How to Write the Discussion of a Scientific Paper

1. DISCUSSION AND VERBIAGE

The discussion (which some journals term a comment, especially for short papers) is harder to define than the other sections. Thus, it is usually the hard­est section to write. And, whether you know it or not, many papers are rejected by journal editors because of a faulty discussion, even though the data of the paper might be both valid and interesting. Even more likely, the true meaning of the data may be completely obscured by the interpretation presented in the discussion, again resulting in rejection.

Many, if not most, discussion sections are too long and verbose. As Doug Savile said, “Occasionally, I recognize what I call the squid technique: the author is doubtful about his facts or his reasoning and retreats behind a protective cloud of ink” (Tableau, September 1972). Another reason some discussions are long and hard to follow is that many authors think they must avoid first per­son. If you mean “I found that . . .” or “We conclude that, . . .” say so. Try to avoid wordier, and sometimes more ambiguous, constructions such as “It was found in the present investigation that . . .” and “It is concluded that.”

Some discussion sections remind one of the diplomat, described by Allen Drury in Advise and Consent (Garden City, NY: Doubleday, 1959, p. 47), who characteristically gave “answers which go winding and winding off through the interstices of the English language until they finally go shimmering away alto­gether and there is nothing left but utter confusion and a polite smile.”

2. COMPONENTS OF THE DISCUSSION

What are the essential features of a good discussion? The main components will be provided if the following injunctions are heeded.

1. Try to present the principles, relationships, and generalizations shown by the results. And bear in mind, in a good discussion, you discuss—you do not recapitulate—the results.
2. Point out any exceptions or any lack of correlation and define unsettled points. Never take the high-risk alternative of trying to cover up or fudge data that do not quite fit.
3. Show how your results and interpretations agree (or contrast) with pre­viously published work.
4. Don’t be shy; discuss the theoretical implications of your work, as well as any possible practical applications.
5. State your conclusions as clearly as possible.
6. Summarize your evidence for each conclusion. Or, as the wise old scien­tist will tell you, “Never assume anything except a 4-percent mortgage.”

Much as the methods and the results should correspond to each other, the introduction and the discussion should function as a pair. At least implicitly, the introduction should have posed one or more questions. The discussion should indicate what the findings say about the answers. Failure to address the initial questions commonly afflicts discussions. Be sure the discussion answers what the introduction asked.

Whereas the content of the introduction commonly moves from the general topic to your specific research, in sort of a funnel format, the discussion tends to do largely the reverse, much like an inverted funnel. For example, a well- structured discussion may first restate the main findings, then discuss how they relate to findings of previous research, then note implications and applications, and perhaps then identify unanswered questions well suited for future research. In the introduction, you invited readers into your research venue; in the dis­cussion, you usher them out, now well informed about your research and its meaning.

3. FACTUAL RELATIONSHIPS

In simple terms, the primary purpose of the discussion is to show the relation­ships among observed facts. To emphasize this point, the story may be told about the biologist who trained a flea.

After training the flea for many months, the biologist was able to get a response to certain commands. The most gratifying of the experiments was the one in which the professor would shout the command “Jump,” and the flea would leap into the air each time the command was given.

The professor was about to submit this remarkable feat to posterity via a scientific journal, but he—in the manner of the true scientist—decided to take his experiments one step further. He sought to determine the location of the receptor organ involved. In one experiment, he removed the legs of the flea, one at a time. The flea obligingly continued to jump upon command, but as each successive leg was removed, its jumps became less spectacular. Finally, with the removal of its last leg, the flea remained motionless. Time after time the command failed to get the usual response.

The professor decided that at last he could publish his findings. He set pen to paper and described in meticulous detail the experiments executed over the preceding months. His conclusion was one intended to startle the scientific world: When the legs of a flea are removed, the flea can no longer hear.

Claude Bishop, the dean of Canadian science editors, told a similar story. A science teacher set up a simple experiment to show her class the danger of alcohol. She set up two glasses, one containing water, the other containing gin. Into each she dropped a worm. The worm in the water swam merrily around. The worm in the gin quickly died. “What does this experiment prove?” she asked. A student from the back row piped up: “It proves that if you drink gin you won’t have worms.”

4. NOTING STRENGTHS AND LIMITATIONS

The discussion is a place to note substantial strengths and limitations ofresearch being reported. Some authors feel awkward about including such content. However, doing so can aid readers, and it can help show editors and referees (peer reviewers) that your work is publishable.

Some authors consider it immodest to note strengths of their work—for example, superior experimental techniques, large sample size, or long follow­up. However, such information can aid readers in determining how definitive the findings are. It also can help persuade peer reviewers and editors that your work deserves publication.

What if research had significant limitations—such as difficulties with a technique, a relatively small sample size, or relatively short follow-up? Some authors might try to hide such limitations. However, doing so runs counter to the openness that should characterize science. And astute reviewers, editors, or readers might well notice the limitations—and assume, either to themselves or in writing, that you were too naive to notice them. It is better, therefore, to identify substantial limitations yourself. In doing so, you may be able to discuss what impact, if any, the limitations are likely to have on the conclusions that can be drawn.

Not every discussion needs to discuss strengths or limitations of the research. However, if research has strengths or limitations major enough to be worthy of note, consider addressing them in the discussion.

5. SIGNIFICANCE OF THE PAPER

Too often, the significance of the results is not discussed or not discussed ade­quately. If the reader of the paper finds himself or herself asking “So what?” after reading the discussion, the chances are that the author became so engrossed with the trees (the data) that he or she didn’t really notice how much sunshine had appeared in the forest.

The discussion should end with a short summary or conclusion regarding the significance of the work. (In some journals, papers include a separate con­clusion section.) We like the way Anderson and Thistle (1947) said it: “Finally, good writing, like good music, has a fitting climax. Many a paper loses much of its effect because the clear stream of the discussion ends in a swampy delta.” Or, in the words of T.S. Eliot, many scientific papers end “Not with a bang but a whimper.”

6. DEFINING SCIENTIFIC TRUTH

In showing the relationships among observed facts, you do not need to reach cosmic conclusions. Seldom will you be able to illuminate the whole truth; more often, the best you can do is shine a spotlight on one area of the truth. Your one area of truth can be illuminated by your data; if you extrapolate to a bigger picture than that shown by your data, you may appear foolish to the point that even your data-supported conclusions are cast into doubt.

One of the more meaningful thoughts in poetry was expressed by Sir Rich­ard Burton in The Kasidah:

All Faith is false, all Faith is true;

Truth is the shattered mirror strown

In myriad bits; while each believes

His little bit the whole to own.

So exhibit your little piece of the mirror, or shine a spotlight on one area of the truth. The “whole truth” is a subject best left to the ignoramuses, who loudly proclaim its discovery every day.

When you describe the meaning of your little bit of truth, do it simply. The simplest statements evoke the most wisdom; verbose language and fancy tech­nical words are used to convey shallow thought.

Source: Gastel Barbara, Day Robert A. (2016), How to Write and Publish a Scientific Paper, Greenwood; 8th edition.