Choose the structure and the reporting guideline that fix what each part carries
A report of original research takes the IMRaD sequence of introduction, methods, results and discussion, in the variant its target journal prescribes. The sequence answers the questions every scientific report has to answer and carries a story from the first section to the last. A reporting guideline's checklist sets out what each part has to disclose, and that detail has to fit inside the journal's word limit. The form a report takes also varies with the functions it serves and with how the research was planned.
The IMRaD sequence
Articles about original research tend to follow a standard format called IMRaD: introduction, methods, results and discussion. Most research manuscripts and most scientific papers are now organised this way.
The introduction gives the background and says why the study was done, the methods how it was done, the results what it found, and the discussion what the findings mean. The introduction also says why the topic is important, what prior research has been done and what question the study set out to answer, the methods say who the subjects were, what was done to them and how the data were analysed, and the discussion says what cautions apply to interpreting the findings. Clinical documents and technical reports may contain every detail of a study and hundreds of tables or figures, but a manuscript needs to describe only the highlights, and unlike a clinical study report it must put the outcome in the context of the current literature.
For a peer-reviewed article reporting research data, IMRaD is often the prescribed model, and every report of original research also needs an abstract, although the abstract is not part of the acronym. A report should follow the general model, because a structure of the author's own invention may invite summary rejection. Not every journal follows the model precisely, and some use variations of their own, so preparing a report begins with the target journal's instructions for authors, whose directions are followed carefully. In 2012 the Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Manuscript Preparation and Submission were a very useful general description of the structure of a scientific article.
The standard structure makes developing and publishing manuscripts more consistent and easier to manage. It helps authors organise their ideas and include all the important elements of the study, helps editors and reviewers evaluate manuscripts, and helps readers locate specific information quickly without going through the whole paper. Researchers sometimes assume that the key to writing a good research manuscript is a good command of language, but it is really understanding the manuscript's structure.
The questions the sequence answers
Every form of scientific communication must start with the answer to the question Why did you start?, and the questions that must be answered next, in order, are What did you do?, What answer did you get? and What does it mean?, in the outline Austin Bradford Hill presented in the 1960s. The answers to these questions are the basis of the IMRaD structure.
One of the main problems in most draft manuscripts is that the question being addressed is not stated clearly. Readers who are not given the context of the study are unable to situate its importance within the field. Once the author has the answer to the first question clearly in mind, the rest of the paper is easier to write.
Where the sequence came from
The first scientific journal, Philosophical Transactions, was founded in 1665, and for more than two centuries afterwards scientific manuscripts had no standardised form or style. The IMRaD story began with Pasteur, who developed the idea that a paper should be written so that the same results could be reproduced if the procedure was followed exactly. A standard structure began to emerge in the 1940s, and in the 1980s it was widely adopted for original research manuscripts in the sciences.
How the sections connect
The sections and their order let readers move logically from an overview of the research in the abstract to the rationale for the study in the introduction, the experiments in the methods, the findings in the results, and the significance of the findings and their implications for the field in the discussion. This structure can be drawn as an hourglass: the ideas run from a broad introduction down to the specifics of the objectives and methods, then back out to the broader clinical relevance of the findings in the discussion, and at important points they are linked back to each other to create flow and strengthen the logical presentation of the research.
Linking the sections of a manuscript and the information within them can be called manuscript flow. Even writers who know what belongs in each section are frequently unsure how to link it all together so that it flows smoothly from one idea to the next. A scientific or medical manuscript essentially tells a story about what happened, and imagining that the reader is being guided through a presentation of what was done and what was found helps create a logical flow of information within each section and from one section to the next.
The sequence as a story
IMRaD helps authors tell a story with the most common story structure, the three-act structure. Its three acts are separated by two plot points. The setup introduces the setting and the characters, the tension, or inciting incident, at the first plot point engages the readers' curiosity or excitement, the action builds momentum to a cliffhanger at the second plot point, and the climax and the resolution, or denouement, relieve the tension. The IMRaD structure of a manuscript maps onto these acts.
In a manuscript the setup is the introduction, where readers learn about the characters and setting of the study: the characters might be proteins, drug compounds, diseases or patients, and the setting a particular cell type, organ, hospital or population. The setup also gives the background readers need, such as statistics on a condition, why existing drugs are ineffective or how the structure of a mutated protein influences a disease, and these details frame the purpose of the study, the problem it aims to address, which in turn helps frame the tension. The tension is the research question at the end of the introduction. It describes what the characters need to accomplish to help solve the problem framed in the setup, and it states the specific question the study proposes to answer and how that question connects to the problem.
The action is the methods and the results. The methods start it by describing the approach, experiments, data and analysis used to address the question. The results build on that momentum, their findings building on each other until readers know them but may not yet know whether they answered the question, a cliffhanger that carries the readers into the discussion.
The climax is the first paragraph of the discussion, which says whether the characters answered the question raised in the tension. To maximise its impact, the paragraph should remind readers of the study question, summarise the key findings and state how those findings answered the question. The rest of the discussion is the resolution, which explains how the findings fit within the existing literature, whether they support or contradict findings published in other studies, and how they advance the field. The conclusions paragraph at its end is the denouement, which ties the study together by reminding the reader of the problem, reiterating the question, highlighting the main findings, stating how they answer the question and emphasising how the study advances the field.
The five stages, setup, tension, action, climax and resolution, are abbreviated STACR. Engaging stories tend to follow this structure, although variations abound. IMRaD follows the STACR flow fairly naturally, so an original research manuscript has a built-in story structure that engages and informs readers. Stories make writing stronger and its message more engaging, and writing a manuscript using STACR helps readers understand the importance of the work and use what they have read.
