How to Start Writing a Research Paper: A 10-Step Guide

Key Takeaways:

• Start with a focused research question, validated against the FINER criteria, before you write a single sentence of your paper.
• Write the Methods section first, not the Introduction; it is the easiest section to draft and builds early momentum.
• Spend disproportionate time on your Title and Abstract; they are the only parts most readers, editors, and reviewers will ever see.
• Plan for revision from day 1: budget at least 25-30% of your total writing time for editing, journal alignment, and co-author feedback.

Glossary of Key Terms

Before diving in, here are the key terms used throughout this guide. Bookmark this table if you are new to academic publishing.

Why Is Starting a Research Paper So Hard?

Starting a research paper is hard because 3 obstacles hit at once: unfamiliarity with academic writing conventions, uncertainty about how to structure data and arguments, and the pressure to sound polished from the first sentence. Each obstacle feeds procrastination, and procrastination compounds the problem as deadlines approach.

The research itself, designing experiments, collecting data, running analyses, often feels energizing. Writing it down feels like a different job entirely, and in many ways it is. The good news: writing a research paper is a learnable, repeatable process, not a talent. This guide breaks that process into 10 concrete steps, with frameworks, examples, templates, and tools at every stage.

1 mindset shift before you begin: separate generating from polishing. First drafts exist to get ideas out; revision exists to make them good. Researchers who accept this write faster and procrastinate less.

Step 1: Define a Clear, Focused Research Question

Knowledge without direction is like a boat on dry land. Everything in your paper, the literature you cite, the methods you choose, the story your discussion tells, flows from your research question. A vague question produces a vague paper; a sharp question practically outlines the paper for you.

What Makes a Good Research Question?

A good research question is specific, answerable with the data and methods available to you, and anchored in a genuine gap in the literature. It should be narrow enough to answer in 1 paper, yet significant enough that other researchers care about the answer.

Test your draft question against these 4 quick checks:

• Specificity: does it name a defined population, variable, or context, rather than a broad theme?
• Answerability: could you describe, in 1 paragraph, the data that would answer it?
• Novelty: has the exact question already been answered? If yes, what angle is still open?
• Scope: can it be answered within your time, budget, and word limit?

Use the FINER and PICO Frameworks

2 established frameworks can pressure-test your question. FINER works across all disciplines; PICO is standard in health sciences and other intervention-based fields.

PICO structures clinical and intervention questions in 4 parts:

• P (Population): who is being studied, e.g., adults over 65 with type 2 diabetes.
• I (Intervention): the treatment, exposure, or condition of interest.
• C (Comparison): the alternative against which the intervention is judged.
• O (Outcome): the measurable result, e.g., HbA1c reduction at 12 weeks.

Weak vs. Strong Research Questions: Examples

Step 2: Conduct a Thorough Literature Review

Before designing your study around a question, verify that the gap you see is real. A literature search tells you what is already known, which methods dominate the field, where researchers disagree, and how your work will fit into the conversation. It also supplies the citations that give your paper credibility.

Which Databases Should You Search?

Search at least 2-3 databases: 1 broad multidisciplinary index plus 1-2 field-specific ones. No single database covers everything, and relying on 1 source risks missing key studies and inflating your perceived gap.

Build Effective Search Strings

Databases reward precise queries. Combine synonyms with Boolean operators, and refine iteratively: start broad, scan the first 50 results, then tighten.

2 more techniques multiply your coverage. Backward citation chaining means mining the reference lists of key papers for earlier work. Forward chaining means using the database’s cited-by feature to find newer studies that build on those papers. Together they often surface literature that keyword searches miss.

Choose a Reference Manager Early

Set up a reference manager before you collect your first PDF, not after you have 80 unsorted files. These tools store papers, generate citations in any style, and sync with your word processor.

Synthesize With a Literature Matrix

Reading is not reviewing. To move from summarizing papers 1 by 1 to synthesizing across them, build a literature matrix: a table with 1 row per study. Patterns, contradictions, and gaps become visible at a glance, and your review section almost writes itself.

Step 3: Formulate a Research Statement

With your question validated and the literature mapped, distill your paper’s purpose into a research statement. This 1-3 sentence core tells readers exactly what you studied, how, and why it matters. Every section you write later should trace back to it.

