CPR for Bad Scientific Writing

When your advisor or a peer reviewer writes on your paper comments such as, “unclear” or “ambiguous” or “obtuse”, do you think that s/he:

a)    is just picking on you
b)    doesn’t understand your unique “style”
c)    is making a subjective judgment
d)    might be right

If you picked d, you have taken the first step toward becoming a good writer. 

It’s true that some professors are a bit obsessed with grammar and punctuation, but the real obstacle to good scientific writing is poor style.  In fact, a paragraph with proper punctuation and grammar might be utterly incomprehensible, whereas one that has not followed grammatical rules is perfectly clear in its message.  Consider the following grammatically correct sentence:

Increasing foreign competition and technological change, in a variety of forms, are now, as they always have been, disrupting various well-established patterns in terms of industrial organization.

Difficult to understand, right?  That’s because the style is poor. 

Next, consider an ungrammatical, but quite understandable sentence:

The material applied to the blades of wind turbines age rapidly in tests.

It should be obvious that the first example would be much more difficult for an editor to fix than the second example.  I’m not saying that you should neglect grammar and punctuation.  My point is that style is crucial.

What you should strive for, above all else, is clarity in your writing style.  Clarity is an aspect of style that is particularly relevant to scientific writing, but is one of the most difficult for some scientists to achieve. So how do you develop clarity and other stylistic goals? 

If you are having extreme difficulty, there are a few things you can do immediately.  By attending to three aspects of writing: concision, precision, and revision (CPR), you can quickly improve your technical writing.

Concision

Begin by eliminating all unnecessary or meaningless words:  “it is noted”, “as we have seen”, “in terms of”.  By dropping extraneous words, we not only reduce the wordiness, but can better see what other revisions are necessary.

Precision

Precision can be improved by selecting those phrases that are not exact in their meaning and rewriting them.  Pay particular attention to those noun clusters that scientists are so fond of: “nutrient use efficiency respiratory rates” or “plant trait plasticity variation” The best way to improve a paragraph’s meaning is to choose clear, meaningful nouns and follow them with verbs that explain the noun’s meaning in the sentence: “Plasticity of plant traits varies among species.”

Revision

After cleansing our writing of excess verbiage and improving the meaning of the remaining words, we can now revise and improve the style of the writing.  The preceding posts provide guidance as to how to make transitions from one sentence to the next and to meet reader expectations in sentence structure.  However, some very simple approaches such as using connecting words (this, also, as well as, recently) will go a long way toward improving the reader’s understanding.

To take the next steps in improving your technical writing skills, you must delve into writing guides, particularly ones that focus on style and clarity.  The best known of these is The Elements of Style by Strunk and White.  There are many others readily found on the internet.

How To Handle Bizarre Behavior

Let's take a break from discussing scientific writing to consider some other concerns that PIs occasionally encounter and must respond to.

Today, I came across the following question in an e-zine for Principal Investigators:


One of my lab technicians has begun talking to himself and has become very argumentative. Several of my staff have confided they are worried he is developing a mental disorder--and thus we could be heading for some episode of "workplace violence". As PI and his supervisor, may I legally interview him about these personality changes? Should I? If I refer the matter to our institute's Human Resources department, could the technician accuse me of "defamation of character"? Do I have any duty to do something or anything?


If you are a PI, how would you handle this?  If you are a technician, post-doc, or student working in a lab with such a person, what would you do, if anything?  What would you want the PI and/or HR do?

Closing the Gap

"It may seem obvious that a scientific document is incomplete without the interpretation of the writer; it may not be so obvious that the document cannot "exist"without the interpretation of each reader."~Gopen and Swan (1990)


This is the final post in the series on the "Science of Scientific Writing" by the authors cited above.  The objective of their paper was to introduce some principles of writing that will help to close the gap between writer interpretation and reader comprehension. 

In the previous posts, we've worked through five principles designed to improve comprehension of scientific writing.  There are two more principles that provide further guidance in writing to meet reader expectations.  Here's number six:

In general, provide context for your reader before asking that reader to consider anything new.

When the writer fails to provide context, the reader is left floundering.  Writers often neglect to provide context because the information seems obvious to them and they fail to recognize that the reader may not be similarly acquainted.

