How to Write a Research Paper

How research papers differ from essays

An essay usually develops an interpretation or argument using assigned readings. A research paper investigates a focused question using sources you locate and evaluate yourself. It may also report original data. The difference affects the scope, evidence, and organization of the paper.

For example, a five-page essay might explain how two assigned authors describe student sleep. A ten-page research paper might ask what studies published from 2019 through 2024 show about the relationship between nighttime social media use and sleep quality among first-year college students. The second task requires a search strategy, several scholarly sources, source comparison, and possibly a methods section.

Weak research-paper plan: “I will introduce social media, explain three negative effects, add quotations from websites, and write the references at the end.” Strong research-paper plan: “I will define nighttime use, explain how the literature measures sleep quality, compare findings from surveys and experiments, assess whether the evidence supports causation, and record every source when I save it.”

• Scope: An essay may interpret one assigned text; a research paper may compare ten studies that use different samples or methods.
• Sources: An essay may rely on a course reading list; a research paper requires database searching and decisions about which evidence is credible and relevant.
• Structure: An essay often uses an introduction, body, and conclusion; an empirical research paper may require an abstract, literature review, methods, results, discussion, and references.
• Responsibility: In a research paper, you must preserve the path back to every fact, quotation, dataset, and study. A note such as “sleep study online” is not enough.

Turn the assignment into an answerable question

Begin with the assignment's limits. Record the required length, number and type of sources, citation style, disciplinary format, and whether you must collect data. An eight-page literature review with eight peer-reviewed sources requires a different plan from a laboratory report based on an experiment.

A workable research question identifies the issue and places boundaries around it. Useful boundaries include population, location, time period, text, policy, variable, or type of evidence. You do not need every boundary, but you need enough to make the question answerable at the assigned length.

Weak question: “How does social media affect mental health?” This could include every platform, age group, diagnosis, country, and research method. Strong question: “What do peer-reviewed studies published from 2019 through 2024 suggest about the relationship between nighttime Instagram use and self-reported sleep quality among first-year U.S. college students?” The stronger version defines a population, platform, outcome, evidence base, and period.

• Test the length: An eight-page paper usually cannot explain the history, benefits, harms, regulation, and psychology of all social media. It could compare how three or four studies measure one outcome.
• Test the evidence: Search the key concepts before committing. If no studies examine Instagram specifically, broaden the platform term to “social media” rather than filling the paper with unrelated sources.
• Test access: A question about confidential university health records is not feasible unless you have approved access. A literature-based question using published studies is feasible.
• Test neutrality: “Why does Instagram destroy students’ sleep?” assumes the answer. “How is nighttime Instagram use associated with sleep quality?” allows for positive, negative, mixed, or inconclusive findings.
• Test the required method: If the assignment prohibits original human-subject research, do not design a student survey. Frame the paper as a review of published evidence.

Search databases with a keyword plan

Search by concepts rather than typing the entire research question into one box. For the question about Instagram and sleep, the main concept groups are college population, social media use, and sleep. Add synonyms within each group and connect the groups with AND.

A useful search could be: (“first-year students” OR freshmen OR undergraduates) AND (Instagram OR “social media”) AND (“sleep quality” OR insomnia OR “sleep disturbance”). Quotation marks hold phrases together, OR retrieves synonyms, and AND requires separate concepts. Database rules vary, so check its search help.

Weak search: “Does Instagram cause bad sleep in college freshmen?” Repeating this natural-language search tends to return popular articles and a narrow set of results. Strong search: Run several documented keyword combinations in PsycINFO, PubMed, or another relevant database, review the subject terms used by good articles, and revise the search with those terms.

• Start with the university library search system for a broad scan. Use a subject database for a more controlled search, such as PsycINFO for psychology, ERIC for education, PubMed for health, or MLA International Bibliography for literature and language.
• Use Google Scholar for discovery and citation tracing, not as proof that a source is peer-reviewed. Check the journal and publication record before using the item.
• Apply filters after testing the search. For example, limit results to peer-reviewed journal articles from 2019–2024 only if those limits match the question.
• Open a relevant article's reference list to find earlier studies. Use “Cited by” tools to find later studies that responded to it.
• Keep a search log. A useful entry reads: “PsycINFO, February 12, 2025; query: undergraduates AND social media AND sleep quality; filters: peer-reviewed, 2019–2024; 63 results; saved 8.” This lets you repeat or explain the search.
• Do not chase a fixed source count by collecting near-duplicates. Eight studies using the same type of survey may provide less useful coverage than a smaller set containing a review, several empirical studies, and a study with a contrasting method.

Recognize scholarly and relevant sources

A scholarly source normally identifies its authors, publication venue, evidence, method, and references. Peer-reviewed journal articles are common scholarly sources, but a PDF file or a result found through a library database is not automatically scholarly. Databases also index editorials, book reviews, news items, and trade publications.

Match the source type to the claim. To report how common sleep disruption is among students, use the original survey or a rigorous review rather than a newspaper summary of the survey. A news article may help explain public discussion, but it should not replace the underlying research.

Weak source set: a wellness blog, a company survey, two news reports, and one journal article, all selected because they agree that phones harm sleep. Strong source set: peer-reviewed studies with different designs, a review article that maps the field, and relevant counterevidence, with each source assessed for sample, measurement, date, and limitations.

• Empirical article: Reports original data and explains participants or materials, procedures, measures, and analysis. Example: a study comparing phone-use logs with sleep diaries.
• Review article: Synthesizes previous studies. Example: a systematic review of digital media use and adolescent sleep.
• Theoretical article: Develops or tests a conceptual explanation. Example: a model explaining why notifications may delay sleep onset.
• Scholarly book or chapter: Often useful in history and the humanities. Check the academic publisher, author expertise, citations, and relevance to the exact question.
• Popular source: Written for a general audience and often lacks methods or full references. Use it only when the public account itself is part of the research question.
• Quality check: A study of 30 volunteers from one class cannot automatically support a claim about all U.S. undergraduates. