Academic publishing has its own format ecosystem, shaped by decades of tradition, journal requirements, and discipline-specific conventions. A computer science paper is written in LaTeX, compiled to PDF, and submitted to arXiv as TeX source. A psychology paper is written in Word using an APA template, submitted as DOCX, and typeset by the publisher's production team. A law review article might arrive as RTF with footnotes in Bluebook citation format.
Navigating this landscape requires knowing not just what each format can do, but what each journal, conference, and preprint server expects. This guide covers the major academic formats, submission workflows, reference management, and practical conversion strategies for researchers who work across disciplines.
The format you write in affects your collaboration workflow, equation handling, citation management, and submission options. Choose wrong and you'll spend hours reformatting before submission. Choose right and the format disappears into the background while you focus on the research.
LaTeX: The Gold Standard for STEM
LaTeX (pronounced "lah-tek" or "lay-tek") is a typesetting system, not a word processor. You write source code that describes the document's content and structure, and a compiler (pdflatex, xelatex, or lualatex) produces a PDF with professional typography. LaTeX handles equations, cross-references, bibliographies, figures, tables, and complex formatting through a markup syntax.
LaTeX dominates in mathematics, physics, computer science, engineering, and statistics. The reason is equations: LaTeX produces publication-quality mathematical notation that Word's equation editor can't match. $\int_0^\infty e^{-x^2} dx = \frac{\sqrt{\pi}}{2}$ compiles to perfectly typeset mathematics with correct spacing, symbol sizing, and alignment.
Beyond equations, LaTeX excels at automated cross-references ("see Section 3.2" stays correct when sections move), bibliography management (BibTeX/BibLaTeX handle citation formatting automatically), consistent numbering (figures, tables, equations all auto-numbered), and reproducible builds (the same source always produces the same PDF).
The tradeoff: LaTeX has a steep learning curve. The markup syntax is more complex than Markdown. Error messages are cryptic. Template modification requires understanding the macro system. Collaboration requires that all co-authors know LaTeX (or use Overleaf, which provides a visual editor). For text-heavy documents without equations, the overhead of LaTeX may not be worth it.
LaTeX Editors and Overleaf
Overleaf is the dominant online LaTeX editor. It provides a browser-based editor with real-time compilation, collaboration (Google Docs-style shared editing), template library, and submission integration with many journals. Roughly 50% of LaTeX papers now originate on Overleaf. The free tier covers most individual use; paid plans add more collaborators and build time.
Desktop alternatives: TeXShop (macOS), TeXstudio (cross-platform), VS Code with LaTeX Workshop extension (cross-platform, increasingly popular). All produce identical output because they all invoke the same TeX engine.
Word Templates: The Publisher's Format
Many journals in the humanities, social sciences, medical sciences, and some engineering fields accept or require DOCX submissions. Publishers provide Word templates with pre-configured styles (title, abstract, body text, headings, references) that match the journal's layout.
The workflow: download the journal's template, write your paper using the provided styles (not direct formatting — never direct formatting), manage references with Zotero or Mendeley, and submit the DOCX. The publisher's production team handles final typesetting.
Word's advantages for academic writing: tracked changes and comments for revision rounds (supervisors and reviewers use Word's review features heavily), real-time collaboration via Word Online or SharePoint, lower learning curve than LaTeX, and broader compatibility with co-authors.
Word's disadvantages: equation typesetting is adequate but not publication-quality (Word's equation editor improved significantly in 2018+ versions, but still doesn't match LaTeX for complex notation), cross-references are fragile (they break when content moves), bibliography plugins (Zotero, Mendeley, EndNote) occasionally corrupt documents, and large documents (100+ pages with many figures) become slow and unstable.
PDF Submission Requirements
Almost every journal and conference ultimately wants a PDF for review. Even DOCX submissions are typically converted to PDF by the submission system before reaching reviewers. The question is whether you submit PDF directly or let the system convert.
Conferences (CS, physics): Most require PDF compiled from LaTeX using the conference's style file (.cls or .sty). The style file controls margins, fonts, column layout, and heading formatting. IEEE, ACM, Springer, and Elsevier all provide their own style files. Using the wrong one is grounds for desk rejection.
Journals (broad): Requirements vary. Some want PDF only. Some want DOCX only. Some want TeX source. Some accept any format for initial submission and require a specific format only after acceptance. Always check the "Instructions for Authors" page — format requirements are non-negotiable.
Thesis/dissertation: Universities have specific formatting requirements (margins, fonts, line spacing, title page layout) and typically require PDF/A for their institutional repository. Check your graduate school's formatting guide before you start writing — reformatting a 200-page dissertation after the fact is painful.
