Research Paper Writing Guide: Format, Structure, and AI-Assisted Review

Journal reviewers spend 2-4 weeks evaluating your research paper. They check structure, methodology, argument quality, citation accuracy, and whether the paper meets the journal's specific requirements. If they find problems, you get a "revise and resubmit" — or a rejection.

Most of those problems are structural, not scientific. Missing sections, weak abstracts, methodology described without justification, results mixed with discussion, conclusions that introduce new material. These are fixable issues that could have been caught before submission.

This guide covers research paper structure, formatting conventions, the most common reasons papers get rejected, and how to build a review process that catches those issues before a journal reviewer does.

The IMRaD Structure

Most empirical research papers follow the IMRaD structure: Introduction, Methods, Results, and Discussion. This isn't arbitrary — it reflects the logic of scientific inquiry and gives readers a predictable structure for finding information.

Some disciplines add sections: a separate Literature Review (common in social sciences), a Theoretical Framework section, or a Limitations section. Some journals require a Highlights section, a Graphical Abstract, or Author Contributions. Always check the target journal's author guidelines.

Section-by-Section Guide

Title

The title is the first thing editors, reviewers, and readers see. It determines whether anyone reads further.

Effective titles are:

• Specific — "The Effect of Mindfulness-Based Stress Reduction on Cortisol Levels in Healthcare Workers: A Randomised Controlled Trial" (not "Mindfulness and Stress")
• Informative — convey the topic, scope, and often the method or finding
• Keyword-rich — include terms that researchers in your field would search for
• Concise — typically 10-15 words, though this varies by discipline

Common title mistakes:

• Too broad ("A Study of Climate Change")
• Too clever (puns and wordplay that obscure the content)
• Too long (exceeding 20 words without justification)
• Missing key terms that enable discovery

Abstract

The abstract is a self-contained summary of the entire paper. Many readers (and all database searches) rely on the abstract to decide whether to read the full paper. Journal editors often make initial screening decisions based on the abstract alone.

Structure (for structured abstracts):

Common abstract mistakes:

• Vague results. "Results were significant" — include actual findings, effect sizes, or key themes.
• No methods. Readers need to know HOW you reached your findings.
• Copying introduction text. The abstract summarises the whole paper, not just the background.
• Exceeding word limit. Most journals specify 150-300 words. Exceeding it signals carelessness.
• Including citations. Abstracts typically don't include references (check journal policy).

Introduction

The introduction builds the case for your research. It follows a "funnel" structure: broad context → specific gap → your research question/hypothesis.

The three moves of an introduction (Swales' CARS model):

1. Establishing the territory — what's the broad area and why does it matter?
2. Establishing the niche — what's missing, contradictory, or under-explored?
3. Occupying the niche — what does your paper do to address the gap?

Common introduction mistakes:

• Starting too broadly. Begin with your research area, not the history of the field.
• Gap not clearly stated. The reader should understand exactly what's missing by the end of the introduction.
• Research questions absent or vague. State specific, answerable questions or testable hypotheses.
• Too long. The introduction frames the research — it doesn't review all the literature (that's the lit review or background section).

Methods

The methods section describes what you did in sufficient detail for another researcher to replicate the study. Clarity and completeness are paramount.

Essential elements:

Common methods mistakes:

• Insufficient detail. "Interviews were conducted" — how many, how long, what type, what questions?
• No justification for choices. Why this method and not alternatives? Why this sample size?
• Missing ethics statement. Even minimal-risk studies need ethics acknowledgement.
• Wrong tense. Methods are written in past tense — you're describing what you did.

Results

The results section presents findings without interpretation. This discipline is critical: results tell the reader WHAT you found; discussion tells them WHAT IT MEANS.

For quantitative papers:

• Report statistical results with effect sizes, confidence intervals, and p-values
• Use tables and figures for complex data — but don't duplicate data across text, tables, and figures
• Report all results, including non-significant findings
• Organise by research question or hypothesis

For qualitative papers:

• Present themes with supporting evidence (participant quotes)
• Use enough quotes to demonstrate the theme but not so many that the section becomes a transcript
• Organise by theme, not by participant
• Include disconfirming evidence

Common results mistakes:

• Interpreting results. "This shows that participants preferred X because..." — save interpretation for discussion.
• Selective reporting. Report all findings, including those that don't support your hypothesis.
• Tables without narrative. Every table needs surrounding text that guides the reader.
• Redundant presentation. Don't present the same data in a table AND a figure AND the text.

Discussion

The discussion is where you demonstrate your scholarly contribution. You interpret findings, connect them to existing literature, and explain implications.