The earthquake example we considered in previous posts did not provide context for the abstract, but instead jumped right into the technical information:

Large earthquakes along a given fault segment do not occur at random intervals because it takes time to accumulate the strain energy for the rupture. The rates at which tectonic plates move and accumulate strain at their boundaries are approximately uniform....rest of abstract.

Isabella's version added the necessary context by explaining in the first sentence that this discourse was about how strain buildup causes earthquakes:

Earthquakes occur when a certain amount of strain caused by the movement of tectonic plates has been accumulated. 

The reader is now prepared mentally to consider more technical aspects of earthquake frequency.  As the writer proceeds through the technical information, she should provide context for each new bit of information introduced in the piece.  This approach requires that the writer ask herself if the reader will understand what is being introduced in each succeeding sentence or if some explanation or backward linkage to "old information" is required.


We've now come to the final principle, which is:

In general, try to ensure that the relative emphases of the substance coincide with the relative expectations for emphasis raised by the structure.

The foregoing principles relate to sentence structure and how it sets up reader expectations.  For example, the reader expects to see the person or thing that the discourse is about in the "topic position" near the beginning of the sentence.

Squirrels hide acorns. The topic is squirrels.
Acorns are hidden by squirrels.  The topic is acorns (or oak trees, seed dispersal).

The writer must choose the appropriate sentence structure that is consistent with both the material being presented and with reader expectations.  If the reader has been told that the discourse is about acorns and seed dispersal, then the second sentence provides the substance to be emphasized in the expected position in the sentence.

We've now covered all seven principles of scientific writing:

 1.  Follow a grammatical subject as soon as possible with its verb.
 2.  Place in the stress position the "new information" you want the reader to emphasize.
 3.  Place the person or thing whose "story" a sentence is telling at the beginning of the sentence, in the topic position.
 4.  Place appropriate "old information" (material already stated in the discourse) in the topic position for linkage backward and contextualization forward.
 5.  Articulate the action of every clause or sentence in its verb.
 6.  In general, provide context for your reader before asking that reader to consider anything new.
 7.  In general, try to ensure that the relative emphases of the substance coincide with the relative expectations for emphasis raised by the structure. 



Keep in mind that these are principles, not rules.  You will not necessarily be able to apply all seven in every sentence or even in every paragraph.  In some cases, you may have to make a choice between two structures. Also, some highly skilled writers can violate reader expectations quite effectively, for example to make a memorable point.  

The key is to recognize when you consistently violate reader expectations in one or more of these principles.  A writer who continually fails to put new information in the stress position during early writing attempts typically continues that structural pattern in subsequent writing.  It becomes a habit that is difficult to break, particularly if the writer is unaware of how it affects reader comprehension.  If you get reviewer comments that your writing is "unclear" or "ambiguous" on a frequent basis, you may be violating one of these principles.

As a further exercise, I suggest you select a couple of papers in your field--one that you think is particularly good (clear, understandable, compelling) and another that is difficult to follow, that requires frequent rereading of sentences to understand.  From those papers select a paragraph or two and dissect them based on what we've covered in this series.  I think you'll find that the better paper adheres to the seven principles, and the difficult paper violates one or more of them.  By doing this exercise with someone else's writing, you develop an "eye" for discourse that needs revision to improve reader comprehension.  Once you've become proficient at spotting problematic structure in other writings, then you are prepared to tackle your own writing.

Editing your own writing is not as easy as it sounds.  It's easy to become enamored of your own words and sentence structures and are loathe to change them.  I find this infatuation with one's own words quite insidious and difficult to overcome.  Students are particularly prone to this condition.  They think that every sentence is a pearl of wisdom that needs no revision.  So they are quite shocked to get their work returned, and it is covered with "red ink".  Some are so obstinate that they refuse to make the suggested changes (which is why I always keep a copy of my marked-up version to compare with their revision). Such students never improve and continue to have increasingly difficult problems.  Other students learn quickly because they take the time to consider what their mistakes were and why they need revision. 