Reference Management: BibTeX, Zotero, and Citation Formats
BibTeX is the reference format for LaTeX. References are stored in a .bib file as structured entries with fields (author, title, year, journal, DOI, etc.). LaTeX's \cite{} command inserts citations, and BibTeX/BibLaTeX formats them according to the specified citation style (APA, IEEE, Chicago, etc.). The .bib file is plain text and version-controllable.
Zotero is the most popular free reference manager. It captures references from web pages (one-click save from library databases), stores PDFs and notes, syncs across devices, and integrates with both Word (via plugin) and LaTeX (via Better BibTeX export). Zotero outputs citations in any style via CSL (Citation Style Language).
Mendeley (Elsevier-owned) provides similar features with tighter Elsevier ecosystem integration. EndNote (Clarivate-owned) is the commercial standard in some fields, particularly medical sciences.
The practical workflow: use Zotero to collect and organize references. For LaTeX papers, export to BibTeX via Better BibTeX (auto-updating export). For Word papers, use Zotero's Word plugin to insert citations directly. For switching between formats, Zotero's export handles the conversion between BibTeX, RIS, and CSL JSON.
Preprint Servers: arXiv, bioRxiv, and Others
arXiv (arxiv.org) dominates in physics, mathematics, computer science, and quantitative fields. arXiv accepts LaTeX source (preferred — they compile it on their servers for consistent rendering), PDF, and Word/RTF (converted to PDF). Submitting LaTeX source has advantages: arXiv can generate HTML versions, the source is available for reuse, and the compiled PDF uses arXiv's standard rendering.
bioRxiv and medRxiv serve biology and medical sciences. They accept PDF and Word. No LaTeX compilation — you submit a rendered document.
SSRN covers social sciences, economics, and law. Accepts PDF and Word.
For arXiv specifically: submit TeX source with all figures, style files, and a compiled PDF. arXiv's auto-compilation uses TeX Live, which may differ from your local installation. Common submission errors: missing fonts, missing style files, incompatible package versions. Test with arXiv's exact TeX Live version (they publish which year's distribution they use) before submitting.
Converting Between Academic Formats
Academic format conversion is common — you might write in LaTeX but a co-author needs Word, or you wrote in Word but a journal wants LaTeX.
LaTeX to PDF: Direct compilation via pdflatex/xelatex/lualatex. This is the native output path and produces the highest quality result.
LaTeX to DOCX: Pandoc handles this: pandoc paper.tex -o paper.docx --bibliography=refs.bib. Equations convert to Word equation objects. Tables and figures map to Word equivalents. Quality is good but not perfect — complex LaTeX packages may not convert. TeX to DOCX on ChangeThisFile provides this conversion.
LaTeX to Markdown (/tex-to-md): Useful for converting papers to blog posts or documentation. Equations can be preserved as LaTeX notation (most Markdown renderers support LaTeX math) or converted to images.
DOCX to LaTeX: This direction is harder. Word's formatting model doesn't map cleanly to LaTeX. Pandoc handles it (pandoc paper.docx -o paper.tex), but the output needs manual cleanup for anything beyond basic text.
DOCX to PDF (/docx-to-pdf): Reliable via Word or LibreOffice. This is the standard path for Word-based papers. Embed fonts before converting.
Markdown to LaTeX (/md-to-tex): Pandoc's core use case. Write in Markdown with LaTeX math notation, convert to LaTeX for compilation, or directly to PDF via Pandoc's LaTeX backend.
Thesis and Dissertation Formats
University thesis requirements are notoriously specific: exact margins (typically 1" or 1.5" on left for binding), specific fonts (often Times New Roman 12pt), double-spacing, particular title page layouts, specific pagination (Roman numerals for front matter, Arabic for body), and PDF/A for the institutional repository.
In STEM fields, most theses are written in LaTeX using the university's thesis class file. These class files implement all formatting requirements automatically — set the options and write your content. The LaTeX approach is reliable for long documents (200+ pages) with many figures, equations, and cross-references.
In humanities and social sciences, Word with the university's thesis template is more common. Word handles long documents but requires more manual formatting attention. Use styles religiously (heading styles for chapters and sections, Normal for body text), insert a table of contents from heading styles, and use Word's cross-reference feature for figure and table references.
Regardless of format: check your university's thesis formatting guide before writing, not after. Reformatting a completed thesis to meet specific margin, font, and pagination requirements is tedious and error-prone.
The format landscape in academic publishing is fragmented because different disciplines optimized for different needs. Physics needs equations and preprints — hence LaTeX and arXiv. Psychology needs tracked revisions and APA formatting — hence Word and journal templates. The friction comes when you cross boundaries: a computer scientist co-authoring with a sociologist, or submitting a LaTeX paper to a journal that wants Word.
The practical strategy: write in whatever format your field expects, manage references in Zotero (which bridges all formats), and use Pandoc when you need to convert. Don't fight the conventions of your field — they exist for reasons, even if those reasons are historical rather than technical.