Academic Writing and Scientific Papers Guide

A fundamental guide from the Study Plan to standardize and elevate the quality of scientific production.

Academic writing is characterized by its formality, objectivity, and scientific rigor. Unlike a literary or journalistic text, the academic paper has a clear objective: to report a finding, test a hypothesis, or systematically review an area of knowledge in such a way that other researchers can reproduce or validate what was said.

✍️ Principles of Scientific Writing

Clarity and Precision: Avoid unnecessary jargon that does not belong to the scope of your niche. Every word must have a function in the argument.
Impersonality: The most accepted standard is to write in the 3rd person singular (e.g., “it was observed”, “it is concluded”), or in the 1st person plural (e.g., “we verified”, although slightly less formal depending on the field). Avoid the 1st person singular (“I thought”).
Cohesion and Coherence: Use appropriate logical connectives between paragraphs to indicate whether the next idea is of addition, contrast, or conclusion.
Foundation: Never state something as a scientific fact without presenting the methodological data that supports it or a reliable bibliographic citation. All formatting of citations and references must follow a strict institutional standard (see more in abnt-formatacao-artigos).
Equations and Formulas: All mathematical formulas laid out in a block must necessarily be accompanied by a numbered caption (e.g., <p class="math-caption">Eq. 1: Formula name.</p>) immediately below the equation, ensuring technical clarity and ease of reference throughout the text.

🏗️ Classic Structure: IMRaD

The vast majority of papers in pure and applied sciences follow the empirical IMRaD format:

Introduction: What is known about the subject? What is not known (the gap in the literature)? What is the hypothesis/objective of your work?
Methods: How was the research done? Describe the materials, sample, data collection tools, and statistical analyses so that the experiment is 100% reproducible.
Results: What was found? Present the raw data through texts, tables, and charts without interpreting them in an opinionated way.
and
Discussion: What do these findings mean? Was the hypothesis confirmed? How do your results compare with the findings of other authors (using the base literature)? What are the study’s limitations?

📝 Tips for the Writing Routine

Write the Abstract last: The abstract is the “trailer” of your paper. It should contain 1 or 2 sentences from the introduction, the core of the method, the main results, and the primary conclusion. Only write this when you have a complete view of the text.
Keep a literature log: Use tools like Mendeley or Zotero, or create notes here in your digital garden, documenting the papers you read.
Peer review: Criticism is part of science. Before submitting to the journal, ask a colleague to read the text mercilessly.

🧪 Validating the Architecture: Ablation Study

In empirical papers (especially in AI), the results section must isolate each proposed component’s contribution by removing them one at a time and measuring the performance drop. This convention has its own name and anatomy — see ablation-study.

If you are in graduate school or doing your final thesis at Unesp, remember to consult the specific rules for delivering full papers: padrao-unesp-teses.