How to Write a Research Paper: Steps, Tips, Examples (2026 Guide)

Glossary of Key Terms

Step 1: Understand Your Assignment and Define Your Scope

The single most common mistake early researchers make is beginning to write before fully understanding what is expected of them.

What to check before you start

• Type of paper: Is this an argumentative essay, an empirical study, a literature review, or a report?
• Word limit: This determines how broad or narrow your topic can be
• Formatting requirements: Font, spacing, heading style, margin size
• Citation style: APA, MLA, Chicago, Vancouver, IEEE, or Harvard?
• Submission method: Institutional portal, journal submission system, or email?
• Deadline: Work backward and assign time for research, drafting, editing, and plagiarism checking

Scope by paper length

Step 2: Choose Your Topic

A good topic is the foundation of a good paper. It should be specific enough to allow deep coverage, broad enough to have sufficient source material, and ideally something that genuinely interests you.

Criteria for a strong research topic

• Original: Does it add something new to existing knowledge or debate?
• Feasible: Can you access the data, sources, or participants required?
• Relevant: Does it align with your field and assignment brief?
• Focused: Is it narrow enough to be covered within your word limit?

Discipline-specific examples of topic narrowing

Techniques for generating topic ideas

• Free writing: Write continuously for five minutes on a broad subject; patterns will emerge
• Scan reference lists: Papers you admire often contain suggestions for future research in their Discussion or Conclusion sections
• Check recent literature: Gaps in current knowledge are highly publishable
• Talk to your supervisor or librarian: Both are underused resources

Step 3: Conduct a Literature Review

A literature review is not simply a summary of what others have written. It is a critical evaluation of existing knowledge that establishes the context for your own contribution.

What a literature review should do

• Identify what is already known about your topic
• Highlight gaps, contradictions, or debates in the existing literature
• Justify why your paper is necessary
• Demonstrate that you have engaged with authoritative sources

Where to search for sources

Primary vs secondary sources

Annotated example (Medicine): Instead of simply citing a review article on antibiotic resistance, note what it contributes: Smith et al. (2022) systematically reviewed 47 studies and found that overuse of broad-spectrum antibiotics in outpatient settings correlated with a 34% increase in resistance rates: a finding directly relevant to this paper’s focus on primary care prescribing behaviour.

Efficient reading strategy

Do not read every source in full during the initial survey. Instead, speed up your search of the literature as follows:

1. Read the title and abstract to assess relevance
2. Skim the introduction and conclusion for key claims
3. Read methods and results in full only for the most relevant papers
4. Keep a reference manager (Zotero, Mendeley, or EndNote) updated as you go

Step 4: Develop a Thesis Statement

Your thesis statement is the spine of your paper. Everything you write should either support or develop it.

Characteristics of a strong thesis statement

• Specific: Not “social media is harmful” but “daily Instagram use exceeding 90 minutes is associated with increased anxiety scores in adolescent girls aged 13–17”
• Arguable: It makes a claim that requires evidence and can be challenged
• Answerable: It responds directly to your research question
• Concise: One or two sentences maximum

Formula

[Subject] + [specific claim or finding] + [reason or method]

Weak: This paper discusses the effects of climate change on agriculture.

Strong: Prolonged drought cycles resulting from climate change have reduced wheat yields in sub-Saharan Africa by an estimated 15–20% since 2000, disproportionately threatening food security in smallholder farming communities.

Discipline-specific thesis examples

Step 5: Build a Research Paper Outline

An outline is a planning tool, not a straitjacket. It saves enormous time during drafting and prevents structural problems that are costly to fix later.

Standard IMRaD structure

Non-IMRaD structure (humanities and social sciences)

• Introduction + thesis
• Thematic or chronological body sections
• Counterargument and rebuttal
• Conclusion

How to build your outline

1. List your main arguments or findings as Roman numerals (I, II, III…)
2. Under each, list supporting evidence or sub-points (A, B, C…)
3. Note the specific sources you plan to cite against each point
4. Flag any sections where you need more research before writing

Step 6: Write a Compelling Title

Your title is the most-read part of your paper. It determines whether researchers find it in database searches and whether they click through to read it.

Title writing principles

Title formats

• Declarative: States the finding directly: Mediterranean diet adherence reduces cardiovascular risk by 28%
• Descriptive: Describes the study: The effects of intermittent fasting on insulin sensitivity: a randomised trial
• Interrogative: Poses a question: Does early bilingual education improve long-term cognitive flexibility?

Step 7: Write the First Draft

Writing the introduction

A strong introduction answers three questions in sequence:

• What?: What is the topic and why does it matter?
• Why?: What gap in knowledge does this paper address?
• How?: What is the thesis and how will it be argued?

Annotated example (Computer Science): “Facial recognition systems are now embedded in border control, policing, and banking applications globally (What). However, studies have demonstrated error rates up to 34% higher for darker-skinned female faces compared to lighter-skinned male faces, a disparity that has received limited attention in production-system design (Why). This paper examines bias propagation in commercially deployed recognition models and argues that racially homogenous training datasets are the primary driver of differential error rates (How/Thesis).”

Writing the body

• Each paragraph should advance one argument or piece of evidence
• Open each paragraph with a topic sentence that connects to the thesis
• Support every claim with evidence: data, citations, examples
• Use transition sentences to link paragraphs: the final sentence of one paragraph should lead into the next
• Integrate tables, figures, and graphs to communicate quantitative findings; every visual must have a caption and be referenced in the text

Writing the conclusion

Step 8: Follow Journal and Formatting Guidelines

If you are submitting to a journal, treat its author guidelines as mandatory. Manuscripts are frequently desk-rejected not for poor science but for non-compliance with basic formatting requirements.