Research Statement vs. Thesis Statement: What Is the Difference?

A research statement describes what your study investigates and how; a thesis statement asserts the answer or argument you will defend. Empirical papers typically open with a research statement, while argumentative humanities papers lead with a thesis.

In practice, many papers use both: the introduction ends with a research statement (aims and approach), and the discussion or conclusion delivers the thesis-like claim your evidence supports.

A Formula You Can Use

Fill in this template, then compress it: “This study examines [specific question] in [population or context] using [method], in order to [contribution to the gap].”

• Before: “This paper looks at social media and sleep among students, which is an important topic today.”
• After: “This study examines whether daily short-video use above 2 hours predicts reduced sleep quality among 300 undergraduates, using wrist actigraphy, to provide the first objective evidence on a platform type absent from prior meta-analyses.”

The revised version is longer but infinitely more useful: it commits you to a population, a measure, a method, and a contribution. If you cannot fill in the template yet, return to Steps 1 and 2 before drafting.

Step 4: Create a Strong Research Paper Outline

An outline is a mind map of your argument: it decides what goes where before you invest hours in prose. Papers drafted from outlines are faster to write, easier to keep coherent, and far easier for co-authors to review at the planning stage, when changes are cheap.

The IMRaD Structure

Most empirical papers follow IMRaD: Introduction, Methods, Results, and Discussion, framed by a Title, Abstract, Conclusion, and References. The table below shows what each section does and a typical share of the word count for a standard 5,000-word paper.

See also: Discussion vs Conclusion: What is the Difference?

Under each section heading in your outline, add 3-5 bullet points naming the specific content: the datasets, the figures, the arguments. A good test: someone reading only your outline should be able to predict every figure and table in the final paper.

Outlining also works in reverse. After finishing a draft, write 1 sentence summarizing each paragraph; this reverse outline exposes repetition, gaps, and paragraphs that wandered off topic.

When Does IMRaD Not Apply?

IMRaD does not fit humanities papers, theoretical or mathematical work, review articles, or case studies; each has its own conventions. Check 3-5 recent papers in your target journal and mirror their structure.

• Humanities and law: thematic sections built around an argument, often without a separate methods section.
• Theory and mathematics: definitions, propositions, and proofs replace methods and results.
• Review articles: organized by themes, chronology, or methodological schools, often following PRISMA guidelines for systematic reviews.
• Case studies and case reports: background, case presentation, discussion, and take-home messages.

Step 5: Begin Working on the First Draft

The first draft has exactly 1 job: getting your research out of your head and onto the page. Do not expect elegance; expect gaps, clumsy sentences, and placeholder notes like [ADD CITATION]. All of that is normal and fixable. What is not fixable is a blank page.

A timing tip many researchers miss: do not wait until the study is finished to start writing. Drafting the Methods while you run the study, and the Results as analyses complete, captures details accurately, exposes design problems early, and shrinks the intimidating final write-up into small, manageable installments.

Which Section Should You Write First?

Write the Methods section first. It is purely descriptive, requires no persuasion, and you know the material better than anyone; most writing experts recommend it as the momentum-builder. A proven order for the full draft:

• Methods: describe what you did while it is fresh.
• Results: build your figures and tables, then write the text around them.
• Discussion: interpret the results against the literature matrix from Step 2.
• Introduction: now you know exactly what you are introducing.
• Conclusion, Abstract, and Title: distill the finished story last.

Drafting Strategies That Beat Writer’s Block

• Timebox: write in 25-45 minute focused sessions with notifications off; 2 sessions a day outperforms rare marathon weekends.
• Set output targets: 300-500 words per session is sustainable; some days you will exceed it.
• Write ugly on purpose: draft in plain sentences and fix style later; separating drafting from editing is the single biggest speed unlock.
• Use placeholders: drop [CHECK STAT] or [CITE] markers instead of breaking flow to hunt for details.
• Never stop at a blank spot: end each session mid-section, with a 1-line note on what comes next, so restarting is frictionless.
• Talk it out: if a paragraph will not come, explain the point aloud as if to a colleague, record it, and transcribe.

Step 6: Write the Methods and Results Sections

These 2 sections are the factual spine of your paper, and reviewers scrutinize them hardest. They are also the most formulaic, which makes them the best place to build early momentum.