Even seasoned writers have this problem of falling in love with their writing (maybe even more so than others).  However, that sentence, which you worked so hard to produce and that are now so proud of, may be confusing to the reader.  If it is, it needs to be revised.  You must learn to be ruthless with your own writing.  If you find yourself balking at changing a sentence you know to be flawed, tell yourself that the new sentence will also be your creation and an even better one than the original.  I also agree with the common advice of putting your writing away for awhile to get some distance from it.  By distancing yourself, you not only can take a fresh look at your writing at a later date, but you minimize your feeling of ownership.  With a less possessive attitude, you can more easily rearrange and discard words.

I hope readers have learned something in this series.  I know I have.  It's made me take an even closer look at my writing tendencies and in particular my favorite sentence structures that need improvement.  I still find myself falling back into bad habits on occasion, but now I can more easily spot those indiscretions.  By becoming aware of how poor constructions affect reader comprehension, we can resist and consciously change such habits.

Reader Eggspectations

This post continues the series on principles of scientific writing as described by Gopen and Swan (1990).   We are now at the fifth principle, which deals with verbs (again).  The following is an example given by Gopen and Swan.

See if you can figure out the problem.

Transcription of the 5S RNA genes in the egg extract is TFIIIA-dependent. This is surprising, because the concentration of TFIIIA is the same as in the oocyte nuclear extract. The other transcription factors and RNA polymerase III are presumed to be in excess over available TFIIIA, because tRNA genes are transcribed in the egg extract. The addition of egg extract to the oocyte nuclear extract has two effects on transcription efficiency. First, there is a general inhibition of transcription that can be alleviated in part by supplementation with high concentrations of RNA polymerase III.  Second, egg extract destabilizes transcription complexes formed with oocyte but not somatic 5S RNA genes.

If you've been following along, you should be able to identify some of the problems with this passage.  However, one of the main problems here is that the topics are apparent, but not what they are supposed to be doing.  Readers expect the action of a sentence to be articulated by the verb.  This passage violates that expectation.

Here are the verbs from that paragraph:

     is
     is...is
     are presumed to be
     are transcribed
     has
     is...can be alleviated
     destabilizes

Just looking at this list, we recognize that the passage tells us very little about what actions are actually taking place.

Here is the revision with the new verbs highlighted:

In the egg extract, the availability of TFIIIA limits transcription of the 5S RNA genes. This is surprising because the same concentration of TFIIIA does not limit transcription in the oocyte nuclear extract. In the egg extract, transcription is not limited by RNA polymerase or other factors because transcription of tRNA genes indicates that these factors are in excess over available TFIIIA. When added to the nuclear extract, the egg extract affected the efficiency of transcription in two ways. First, it inhibited transcription generally; this inhibition could be alleviated in part by supplementing the mixture with high concentrations of RNA polymerase III. Second, the egg extract destabilized transcription complexes formed by oocyte but not by somatic 5S genes.

This revision does not fix all the problems, but goes a long way toward improving comprehension.  The verbs are now telling us something about what actions were taken in this study.  Without knowing more about the topic and what the authors were really trying to get across, it's difficult to do more with it.  However, it serves to illustrate the fifth principle:

Articulate the action of every clause or sentence in its verb.

How to Improve Comprehension of Scientific Writing

The previous series of posts has focused on what features of scientific writing interfere with reader comprehension and how we can change our writing habits to better meet reader expectations.  These ideas are taken from the paper, "The Science of Scientific Writing", by Gopen and Swan (199).

In this post, I will show how Gopen and Swan revised the previous writing example to improve comprehension.  The writing principle we are considering requires that for each sentence, a backward linkage is made to "old information" and new information is positioned for contextualization forward.  The first step is to identify the new material in each sentence that should be emphasized.  I've highlighted these phrases in yellow:

Large earthquakes along a given fault segment do not occur at random intervals because it takes time to accumulate the strain energy for the rupture. The rates at which tectonic plates move and accumulate strain at their boundaries are approximately uniform. Therefore, in first approximation, one may expect that large ruptures of the same fault segment will occur at approximately constant time intervals. If subsequent main shocks have different amounts of slip across the fault, then the recurrence time may vary, and the basic idea of periodic mainshocks must be modified. For great plate boundary ruptures the length and slip often vary by a factor of 2. Along the southern segment of the San Andreas fault the recurrence interval is 145 years with variations of several decades. The smaller the standard deviation of the average recurrence interval, the more specific could be the long term prediction of a future mainshock. 

