage states “two of the worst written papers I have seen in
my time at Nature came from native-English speakers based at a major UK university.” I too have come across papers that resist reading written by native English speakers in major universities in the UK and the US. The problem lies primarily in the complex writing style adopted by most scientists and transmitted through mentoring from one generation of researchers to the next. I have discussed this problem in an article I published in EMBO Reports.
When scientists who are not native speakers of English try to write in the commonly used writing style (most of them do), the number of grammatical mistakes multiplies. Grammatical mistakes are of two functional types. Benign mistakes are obvious to someone who is fluent in English language. They can be easily corrected and many of them do not impede comprehension. On the other hand, insidious grammar mistakes are not grammar mistakes as such. They are mistakes in the sense that they convey a meaning other than what the author intended. Insidious grammar mistakes can have a substantial negative effect on the publication process, and I will discuss them in a future post.
What should you avoid and what should you do to improve the clarity and precision of your research writing?
An article in Science states that scientists use “sophisticated words and complex grammatical constructions that can disrupt reading comprehension.” Most of this complexity is not caused by the complexity of science but by adherence to a complex and awkward writing style. When Watson and Crick described the structure of DNA in Nature in 1953, they did not use “sophisticated” words. Moreover, the average length of the sentences in their paper is one and a quarter lines, which is too short for any “complex grammatical constructions.”
“All the bases are flat, and since they are stacked roughly one above another like a pile of pennies, it makes no difference which pair is neighbor to which.”
A major problem in scientific writing is poor sentence construction. Even if the grammar is correct, poorly constructed sentences are more difficult to read and understand, they might be interpreted in different ways, and they might even misrepresent the intention of the author.
Long sentences: Long sentences are not necessarily difficult to understand, but writing them requires good writing skills and a strong command of the language. If either of these is absent, long sentences often end up being defective in grammar and/or structure.
Long subject‒verb distance: If the subject is separated from its verb by many words, one of two things can happen. (1) The reader focuses on the intervening text and loses track of the main sentence. (2) The reader simply scans over the intervening text while anticipating the verb in order to make sense of the sentence. This separation is particularly risky with the passive voice because the verb comes at or near the end of the sentence. This is the main reason why journals ask their authors to write in the active voice.
Long subordinate clauses: A subordinate clause contains a subject and a verb but cannot stand on its own. It is used to provide additional information but is not essential for the sentence. It is separated from the main sentence by a pair of commas or parentheses. If these clauses are too long or too many, they interrupt the sense of the sentence.
Rambling sentence: A rambling sentence usually contains several independent clauses, each of which can stand on its own like a sentence. If there are more than two such clauses they give the sense that the sentence has no end, which can be disorienting. In the following example, the colon (:) and semicolon (;) are used correctly to separate the clauses, but the presence of 3 clauses, 78 words and 8 verbs in one sentence is not a formula for clear writing.
“Downstream of cadherin-bound PG, PVIgG negatively interfere with a signaling pathway that relates to terminal differentiation and involves GSK3b: the transiently enhanced turnover of Dsg3 and PG in response to PVIgG was followed by delayed terminal differentiation and consequently sustained proliferation; GSK3b inhibition was sufficient to abrogate blister formation in PVIgG-injected neonatal mice, supporting the possibility that the PVIgG-targeted signaling pathway involves the PI3K/GSK3b axis that in turn is in control of PG nuclear trafficking and consequently c-Myc.”
Scientific jargon: Jargon is essential for science, but it should be used for necessity rather than routinely. Journals, especially those that are for wider audiences, advise against excessive use of jargon. “Produced a higher level of protein X” is ubiquitous in the scientific literature. It simply means “produced more protein X.” “The number of cells was quantified” means “The cells were counted.”
Incorrect position of modifiers: Misplacement of adjectives, adverbs and phrases that play those roles can change the meaning of the sentence completely. In the following example the four sentences have different meanings depending on the placement of “only.”
– Only the surgeon examined Peter yesterday.
– The surgeon examined only Peter yesterday.
– The surgeon examined Peter only yesterday.
– The surgeon only examined Peter yesterday.
Series of “and” and “or”: It is legitimate to have several instances of “and” and “or” in a sentence, but unless they are ordered and written correctly the meaning becomes incomprehensible or comprehensible but wrong.