Writing the Methods

The gold standard is replicability: a competent peer should be able to repeat your study from this section alone. Write in past tense, use passive voice where the focus is not on who is doing the action, and organize under conventional subheadings.

• Participants or materials: who or what you studied, with inclusion and exclusion criteria, sampling strategy, and sample size justification.
• Design and procedure: what was done, in what order, under what conditions.
• Measures and instruments: every variable, how it was operationalized, and the validity and reliability of each instrument.
• Analysis: statistical tests or analytical methods, software and versions, and how you handled missing data and outliers.
• Ethics: approval body and protocol number, consent procedures, and data protection measures.

Presenting Results, Figures, and Tables

Report findings; do not interpret them (interpretation belongs in the Discussion). Lead with the result that answers your research question, then present secondary findings. Report effect sizes and confidence intervals alongside p-values, and report results that did not support your hypothesis with equal transparency.

Decide early what becomes a figure, a table, or text:

• Figures show patterns: trends, group differences, distributions, and relationships.
• Tables show precision: exact values, descriptive statistics, and model outputs.
• Text handles the rest; never repeat in prose what a figure or table already shows, just point to the key takeaway.
• Every figure and table needs a self-contained caption: a reader should understand it without the body text.
• Number figures and tables in order of mention, and cite each 1 explicitly in the text.

Step 7: Write a Compelling Introduction

The introduction is your paper’s gateway, but a research paper is not a magazine feature: editors and reviewers are rarely won over by dramatic anecdotes. What persuades them is a tightly argued case that a specific, important gap exists and that your study fills it. Fortunately, there is a well-studied formula for exactly that.

The CARS Model: 3 Moves

The Create a Research Space (CARS) model, derived from analysis of hundreds of published introductions, structures the section as 3 moves:

Common introduction mistakes to avoid: opening with sweeping claims (“Since the dawn of time…”), reviewing literature exhaustively instead of selectively, burying the research question in the middle, and previewing your results in detail. End the introduction with your research statement from Step 3; readers expect to find it there.

Step 8: Spend Time on Your Abstract and Title

Researchers often perfect the Methods and Results while dashing off the Title and Abstract in the final hour. That is backwards: for every person who reads your full paper, dozens will read only these 2 elements. They determine whether your paper is found by search engines and databases, opened by readers, sent to review by editors, and ultimately cited.

How Long Should an Abstract Be?

Most journals set abstract limits between 150 and 300 words; 250 words is the most common ceiling. Always check your target journal’s guidelines, because exceeding the limit can block submission outright. Whatever the length, the abstract must stand alone: no citations, no undefined abbreviations, and no claims absent from the paper.

A reliable recipe for an unstructured abstract: 1-2 sentences of context and gap, 1 sentence on aim, 2-3 sentences on method and sample, 2-3 sentences on key results with numbers, and 1 sentence on the main implication. Write it last, when the story is settled, and include your main keywords naturally for discoverability.

Title-Writing Patterns That Get Papers Read

An effective title is specific, honest, and rich in the keywords your audience actually searches. Aim for 10-15 words, avoid filler like “A study of” or “An investigation into,” and skip abbreviations and question marks unless your field favors them.

Step 9: Write the Discussion and Conclusion

The Discussion is where your paper earns its citations: it converts results into meaning. A dependable structure moves from the specific to the general: restate the key finding, compare it with prior studies (your literature matrix pays off again here), explain agreements and discrepancies, acknowledge limitations candidly, and state implications for research or practice.

• Interpret; do not repeat. Reference results, but do not restate every number from Step 6.
• Address limitations yourself, with their likely direction of bias; reviewers will find them anyway, and preemption reads as rigor.
• Avoid overclaiming: match the strength of your language to the strength of your evidence, and never present correlation as causation.

The Conclusion then answers your research question in plain, memorable language. Keep it short: summarize the answer, its 1-2 main implications, and a focused direction for future work. Introduce nothing new here; a conclusion that raises concepts never discussed in the body signals a structural problem and irritates reviewers.

Step 10: Revise, Polish, and Align With Journal Guidelines

Poor language and mismatched formatting are among the most common, and most avoidable, reasons for desk rejection. Budget real time for this step: 25-30% of your total writing effort is a sensible allocation, spread across multiple passes rather than 1 heroic all-nighter.

