In short: Writing a scientific manuscript efficiently starts before writing. It starts with checking whether your figures tell a story, choosing a narrative and outlining key points. A practical writing order is Methods, Results, Discussion, Introduction, References, Abstract and finally Title, followed by several rounds of rereading and mock review. Chaperone delivers manuscript writing and publication workshops and courses for universities, research institutes and funders, taught by trainers who have led research groups and published extensively.
Ask researchers what the hardest part of publishing is and few will say the experiments. They will say the writing.
The blank page. The supervisor who takes six weeks to return comments. The reviewer who says the conclusion is not supported by the data. English as a second language. The fourth rewrite of the same paragraph at midnight.
Communicating results is one of the most important parts of research, and the European Commission’s research competence framework lists it among the core competences of every researcher. Yet most scientists learn to write papers by trial and error.
This guide distils what we teach in our manuscript writing and publication workshops, developed and delivered by trainers who built their own careers from PhD to group leader before becoming career advisors.
What makes a “good” manuscript?
A good manuscript is one a reader can follow without effort. It asks a clear and relevant question, tells a coherent story with convincing data, and ends with a take home message that is fully supported by the evidence.
We open every workshop by asking participants to write down what defines a good manuscript from their perspective as readers. The answers are always remarkably similar: clarity, logical flow, honest data, a clear message. Everyone knows what good looks like. The challenge is producing it.
Two principles frame everything else:
- 1) Prepare every manuscript carefully, whatever the impact factor of the target journal. You and your work set the quality standard, not the journal.
- 2) Every author is responsible for the paper, whatever their position on the author list. If you have concerns, discuss them with your co authors. Ultimately, you can ask to remove your name.
Step 1: Do you have enough data for a paper?
Before writing a single sentence, analyse your data and lay out your figures.
- Do the figures tell a story? Draw a scheme of your ideas. This often exposes a hole in the data.
- Is the data convincing? Are sample sizes meaningful? Do results reach statistical significance?
- Discuss the story with your supervisor before investing weeks in prose.
A very practical exercise: print all your figures, lay them on a desk or the floor, and arrange them in the order that best tells the story. Rearrange until you find the simplest, most objective path to the message. Then present it to colleagues. Restart if needed.
This is more powerful than it sounds. The same set of figures can support very different papers.
Step 2: Choose your narrative
Every good paper has a narrative arc. Common ones in the life and health sciences include:
- Problem to solution. Identify a problem, propose an approach, demonstrate effectiveness, discuss implications. Typical for new diagnostic tests, drugs or public health interventions.
- Comparison. Baseline practice, new approach, direct comparison of performance or cost, conclusion. For example, rapid tests versus PCR.
- Mechanistic insight. Observation, hypothesis, evidence, mechanistic model. Typical for fundamental biology.
- Observation to explanation. An unexpected pattern, possible explanations, the most plausible interpretation.
- Characterisation to prediction. Describe a system, identify trends and drivers, forecast future behaviour.
In our workshop, participants work with one set of seven figures about a fictional pathogen and a new mutation. One group builds an observation to explanation paper: cases are rising, testing also increased, but a new variant emerged and spread, so genetic surveillance could provide early warning. Another group builds a mechanistic paper from the same data: the mutation enhances cell entry, which explains its dominance and points to a drug target.
Notice what happens. In the first story, the figures on cell entry and drug susceptibility disappear. In the second, the testing figure disappears. Choosing a narrative is also choosing what to leave out.
Step 3: Write a title that evolves with the paper
Write a tentative title early, then assess it on a scale from generic to revealing. Titles typically mature in three stages:
- Project stage: the title describes the topic. “Surveillance of Pathogen X in Country Y.”
- Draft stage: the title suggests a question or pattern. “Emergence and Geographic Spread of Mutation M in Pathogen X.”
- Writing stage: the title communicates the main finding. “Mutation M Became the Dominant Variant During Rapid Expansion of Pathogen X.”
Aim for the last one. Check titles in your target journal to calibrate how catchy or how specific yours should be. Above all, be informative.
The structure of a scientific manuscript
Picture the classic manuscript as an hourglass. The Introduction starts broad and narrows down to your specific aims. Methods and Results present your work. The Discussion starts with your key findings and widens back out to the big picture.
How to write the Introduction
The Introduction is an inverted pyramid leading the reader to your question. A reliable structure has three moves:
- What is known: the context, relevance and the scientific question.
- What is unknown: the gap, the unclear or poorly defined issues.
- Why the study was done: your rationale, hypothesis and objectives.
Three tips we insist on:
- One paragraph, one idea. The central idea belongs in the first sentence. A reader should understand the whole argument by reading only the first sentence of each paragraph. Try this test on your own draft.
- Be critical. You are discussing other people’s work, but it is still your manuscript.
- An Introduction is not a mini review. As Wenzel and colleagues put it, if the introduction does not instil enthusiasm for your study, a journal is unlikely to consider it.
How to write the Methods
The Methods section exists so that a competent researcher could repeat your work. Include exact specifications, quantities, sources, equipment and statistical methods. Group related methods together rather than listing them chronologically, and use tables when many components are involved. Answer every “how” and “how much” so the reader never has to guess.
A useful tip: find an authors’ checklist from a recent paper published by your lab or your target journal, and use it from day one.
How to write the Results
The Results section is the organised presentation of your data. The key word is organised.
- Start with the main finding. Do not make the reader search for it.
- One paragraph, one message. Each result should build on the previous one.
- Figures show the data. Text explains why it matters.
- Present results in logical order, not the order in which you ran the experiments. A manuscript is not your lab book.