Each stage can be built by answering its own questions:
- Setup. Who are the characters in this story (proteins, compounds, a disease state, a device)? Where does the story take place (proteins within a particular cell, a disease that affects a certain organ, patients in a hospital or outpatient location)? What problems surround the characters (the prevalence of the disease, earlier devices that proved insufficient, no drug available)?
- Tension. What do the characters need to accomplish, and how can they solve the problem set up (disease progression needs to be slowed; a gene mutation may change protein function)? What question does this story answer (does this drug have fewer side effects than others available, does this device work better)? How does the tension connect with the problem from the setup? Making the connection helps engage the readers' curiosity.
- Action. What happens to address the questions raised in the tension (what methods were used)? How did the characters respond to a change in the setting (what were the results)?
- Climax. Did the characters answer the question (what were the key findings, and how did they answer the questions posed)? How will the readers' curiosity be satisfied (by referring to elements raised in the tension)?
- Resolution. How have the characters and setting changed as a result of the action (how does this work advance the field)? What value does this story offer, and why does it matter (what is the significance of the study)?
The same structure can be applied to paragraphs and sentences. In a sentence the arc is compressed into a setup, a combined tension, action and climax, and a resolution: in a basic sentence the subject introduces a character as the setup, the verb carries the tension, action and climax, and the object resolves the curiosity about what happened to the character, as in Crystal loves dogs. A great paragraph is made of such micro-story arcs, arranged in an order that supports the overall story of the paper, which helps ensure that each paragraph has a coherent purpose within that story. A manuscript built on story structure at the level of the sentence, the paragraph and the whole can engage readers while informing them, and so have greater influence in the field.
Drafting from the reporting checklist
Reporting guidelines for different types of study are available on the internet to help a writer put the right information in each section, and they contain checklists of what information has to go in. A checklist, such as the one for an abstract, shows exactly what needs to be included, so the article can be built from it, and the checklist then gives the outline. A word is written against each box, as in parallel, cluster, double-blind or randomised against trial design, and where the clinical study document holds a well-formulated sentence, that sentence can be copied in. Read as questions, the items form a simple list to go through: whether the randomisation procedure, the ethics, the primary objective and any secondary objectives are included, with a key word or phrase written for each.
A checklist such as CONSORT's or STROBE's serves twice: once to make sure the information is there, and again as a list of sentences to write, one for each item, such as the approval of the ethics committee. With the abridged CONSORT abstract checklist, the first pass, checking that the information needed is available, is part of planning. The guidelines are also useful to writers who are not native speakers of English when they start writing articles, because a checklist shows what has to go in.
Detail against the word limit
An original research manuscript for a medical journal is commonly subject to a word limit of 3,500 to 4,500 words, and at the same time it must contain enough detail for other researchers to verify the research, so it has to be both comprehensive and tightly written. An IMRaD template for journal manuscripts can simplify the writing.
A journal's word limit often makes it impossible to include all the CONSORT items in the abstract or the whole manuscript, and for medical writers this is a daily problem. With a limit of 2,500 words, for instance, reporting everything about a randomised trial is practically impossible. Dealing with the problem starts from knowing the journal's instructions for authors.
The details can go into appendices, which practically any journal allows to be uploaded in addition to the article, its figures and its tables, and this is happening especially with open-access journals. The word limit itself is becoming less common as journals, open-access and more modern ones especially, move online, because editors printing a journal on paper cared about the space each article took, and on a website that space matters far less.
A study that is very difficult to describe within the limit probably has a problem in its design, which is too complex, or its authors did not follow the SPIRIT guidelines when writing the protocol.
An article that cannot be brought within the limit can be split in two, reporting the efficacy results in one article and the safety results in the other, a solution often seen. Looking for another journal is the last option, for an article that cannot stay within the limit because the journal does not allow additional material such as appendices, because it cannot be shortened, or because the authors do not agree to shorten it. The authors are the writer's clients, and what matters is that the writer keeps all the possible options in mind, shares them with the authors and helps them take the best decision.
Form, function and how the research was planned
In Tom Lang's account, scientific journals have three primary functions, disseminating scientific information, legitimising scientific validity and archiving the important details of research, and because one form cannot necessarily meet all three, journal articles take different forms in length, level of detail and rhetorical features. He holds that the functions of an article and its parts should be clear and widely held, and that the forms of the article and its parts should fit those functions, but in 2009 there was no universally held definition of what the functions are. He sees mismatches between forms and functions in publication guidelines: many journals impose character limits on titles, word limits on abstracts and articles, number limits on references and item limits on reporting guidelines, and these limits and differences in form might compromise the intended functions of an article and its parts. In his view, variations in the forms and functions of scientific articles are not necessarily rational, and evidence-based medicine and the standards movement require certain forms and functions. Knowing how to apply manuscript standards can help writers provide additional value to clients.
The organisation of a research report can also vary with the type of research. Traditional, or preplanned, research runs parallel experiments that test hypothesised differences, and its prototype is the clinical trial. Variant, or sequentially planned, research is a series of experiments in which the outcome of one determines the design of the next, and its prototype is the characterisation of a receptor and its function. Both pose a question, design a study to answer it, generate data and draw conclusions from the data, but their reports are organised differently. A report of traditional research has an introduction with no findings, a methods section describing the study design, results for each measured outcome, and a discussion that states the overall conclusions and provides context. A report of variant research gives findings in its introduction, has no study design, presents a series of mini-reports, one for each experiment, with conclusions in subheadings, and ends by stating the overall conclusions and providing context.
In 2009 one experienced writer anticipated that, with the emergence of translational medicine, more reports would vary in organisation. Guidelines on organising reports of sequentially planned research were then absent: a comprehensive search of books, journals and websites for advice on reporting such research found only one book that gives some. A medical communicator who knows the organisational differences can better help the authors of sequentially planned research.