You, a young researcher, stand at a crossroads. Behind you are hours of labor and reams of data you've collected. Before you is a blank page. For many scientists, that blank page is daunting. You probably never took a class in scientific writing. I certainly never did, and my 18 years of on-the-job experience only served to cultivate a dry, formulaic accounting of procedures and outcomes. Writing scientific articles was not something I looked forward to, even after doing it some 200 times.

Then I went on sabbatical and did something a little offbeat: I enrolled in a graduate program in writing children's literature. My bag? Science-based mystery stories for kids.

I cannot report that my book sales rival those of J. K. Rowling, but I did learn some important things. I learned that writing, whether it's a children's book or a scientific article, is all about uncovering, and then telling, stories.

The story

Prior to my stint as a children's book writer, my writing revolved around checklists. My checklist included a directive that the Introduction should be a brief accounting of the previous literature. Next on the list came the Methods, with its description of all aspects of experimental design and conduct. Results should contain descriptive statistics, my checklist said, along with main outcomes and secondary outcomes. Discussion must put the research findings into context, describe strengths and weaknesses, and end with a concise summary of the most compelling findings.

After my fiction-writing gig, I didn't give up those checklists. Scientific writing remains an exercise in convincing another scientist of the validity of your work, and that requires a detail-intensive approach. But I added a new dimension to my writing: the story. Will it be interesting to the anticipated audience? Will it surprise them or reassure them? Will they walk away feeling they've discovered something along with me? Mind you, the audience might be both small and highly specialized. It might be a bunch of balding guys who get really excited about things that would be a complete yawn for most people. But these are specialists in an area in which I also get excited about ridiculous minutiae, and so I know what will turn them on. They are my audience, and they are like me--except maybe the parts about balding and being a guy.

Of course, there is no direct mapping between the elements of a scientific paper and the elements of a mystery story. Indeed, there are many approaches to writing a mystery story and many purposes to a scientific paper. Moreover, a mystery story answers one big question--Whodunit?--whereas scientific papers commonly address modest slices of a question, such as (to put it in a mystery-story context) how they jimmied the door, or whether the paint was damaged when they did. Nonetheless, the need to uncover a story is common to both, so I think some parallels can usefully be drawn. For example, Methods is like the detective work at the crime scene. The Results are the findings leading to the solution. Discussion is the conflict resolution.

Detective work at the scene of the crime: Methods

Details about how the inspector undertook examination of the crime scene, be they visual or in a laboratory, provide the reader with confidence about the ultimate resolution of a mystery. What approach did the detective take to assessing the path the bullets took? What methods were used to produce the evidence that an intruder was involved? As a scientific author you must describe, with the same richness, your approach--i.e., how you went about evaluating the "crime scene."

Experiments and observations may be conducted on humans, animals, cell cultures, chemical platforms, etc. They may include assays, interventions, or manipulations--whatever techniques you used to collect relevant data. Methods describes the study subjects (be they mice or men), as well as how you assessed exposures and outcomes (or whatever parameters are relevant to your scientific work), using enough specific detail and precision that another scientist could replicate what you did. Important details include how your key measurement strategies stand up to scrutiny: comparisons made, and their reliability and validity. It may not sound like the stuff bestsellers are made of, but for your audience, as for you, it's what sells a scientific paper.

The climax: Results

Classically, in literature, there are two ways to get to the story climax. One is to build up to it without getting lost along the way. The other is to divulge upfront how things turn out and then tell how you got to it. The latter is my favorite for scientific writing, because it ensures that you deliver your main point before losing the reader. In any case, it's probably the only possibility, because you are certain to give away the climax in the abstract anyway. When it's done well, the story need not lose any of its impact.

After you've gotten to the point--your main results--you need to provide an insider's perspective. One aspect is describing negative control results, which, in the parlance of a "whodunit," establishes who didn't do it. Other inside information may include ancillary observations or additional experiments, all leading to the same conclusion. You may also want to take a step back from your main findings by examining them in subgroups and anticipating reviewers' skepticism by presenting experiments, data analysis, or statistical models that quash errant disbelief.

Of course, you cannot make it up the way a fiction writer can, but you can present your story honestly and skillfully so the reader is as convinced as you are of the interpretation you are now going to present.

Resolution of the conflict: Discussion

Which brings us to the Discussion section. The end of a mystery story must resolve the conflict by revisiting the (newly revised) status quo. Your paper's Discussion section must provide a feeling of resolution to the conflict created by your results, in much the same way as a work of fiction ties up loose ends and provides a feeling of resolution. You may not have reached a definitive end--you may not know yet "whodunit"--but you have narrowed the possibilities with your research.

Start this section with a brief summary of your main results. Next, make connections between these new results and the prior literature mentioned in the Introduction. How do known physiology, animal experiments, or ecological data provide a context for your findings? How does it support or contrast with your findings? Are there any aspects of your study design, participant characteristics, or measurements that can explain divergent results? Finally, come clean with any and all limitations in the way you conducted your research and explain to the reader how this may or may not have biased your findings.

Final notes of advice

Now that I've described an ideal, let's get practical. Figures and tables in a paper should tell the story without forcing the reader to read the text. Yes, they're missing out if they don't read the whole, exciting yarn, but a lot of people prefer to save the time and jump right to the pictures. You can get the whole exercise started by drafting the figures, then writing around them.

Second, although J. K. Rowling may not agree, brevity is a virtue. Say as much as you must to illustrate your logic and convince readers that your interpretation is correct--but not one word more. Rambling is the enemy of suspense. When I write a paper, I complete a final draft and then shorten it by as much as a third.

Next, remember the basics. Paragraphs should begin with a topic sentence and contain only information on that topic. Write in an active voice; instead of writing, "It was found," go ahead and say who found it: "We found," or "they found." You can't tell a good crime story without detectives.

Finally, writing well comes from writing a lot. Some writers begin with an outline, and others don't. Some write everything in their head, and others begin to formulate the story only after sitting down to type. No matter what technique you use, good papers require dozens of drafts. So get comfy with your computer.

Now, back to my own experience. Where did fiction writing get me? My children's books never saw the inside of a bookstore. My yearly scientific publication rate did go up a good bit, however. It's hard to know whether the increase is attributable to narrative improvement or to age and seniority. But without question, I grew to love the process of writing in a way I never had before. Editors and colleagues began to comment that reading my papers made them feel the same excitement that I must have felt when I first saw my data. In my scientific writing, I still take the pride I have always taken in ideas and data. But now I also take pride in a story well told.

Roberta Ness is a professor and chair of the epidemiology department at the Graduate School of Public Health at the University of Pittsburgh in Pennsylvania. She has written about 200 scientific articles.

Comments, suggestions? Please send your feedback to our editor.

DOI: 10.1126/science.caredit.a0700047

10.1126/science.caredit.a0700047