Pre-submission checklist

• Word count within the specified range
• Citation style matches journal requirements
• Figures and tables formatted to journal specifications
• Abstract length and structure matches the required format
• Cover letter prepared with key contributions stated
• Author affiliations, ORCIDs, and conflict-of-interest statements included
• Keywords selected from the journal’s approved list (where applicable)

Common citation styles at a glance

Step 9: Edit, Revise, and Proofread

Editing is a separate skill from writing. Allow at least 24 hours between finishing a draft and beginning to edit.

Two-pass editing strategy

Pass 1: Structural edit:

• Does the paper answer the research question?
• Is the argument logically sequenced?
• Are there paragraphs that overlap or contradict each other?
• Does the conclusion follow from the evidence in the body?

Pass 2: Line edit:

• Cut every word that is not carrying weight
• Replace passive voice with active voice where possible
• Check consistency: tense, capitalisation, hyphenation, abbreviation use
• Verify every in-text citation has a corresponding reference list entry

Before: “Brain injury incidence shows two peaks: the rates are the highest in the very young and the elderly.” After: “Brain injury incidence peaks in the very young and the elderly.”

Common language pitfalls

You can also check out our resources on the right tense to use in scientific writing, whether to use active or passive voice, and tips on sentence structure.

Step 10: Check for Plagiarism

Plagiarism (intentional or accidental) can result in a failed grade, failed course, rejection, retraction, and lasting reputational damage.

Types of plagiarism

Best practice

• Use a plagiarism checker (such as the free checker from Paperpal) before submission
• Keep meticulous notes during research, always recording source details
• When paraphrasing, close the source document, write from memory, then verify accuracy
• When in doubt, cite

Frequently Asked Questions

How long should a research paper be?

Length is determined by the assignment or journal. Typical undergraduate papers run 2,000–5,000 words; master’s dissertations 10,000–20,000 words; journal articles 4,000–8,000 words. Always check your specific guidelines first. The guiding principle: as long as the argument requires, no longer.

Can I write the introduction last?

Yes,many experienced researchers do exactly this. The introduction frames what follows, but you often cannot frame it precisely until you know what the paper actually says. Writing the body first and returning to the introduction is a legitimate and widely recommended strategy.

How many sources do I need?

There is no universal rule, but a rough benchmark: 8–15 sources for a 3,000-word undergraduate paper; 30–50+ for a master’s thesis; 50–100+ for a doctoral dissertation or standalone literature review submitted to a journal. Quality and relevance matter far more than quantity.

What is the difference between a research paper and a literature review?

A research paper presents original findings: data you collected, an argument you constructed, or an analysis you conducted. A literature review synthesises what other researchers have already published. Many research papers contain a literature review section, but the two are distinct document types.

Do I need an abstract for a course paper?

Check your assignment brief. Many undergraduate papers do not require an abstract, but journal submissions always do. If your paper is 3,000+ words and may reach an external audience, an abstract is good practice: it helps readers decide whether the paper is relevant to them.

My topic has been “done before”; is it still worth writing about?

Almost certainly yes. Replication with a different population, time period, methodology, or dataset is valuable scholarship. What matters is that your paper makes a distinct contribution. Confirming or updating an existing finding has real academic value; state explicitly in your introduction how your paper adds to or departs from prior work.

How do I know if a source is credible?

Apply the SIFT test: Stop before citing; Investigate the source (author credentials, publisher reputation, peer-review status); Find corroborating sources; Trace claims back to the original. Prefer peer-reviewed journals over news articles, government or institutional data over blogs, and recent publications for fast-moving fields.

Can I cite Wikipedia in a research paper?

Not as a primary citation, but as a starting point. Wikipedia articles often link to primary sources in their footnotes: follow those links and cite the original. Wikipedia is not peer-reviewed and edits continuously, making it unsuitable as a scholarly reference in most academic and all journal contexts.

How do I avoid confirmation bias in my literature search?

Actively search for studies that contradict your hypothesis. Use search terms like “critique of [your topic]”, “limitations of [your method]”, or “contrary evidence [your field]”. A paper that engages seriously with opposing evidence is more convincing: and more publishable: than one that only supports its thesis.

My paper was desk-rejected. What should I do?

A desk rejection means the editor decided not to send the paper for peer review. Common reasons include: the topic falls outside the journal’s scope, the manuscript does not meet formatting requirements, the contribution is insufficiently novel, or there are significant methodological flaws. Read the rejection note carefully, revise accordingly or target a better-matched journal, and resubmit. Desk rejection affects approximately 21% of all submissions to major journals: it is not a verdict on the quality of your research. A desk rejection shield service can also help you fix avoidable issues that lead to desk rejection, prior to submission.

Can I use AI to write a research paper?

You can use AI academic writing assistants like Paperpal to help you with outlines, turn rough notes into drafts, and to polish language, tone, and formatting. However, you should not use an entire paper generated by AI; that is considered unethical. Remember that free AI tools like ChatGPT are prone to hallucinations: they can completely fabricate data and citations. Go by the ICMJE guidelines on AI usage in research papers or by your institute’s policies.

Quick-Reference Checklist

This article was originally published on September 21, 2022, and updated on June 9, 2026.