That & which: The wrong choice between “that” and “which” can flip the meaning. “We removed the tumors that were on the superior surface” would raise a question about why the tumors on the inferior surface were not removed.” Though that and which are sometimes used interchangeably in general writing, in science they should be used specifically.
Punctuation: Correct punctuation facilitates reading and clarifies the meaning. But incorrect punctuation can even change the meaning. The comical example is “Let’s eat, John” versus “Let’s eat John.” Here, our common sense saves the day, but in scientific writing incorrect punctuation can cause confusion.
Verbosity: Scientific writing should be concise. But scientists favor phrases over single words that have exactly the same meaning. “Due to the fact that we observed multiple cases of hepatitis C infections ..” can be shortened by half to “Because we observed many hepatitis C infections …”
Excellent examples of confused writing transformed by editing
To illustrate what an editor does to render impenetrable text into comprehensible prose, I will show you two poorly written sentences and their edited versions. I emailed the first sentence to three friends who are research scientists experienced in the relevant field: each one had to read it three times in order to understand it, and one of them understood it incorrectly. So before reading the edited version, challenge yourself by reading each original sentence continuously from beginning to end to see how many times it will take you to understand it.
Example 1
“Herein, we show enhanced radiation-induced (10 Gy) tumor growth delay in a syngeneic model (C3H) but not immunosuppressed (Nu/Nu) squamous cell carcinoma tumor-bearing mice treated post-IR with the constitutive NOS inhibitor N(G)-nitro-l-arginine methyl ester (L-NAME).”
Such sentences are common in the scientific literature. It is the third sentence in the abstract, certainly not a good start. The sentence defies reading because it has several problems: reversal of topic and stress positions, overly long in the sense that it crams too much information, serious abuse of the noun phrase, and a grammar mistake.
Do not think that such awkward sentences are written only by inexperienced authors or those working in countries where English is not the native tongue. This sentence comes from a paper published in a highly respected journal (Cancer Research), and it has 18 authors distributed among four different departments of the National Cancer Institute in the United States.
Had the manuscript gone through my hands, I would have edited as follows:
“In a syngeneic model, growth of squamous cell carcinoma in C3H and immunosuppressed (Nu/Nu) mice was delayed by 10 Gy irradiation. Post-irradiation treatment with the constitutive NOS inhibitor N(G)-nitro-l-arginine methyl ester (L-NAME) further slowed tumor growth in CH3 mice but not in Nu/Nu mice.”
Example 2
“We adopt this broad-scale approach to determine that relationships occur both at the level of the population (and hence not confounded by (1) potential environmental variation and/or (2) statistical nonindependence of individuals) and also across individuals (because (1) relatively recent colonization of the UK by rabbits [15], and (2) previous work [18] demonstrating extremely fine-scale genetic structuring in UK rabbits over short spatial scales both make it difficult to define what constitutes a ‘population’ for analysis).”
This is another example of exceptionally bad writing in the scientific literature. This paper, published in Current Biology, is authored by seven scientists working at two respected universities in the United Kingdom.
This sentence too has many problems: too long (76 words) and packed with too much information, unneeded serial numbers for two phrases and one clause intermingled with the necessary reference numbers, most of the text is in parentheses, the main sentence is split by a long dependent clause, and the terminal end of the sentence is followed by long parenthetic material, which leaves the reader expecting that the main sentence might continue after the end of the parentheses.
My edited version below is somewhat longer, but that is a cheap price to pay for clarity.
“We adopted this broad-scale approach to determine that relationships occur at the level of the population and also across individuals. We determined that relationships occur at the level of the population and are thus not confounded by potential environmental variation and/or statistical nonindependence of individuals. We also worked at the level of individuals because it is difficult to define what constitutes a ‘population’ for analysis. This difficulty is due to the relatively recent colonization of the UK by rabbits [15] and because previous work demonstrated extremely fine-scale genetic structuring in UK rabbits over short spatial scales [18].”
One message I want to leave with you is that an editor’s job is not limited to dotting i’s and crossing t’s. The other message is that even papers from top institutions in English speaking countries are plagued by poor writing caused by inadequate writing skills or unawareness of the qualities of good writing.