What needs to be done for each sentence is listed below:

   1. The backward-linking old information appears in the topic position.
   2. The person, thing or concept whose story it is appears in the topic position.
   3. The new, emphasis-worthy information appears in the stress position. 



Here is the revision suggested by Gopen and Swan based on the above guidelines, with the new information (in the stress position) highlighted in yellow and the old information (backward linkage) highlighted in pink:

Large earthquakes along a given fault segment do not occur at random intervals because it takes time to accumulate the strain energy for the rupture. The rates at which tectonic plates move and accumulate strain at their boundaries are roughly uniform. Therefore, nearly constant time intervals (at first approximation) would be expected between large ruptures of the same fault segment. [However?], the recurrence time may vary; the basic idea of periodic mainshocks may need to be modified if subsequent mainshocks have different amounts of slip across the fault. [Indeed?], the length and slip of great plate boundary ruptures often vary by a factor of 2. [For example?], the recurrence intervals along the southern segment of the San Andreas fault is 145 years with variations of several decades. The smaller the standard deviation of the average recurrence interval, the more specific could be the long term prediction of a future mainshock. 

So, how did you do?  Check out Isabella's version of this paragraph--I find hers easier to understand than Gopen and Swan's revision.  Her revision is written in simple, straightforward language and only needs a bit of polishing.  Here is my modification to her version:

Earthquakes occur when a certain amount of strain caused by the movement of tectonic plates has been accumulated. The rates of these movements at the boundaries of tectonic plates are roughly uniform, leading to the expectation that large earthquakes along a given fault segment occur at nearly constant and nonrandom time intervals (at first approximation), because it takes time to accumulate the strain energy for the rupture. However, if subsequent main shocks have different amounts of slip across the fault, then the recurrence time may vary, and the basic idea of periodic mainshocks must be modified.  For great plate boundary ruptures, the length and slip often vary by a factor of 2, resulting in a large standard deviation of the average recurrence interval, which leads to a less specific long-term prediction of a future mainshock. For example, along the southern segment of the San Andreas fault the recurrence interval is 145 years with variations of several decades.

Gopen and Swan state that in their experience, "the misplacement of old and new information turns out to be the No. 1 problem in American professional writing today."

The fourth principle can be stated as:

Place appropriate "old information" (material already stated in the discourse) in the topic position for linkage backward and contextualization forward. 

I find this principle the most difficult one to apply routinely to my writing because as a writer I am anxious to get my idea (the new stuff) down before I forget my brilliant thought.  Consequently, the temptation is to rush the new information and forget to provide backward linking and context for it.  However, by doing this we only address our needs, not the needs of the reader.

Scientific Writing Principles Cont'd

This post continues a series on principles of scientific writing, as proposed by Gopen and Swan (1990).  We've already covered the first three principles (to understand the material in this post, you need to read the previous material starting here).
  
Readers expect a unit of discourse (a sentence) to be a story about whoever or whatever is mentioned first, i.e., in the topic position.  The example I've been using is "squirrels hide acorns" versus "acorns are hidden by squirrels".  The first version suggests the focus is on squirrels, whereas the second one indicates the focus is on oak trees and/or seed dispersal.  These examples also illustrate active versus passive voice.

The importance of the topic position forms the third principle of scientific writing:

Place the person or thing whose "story" a sentence is telling at the beginning of the sentence,  in the topic position. 

Another expectation for material in the topic position is that it provides a linkage backward (to previous information).   Previously introduced material is called "old information", and its placement in the topic position helps readers follow the author's logic.  Now we consider the fourth principle, which can be stated this way:

Place appropriate "old information" (material already stated in the discourse) in the topic position for linkage backward and put in the stress position the information you want the reader to emphasize (for contextualization forward).

Have we done this with our example?  In the abstract below, I've highlighted in pink the "old information" that should be in the topic position.  New information (yellow) should be in the "stress position" for contextualization forward.