See also: 10 Best Scientific Editing Services for Research Papers in 2026

Journal Alignment Checklist

Choose your target journal before final revisions, then verify every item below against its author guidelines:

• Scope: does your paper fit the journal’s stated aims? Skim 3-5 recent issues to confirm.
• Article type and word limit: original article, short communication, or review, each with different ceilings.
• Structure and headings: some journals mandate specific section names and ordering.
• Citation style and reference limits: switch your reference manager to the journal’s style with 1 click.
• Figures and tables: file formats, resolution, color policies, and maximum counts.
• Declarations: funding, conflicts of interest, ethics approval, author contributions, and data availability statements.
• Supplementary material: what the journal allows, and how it must be labeled.

See also: 10 Best Manuscript Formatting Services for Journal Submissions (2026)

Edit in 3 Passes, Not 1

• Pass 1, structure: use a reverse outline to check flow, remove repetition, and confirm every paragraph serves the research question.
• Pass 2, clarity: shorten sentences (aim under 25 words on average), cut jargon your target reader will not know, prefer active voice, and define every abbreviation at first use.
• Pass 3, correctness: subject-verb agreement, article use, capitalization, colons and semicolons, number style and units, US vs. UK English consistency, numerical accuracy between text and tables, and citation-reference format.

2 finishing moves help enormously: read the paper aloud (awkward phrasing becomes audible), and let the manuscript rest for 48 hours before the final pass so you read what is actually on the page rather than what you meant to write. Then run a plagiarism check to catch unintended overlap, including with your own earlier publications.

Beyond the 10 Steps: What Else Determines Success

The 10 steps produce a manuscript; the practices below determine how smoothly it travels through co-authors, ethics requirements, and peer review.

Citations, Referencing, and Plagiarism

Cite the source of every idea, statistic, and method that is not your own, including paraphrases. Your reference manager automates the formatting; your job is choosing the right style and citing honestly.

3 plagiarism traps catch honest researchers: paraphrasing too closely to the original wording, forgetting quotation marks around copied definitions, and self-plagiarism, reusing chunks of your own published text without disclosure. When in doubt, rewrite from memory with the source closed, then cite it anyway.

Authorship and Co-Author Collaboration

Authorship disputes are among the most corrosive conflicts in academia, and nearly all are preventable by deciding early. The widely used ICMJE criteria state that an author must contribute substantially to the work, help draft or revise it, approve the final version, and accept accountability for it.

• Agree on the author list, order, and corresponding author before drafting begins, and revisit if contributions change.
• Assign sections with deadlines; a shared outline (Step 4) makes ownership visible.
• Use 1 master document with tracked changes or a shared platform to avoid version chaos.
• Credit non-author contributors (data collection help, language editing) in the Acknowledgments.
• Never add honorary authors or omit deserving ones; many journals now require a signed contribution statement.

Ethics, Declarations, and Data Availability

Handle ethics before data collection, not at submission. Journals increasingly desk-reject papers with missing or retrofitted declarations.

• Ethics approval: obtain institutional review board (IRB) or ethics committee approval for human or animal studies, and quote the protocol number in the Methods.
• Informed consent: describe how participants consented, including for identifiable images or case details.
• Conflicts of interest: declare funding sources and any financial or personal relationships that could appear to bias the work; “none” is itself a required statement.
• Data availability: state where your data and code can be accessed, or the justified reason they cannot; many journals and funders now mandate repository deposit.
• Reporting guidelines: follow the checklist for your study type, such as CONSORT for trials, STROBE for observational studies, or PRISMA for systematic reviews.

Can You Use AI Tools When Writing a Research Paper?

Yes, within limits: most major publishers now permit AI assistance for language polishing (e.g., Paperpal), literature discovery (e.g., R Discovery), and formatting, provided you disclose the use and take full responsibility for accuracy. What is broadly prohibited is listing an AI tool as an author or presenting AI-generated content, data, or citations as your own verified work.

• Acceptable in most policies: grammar and clarity editing, translation support, keyword and search-string suggestions, and reference formatting.
• Risky or prohibited: generating results, fabricating or unverified citations, writing whole sections you have not checked, and uploading confidential or unpublished data to public tools.
• Always: read your target journal’s AI policy, disclose usage where required (often in Methods or Acknowledgments), and verify every AI-suggested fact and reference yourself.