- Use subheadings to reinforce the narrative and save interpretation for the Discussion.
- Every figure needs a purpose. If it does not advance the story, remove it.
How to write the Discussion
Open with a brief summary of the main findings and their possible explanation. Then state what is novel and why it matters. Address strengths and limitations honestly, and close with a clear, concise take home message.
Avoid three common pitfalls: repeating the Results, reviewing the literature beyond what is needed for context, and claiming to be first. That last one, as experienced editors will tell you, mostly invites angry letters from people who believe they were.
References
Assume your reviewers are the authors of the papers you cite. Never cite a paper you only read the abstract of, because reviewers can tell when their work has been misinterpreted.
Use a reference manager such as Zotero, Mendeley, EndNote or Paperpile. Build your library as you read, cite while you write, and let the software format everything at the end.
Abstract and title
Write the abstract last. Check the journal’s guide for authors, give one sentence of background, then the key methods and results. Many readers will only ever see your abstract, so it must be useful on its own.
In what order should you write a manuscript?
Most researchers write in the order the paper is read. That is rarely the most efficient approach. The order we recommend:
- Methods
- Results
- Discussion and conclusion
- Introduction
- References
- Abstract
- Title
- Read and rewrite, as many times as needed
Starting with Methods builds momentum, because it requires little creative effort. Writing the Introduction after the Discussion ensures it sets up exactly the story you ended up telling.
Before submitting, recruit mock reviewers: yourself (reading aloud after some days away), your lab colleagues and PIs close to your field.
Common challenges in manuscript writing and how to overcome them
Writing in English as a second language. Use writing assistance tools, use AI wisely as a support rather than an author, and consider a dedicated scientific writing course.
Writer’s block. Start with something that requires little thinking, such as the Methods or a figure legend. Momentum beats inspiration.
Time management. Reserve protected writing time and divide the big cake into small slices, then into tiny bites. The Pomodoro technique works well: 25 minutes of uninterrupted work on one bite sized task, a 5 minute break, and a longer break every four rounds.
The PPP cycle: perfectionism, procrastination, paralysis. Many researchers recognise themselves here. Learn to spot the moment you enter the cycle, lower the bar for a first draft, and use small tasks to get moving again.
The supervisor bottleneck. Book a meeting dedicated only to the manuscript to agree on ideas, division of labour and a feedback schedule. Write your version even if your supervisor plans to rewrite it. Remember that your paper is your top priority but one of many for them, and kindly remind them of the agreed schedule.
Defining the author list. There are no universal rules. Ask about lab and institutional practices before the research starts, not at submission. The CRediT taxonomy (Contributor Roles Taxonomy) is a useful way to describe each author’s contribution.
Getting accepted. Papers are commonly rejected because the question is not interesting, already answered or obvious; the hypothesis or methodology is flawed; the study is underpowered; or the conclusion is not supported by the data. Almost all of these can be caught at the figure layout stage of Step 1.
Why institutions invest in manuscript writing training
Publications remain central to how researchers, departments and institutions are evaluated. Yet scientific writing is rarely taught systematically. Most researchers learn by imitation and painful revision cycles.
Structured training shortens that curve. It helps PhD students publish earlier, reduces the supervision load on group leaders, improves the quality of submissions and lowers the stress associated with writing. For funders and doctoral programmes, it is a direct investment in research output.
How Chaperone’s manuscript writing training works
Our manuscript writing and publication training is available as a workshop or a longer course, online or in person, and adapted to the field and career stage of participants. It is highly practical. Researchers draft and assess titles, test the first sentence rule on real papers, organise figures into competing storylines, write take home messages and work on their own challenges.
We have delivered this training for institutions such as the Gulbenkian Foundation, and as part of long courses for more than 100 researchers based in Africa, alongside grant writing, project management and data management. Researchers who want individual feedback on a manuscript or thesis can also book 1:1 sessions with our consultants.
Key takeaways
- Start with figures and a story, not with text.
- Choose a narrative. It decides what stays in and what goes.
- Write Methods first and the Abstract and Title last.
- One paragraph, one idea, with the main point in the first sentence.
- Use mock reviewers before the real ones.
Frequently asked questions
What is the best order to write a scientific paper?
A practical order is Methods, Results, Discussion, Introduction, References, Abstract and Title, followed by several rounds of rereading. Starting with Methods builds momentum, and writing the Introduction late ensures it matches the final story.
How do I know if I have enough data for a paper?
Lay out your figures and check whether they tell a coherent story that answers a clear question with adequate sample sizes and statistics. Drawing the storyline often reveals gaps. Discuss it with your supervisor before writing.
How can I overcome writer’s block when writing a manuscript?
Start with a section that needs little creative effort, such as the Methods. Break writing into small tasks, use the Pomodoro technique and lower the bar for the first draft.
Does Chaperone offer manuscript writing workshops for research institutions?
Yes. Chaperone delivers manuscript writing and publication workshops and courses for universities, research institutes, doctoral programmes and funders, online or in person, led by trainers with extensive research and publishing experience.
About Chaperone
Chaperone is a social impact company democratising access to high quality career development in science. With over 50 international career consultants and trainers, Chaperone provides group training, 1:1 career sessions through the first online marketplace of its kind for scientists, institutional consultancy and scientific talent recruitment. Chaperone has empowered over 5,000 scientists from more than 40 countries and is trusted by institutions such as EMBL, Karolinska Institutet, the European Commission Joint Research Centre and the Gulbenkian Foundation.
Want to help your researchers publish with more confidence? Email hello@chaperone.online and we will design a training for your community.