Discussion structure:

1. Summary of key findings — brief restatement of main results (1-2 paragraphs)
2. Interpretation and literature connection — what findings mean in context of existing research
3. Theoretical implications — what findings mean for theory
4. Practical implications — what findings mean for practice or policy
5. Limitations — honest assessment of study limitations
6. Future research — specific directions for future work

Common discussion mistakes:

• Repeating results. Briefly reference results but focus on interpretation, not re-presentation.
• Ignoring contradictory findings. If your results contradict the literature, engage with why.
• Overclaiming. Your study suggests or indicates — it doesn't prove. Use appropriate hedging.
• Generic limitations. "Small sample size" is a limitation, but explain WHY it limits your specific findings and HOW future research could address it.
• No new contribution. If your discussion only confirms what the literature already said, what's the point of the study?

Conclusion

The conclusion summarises contributions and implications. It should be concise and forward-looking.

Include:

• Direct answers to your research questions
• Key contribution to knowledge
• Most important practical implication
• One or two specific future research directions

Exclude:

• New arguments, data, or citations
• Lengthy repetition of results
• Vague statements ("more research is needed")

Citation Styles

Different disciplines use different citation styles. Using the wrong style signals unfamiliarity with the field.

Common citation mistakes:

• Mixing citation styles within a paper
• In-text citations that don't match the reference list
• Missing page numbers for direct quotes
• "Et al." used incorrectly (APA: use for 3+ authors from first citation)
• Reference list entries with inconsistent formatting

Common Rejection Reasons

Understanding why papers get rejected helps you prevent those issues:

Studies of editorial decisions consistently show that methodology and writing quality are the most controllable factors. You can't always control whether your contribution is novel enough for a top journal, but you can control whether your methods are rigorous and your writing is clear.

Building a Pre-Submission Review Process

Before submitting to a journal, review your paper systematically:

Review Checklist by Section

Using AI Review for Pre-Submission

Configure a reviewer with journal-specific criteria:

Upload the journal's author guidelines as a knowledge base so the reviewer checks against journal-specific requirements.

The Two-Round Review

Round 1: Structure and content

• Does the paper follow IMRaD (or the required structure)?
• Are all required sections present and complete?
• Is the methodology justified?
• Are results separated from discussion?
• Is the contribution clearly stated?

Round 2: Writing quality and formatting

• Is the academic register consistent?
• Are claims appropriately hedged?
• Are citations formatted correctly?
• Is the abstract within word limits?
• Are tables and figures numbered and referenced correctly?

Frequently Asked Questions

How long should a research paper be?

Most journal articles are 5,000-8,000 words (excluding references). Some journals accept shorter communications (2,000-3,000 words) or longer review articles (8,000-12,000 words). Always check the target journal's word limit — exceeding it often results in desk rejection.

Should I write the abstract first or last?

Last. The abstract summarises the entire paper, so you need to know what the paper contains. Some researchers write a draft abstract first to clarify their thinking, but the final abstract should be written (or substantially revised) after the paper is complete.

How do I choose the right journal?

Consider: (1) Scope — does the journal publish work in your area? (2) Audience — who reads this journal? (3) Impact — is this an appropriate level for your work? (4) Turnaround — how long is the review process? Read 3-5 recent papers from the journal to calibrate your expectations.

What's the difference between "revise and resubmit" and "reject"?

A "revise and resubmit" (R&R) is generally positive — the editor sees potential but wants changes. Address every reviewer comment, even if you disagree (explain why). A "reject" may include feedback that helps you improve for another journal. Many eventually-published papers were rejected from their first-choice journal.

Can I submit to multiple journals simultaneously?

No. Simultaneous submission violates the policies of virtually all academic journals. Submit to one journal at a time. If rejected, you can submit elsewhere.

How many references should I include?

There's no universal number. Empirical papers typically cite 30-60 sources; review papers may cite 100+. Cite what's relevant — every reference should serve a purpose. Over-citing to appear thorough is as problematic as under-citing.

Should I use AI to write my research paper?

That's between you and your institution's academic integrity policy. What's universally acceptable is using AI to REVIEW your paper — checking structure, clarity, citation consistency, and rubric alignment. The writing should be yours; the quality review can be assisted.

Key Takeaways

• IMRaD is the standard structure for empirical research papers: Introduction, Methods, Results, and Discussion. Check your journal's requirements for variations.
• The abstract is your paper's advertisement — include all five elements (background, aim, methods, results, conclusion) within the word limit.
• Methods need justification, not just description — explain why you chose this approach over alternatives.
• Separate results from discussion — present findings objectively, then interpret them in a separate section.
• Most rejections are preventable — poor writing quality, weak methodology, and incomplete sections are controllable factors.
• Use the target journal's author guidelines as a review checklist. Upload them as a knowledge base for journal-specific review.
• Review in two rounds: structure and content first, then writing quality and formatting.
• Hedge your claims — your study suggests, indicates, or provides evidence for. It doesn't prove.

---

This article is for informational purposes. Publication requirements vary by journal, discipline, and institution. Always follow your target journal's author guidelines and your institution's research integrity policies.