Plan Your Writing Timeline

Work backward from your submission deadline using realistic estimates. For a first empirical paper written alongside other duties, the following is a sensible baseline; experienced writers compress it considerably.

After submission, expect 1-6 months to a first decision, and treat “revise and resubmit” as the good news it is: most published papers pass through at least 1 revision round. Respond to every reviewer comment point by point, courteously, even when you disagree.

Pre-Submission Checklist

Run through this list in 1 sitting before you press submit:

• Title is specific, keyword-rich, and within the journal’s length limit.
• Abstract stands alone, respects the word limit, and matches the paper’s actual claims.
• Introduction ends with a clear research statement; the Conclusion answers it directly.
• Methods are detailed enough for replication, with ethics approval and analysis software stated.
• Every figure and table is cited in the text, numbered in order, and has a self-contained caption.
• All in-text citations appear in the reference list, and vice versa, in the journal’s style.
• Declarations are complete: funding, conflicts, contributions, data availability, and any AI use.
• The manuscript has been proofread after a 48-hour rest, checked for plagiarism, and approved by every co-author.

Frequently Asked Questions

How long does it take to write a research paper?

For a first empirical paper, plan on 2-3 months of part-time work from literature review to submission-ready manuscript; experienced researchers often manage 3-6 weeks. The biggest variables are how clean your data are, how many co-author review rounds you need, and how much uninterrupted writing time you can protect each week.

How long should a research paper be?

Most original research articles run 4,000-8,000 words excluding references, with 5,000-6,000 as a common sweet spot; short communications run 1,500-3,000 words and review articles can exceed 10,000. Your target journal’s guidelines override any general rule, so check its stated limit for your article type before drafting.

Which section of a research paper should you write first?

Write the Methods section first: it is descriptive, requires no argumentation, and builds momentum fast. Then write the Results around your figures, the Discussion, and the Introduction, leaving the Conclusion, Abstract, and Title for last, when the paper’s story is fully settled.

What is the difference between an abstract and an introduction?

The abstract is a standalone 150-300 word summary of the entire paper, including its results and conclusions; the introduction is a full section that builds context, establishes the gap, and states your aims, without revealing detailed results. Databases index the abstract; the introduction is read only by people who open the paper.

See also: Introduction vs Literature Review in a Research Paper: What Is the Difference? 

How many references should a research paper have?

A typical original article cites 25-50 sources; reviews often cite 80-150 or more, and short communications may cite under 20. Quality beats quantity: cite the most relevant, recent, and authoritative sources for each claim, follow your journal’s reference cap if 1 exists, and never pad the list with papers you have not read.

Can you publish a research paper without a PhD?

Yes: journals evaluate manuscripts on the quality of the research and writing, not on the author’s degree. Undergraduates, master’s students, and independent researchers publish regularly, especially with a mentor or experienced co-author guiding journal selection, ethics requirements, and responses to peer review.

How do you write a research paper introduction that hooks readers?

Open with a specific, evidence-backed statement of why the topic matters (Move 1 of the CARS model), then expose the gap and state your aims. In academic writing the “hook” is relevance and a sharp gap, not drama: precise numbers, a striking documented trend, or an unresolved contradiction in the literature outperform anecdotes and sweeping generalizations.

See also: Introduction vs Background in a Research Paper: What Is the Difference?

R Discovery is a literature search and research reading platform that accelerates your research discovery journey by keeping you updated on the latest, most relevant scholarly content. With 250M+ research articles sourced from trusted aggregators like CrossRef, Unpaywall, PubMed, PubMed Central, Open Alex and top publishing houses like Springer Nature, JAMA, IOP, Taylor & Francis, NEJM, BMJ, Karger, SAGE, Emerald Publishing and more, R Discovery puts a world of research at your fingertips. 

Try R Discovery Prime FREE for 1 week or upgrade at just US$72 a year to access premium features that let you listen to research on the go, read in your language, collaborate with peers, auto sync with reference managers, and much more. Choose a simpler, smarter way to find and read research – Download the app and start your free 7-day trial today! 

This article was originally published on July 26, 2023, and updated on July 11, 2026.