Write a scientific paper is a long-term job. Whether for a young doctoral student or for a confirmed researcher, writing a scientific paper requires rigor. A scientific paper is often constructed according to a recurring pattern, the secrets of which are as follows.
An article starts with the title, abstract and keywords.
The text of the article follows the IMRAD format (IMRED in French), which answers the questions below:
- Introduction : What did you do/what others did? Why did you do it?
- Methods: How did you do it?
- Results: what did you find?
- And
- Discussion : What does all this mean?
The main text is followed by the conclusion, acknowledgements, references and supporting documents. The writing of a paper is not done in the same order as the plan. Writing is like a puzzle for which each researcher has his own resolution preferences.
A scientific paper is constructed according to the following plan:
- Title
- Authors and Affiliations
- Abstract
- Keywords
- Introduction
- Background
- Related Works
- Problem definition
- Materials and Methods
- methodology
- Case Study
- Experiment
- Results
- Discussion
- Findings
- References
Sections in bold are almost always present while others may not appear in a scientific paper.
Here are the main guidelines for each section.
Title and header
The title must:
- Identify the subject of the paper
- Reduce scope of work
- Don't be too wide or too narrow (stay eye-catching)
The title is essential as it informs the readers about the issue being discussed and the benefits they will get from the article.
You can include a subtitle after a separator such as a colon.
Examples:
- MEgo2Vec: Embedding Matched Ego Networks for User Alignment Across Social Networks
- SSDMV: Semi-Supervised Deep Social Spammer Detection by Multi-view Data Fusion
- Formal vs. Case-Study-Based Approaches for the Identification of Cultural Influences in Requirements Engineering
- Big data analytics in supply chain management: A state-of-the-art literature review
- Systematic Literature Reviews in Software Engineering: Preliminary Results from Interviews with Researchers
For the header, its writing depends on the format of the article or the newspaper. You must refer to the LaTeX template to know the rules.
Abstract (see the dedicated course for more details)
Write the summary after the content of the article!
The abstract has a word limit (usually between 100 and 250 words). It's a single paragraph with no quotes or abbreviations. It is content with itself (no need to read the introduction or another article to understand).
The summary should state:
- What is the goal/problem? (in yellow)
- Why is it important (social/industrial/scientific, etc.)? (in blue)
- How do you handle the problem? (in green)
- Your main results
How will this contribute/benefit? (in red)
Keep in mind that the summary is a kind of summary of the article, a kind of preface.
Keywords
Follow the template instructions or conference/journal listings.
Keywords are used for indexing and searching the database, so choose them carefully. Use only one keyword or scientific field, be broad and narrow (Social Network, Tourism, Clustering, distance measurement, profiling).
Your article ranks high on the list of any keyword searches that potential readers perform in databases.
Introduction (see the dedicated course for more details)
The introduction provides:
- What is the article about and why should the reader care
- The context of the research (subject, motivation, scope)
- An overview of the current state of the art
A good introduction should answer the following questions:
- What is the problem to be solved?
- Are there existing solutions?
- What is the best ?
- What is its main limitation?
- What do you hope to achieve?
The introduction provides the "what?" " so what? ". Be clear and stay objective.
Follow the following structure (each point can be a few paragraphs).
- Define the problem
- Motivation for the article (context) – first two points can be reversed –
- How will the answer to your problem contribute to the context?
- A bit of context to understand the issues
- Short link to related works (to better identify the gap in the next point)
- An account of what has not been properly addressed by previous researchers in this context (what is the gap?)
- How will the answer to your problem contribute to the area of study?
- The solution in brief and what is new and interesting in the search (list of contributions)
- Some technical details of the proposed solution
- Paper Preview
TL;DR: definition of the concept -> description of the context -> some justifications -> the diagram
DON'T: provide too much/unnecessary information
Literature review (see the dedicated course for more details)
First, write a paragraph with the most important related works (survey, books) relevant to your field/research context.
Organize the paragraphs thanks to the mindmap of your literature review ! Use \subsection* or \paragraph with a subtitle in a different font/shape to make it clear!
Talk about related work in a specific area (a node/leaf of your mental map) with reference. Avoid plagiarism. Discuss the related word, how it works in your context, and the gap in your context.
At the end, summarize the gaps and explain how you avoid/solve them.
Materials & Methods (see the dedicated course for more details)
When writing a lab report, it is often a good idea to start by writing the Materials and Methods section.
This section is usually very straightforward, and writing it first helps many people establish the thought process and understanding of the work that will allow the rest of the report to flow more smoothly.
This section gives a detailed account of the procedure that was followed to carry out the experiment or experiments discussed in the report.
Such a narrative is very important, not only for the reader to have a clear understanding of the experiment, but a well-written Materials and Methods section also serves as a set of instructions for anyone wishing to replicate the study in the future.
Given the importance of “reproducible results” in science, it is quite obvious why this second application is so vital.
Example :
The materials present the data, its context and its particularities
Methods presents a detailed flowchart of the methodology research with paragraphs briefly describing the overall process.
Methodology (see the dedicated course for more details)
The purpose is to describe how your study was conducted to prove the hypotheses or answer the research question.
Describes the experiments in detail (with short examples and figures).
You must also justify the experimental design.
You should critique your arguments because reviewers pay close attention to methods. It needs to be replicable, sometimes need a Git to prove your points.
Add references to previous work by your teams or others.
The explanation greatly depends on the search domain and search type – please refer to the parent link at the top of the page.
Results, Validation and Discussion (see the dedicated course for more details)
We will mainly focus on the discussion. The presentation of the results requires a lot of subtlety and I invite you to go and see the dedicated course.
Consider the following tips:
- Avoid statements that go beyond what the results can support.
- Avoid the sudden introduction of new terms or ideas; you have to present everything in the introduction, to be confronted with your results here.
- Speculation about possible interpretations is allowed, but these should be rooted in fact rather than imagination
- How do these results relate to the original question or objectives described in the Introduction section?
- Does the data support your hypothesis?
- Are your results consistent with what other researchers have reported?
- Discuss weaknesses and gaps. If your results were unexpected, try to explain why
- Is there another way to interpret your results?
- What further research would be needed to answer the questions raised by your findings?
- Explain what is new without exaggerating
Reviewing results and discussing is not just paperwork. You can do other experiments, derivations or simulations.
Sometimes you can't clarify your idea in words because some critical elements haven't been researched substantially.
The conclusion (see the dedicated course for more details)
The conclusion is usually short.
It summarizes what you have done (note! This is not a summary).
It recalls the significant contribution and discoveries.
The reader can read the title, quickly see the summary, the structure of the article, then the introduction and the conclusion. Give the key information in the conclusion.
You can add other works, but avoid developing new ideas.
Do not introduce new arguments. Do not plagiarize your results and your discussion.
EN 
FR 
RU 
ES