The discrete-dipole approximation (DDA) is [often] used in scattering calculations, but its accuracy is unclear in relation to that of other computational methods such as complex-conjugate gradient algorithms and fast-Fourier-transform methods.  The accuracy of the DDA was tested in computations of scattering and absorption by different targets: isolated, homogeneous spheres and two contiguous spheres. For dielectric materials (¦m¦ ≲ 2), the DDA permitted calculations that were accurate to within a few percent.

 We've done a pretty good job of placing the old information in the topic position for linkage backward and new information in the stress position.  The only quibble might be with the phrase "For dielectric materials".  This term is suddenly introduced without explanation, but presumably would be understandable to experts in this field. 

To illustrate this principle further, I will use the example given in Gopen and Swan:

Large earthquakes along a given fault segment do not occur at random intervals because it takes time to accumulate the strain energy for the rupture. The rates at which tectonic plates move and accumulate strain at their boundaries are approximately uniform. Therefore, in first approximation, one may expect that large ruptures of the same fault segment will occur at approximately constant time intervals. If subsequent main shocks have different amounts of slip across the fault, then the recurrence time may vary, and the basic idea of periodic mainshocks must be modified. For great plate boundary ruptures the length and slip often vary by a factor of 2. Along the southern segment of the San Andreas fault the recurrence interval is 145 years with variations of several decades. The smaller the standard deviation of the average recurrence interval, the more specific could be the long term prediction of a future mainshock.

At first read, one has the impression that this is a fairly straightforward description of earthquakes and tectonic plates and is written well.  However, by the time we reach the end of the passage we are feeling somewhat befuddled and would be hard-pressed to say exactly what the point of this paragraph was.  Gopen and Swan analyze this example and conclude that the main problem is that virtually every piece of new information makes its first appearance in the spot we expect to find the old, familiar information.

If your writing continually begins sentences with new information and ends with the old information, you will definitely disorient your readers and reduce comprehension, as the above example does.  So how would one go about fixing the above paragraph?  Take a second look at the fourth principle stated above and see if you can rewrite the paragraph.

I'll provide the revision suggested by Gopen and Swan in the next post along with some additional explanations.

A Slight Detour

Active versus Passive Voice

We are discussing some principles of scientific writing as proposed by Gopen and Swan (1990).  But let's make a brief detour at this point to consider active versus passive voice.  In our previous examples, we did some shifting from passive to active voice and vice-versa.  Some people misunderstand what is meant by active voice, thinking that it only involves first-person subjects:  "we discuss this...I review that..." or involves a form of the verb to be: "I am a scientist."

In actuality, a sentence written in passive voice has the subject acted upon by the verb.  For example:

"Accurate calculations are permitted by DDA."  "The method is reviewed." "John is loved by Mary."

In active voice, the subject is doing the action.  The same sentences written in active voice:

"DDA permits accurate calculations." "We review the method." "Mary loves John."

Use of the passive voice is prevalent in scientific writing and is a big part of why technical writing can sound stilted (or even "tortured" as CPP puts it):  "Data analysis was preceded by log transformation."  There are many other examples, perfected by politicians, in which this is carried to the extreme:  "Mistakes were made." As illustrated above, the active voice makes sentences come alive--and also tends to require fewer words.


However, there are instances in which the passive voice is the proper choice.

--When it makes sense to emphasize the receiver of the action.  The example I gave in the previous post was "Squirrels hide acorns." (active) versus "Acorns are hidden by squirrels." (passive). If your topic is oak trees (and acorns), you want to select the passive voice version because this tells the reader that you are going to be talking mainly about acorns, not squirrels.

--For variety or when the flow of the paragraph would be improved: "One aspect of the nitrogen cycle can be understood through an examination of microbial processes. These processes . . ."

--In the Methods or Experimental section of a paper where repeated use of first person or people's names would be awkward: "Initially, the experimental plots were clipped to remove standing vegetation and then amended with a slow-release NPK fertilizer."

--In an abstract: "Feeding trials were conducted to assess the role of phenolic compounds in plant palatability."

So to summarize active vs. passive voice, you can use active voice to liven up your writing and be less wordy at the same time.  These considerations must be balanced, however, against what subject you wish to emphasize and if passive construction is more appropriate.  The key is to make your selection coincide with the reader's needs and expectations.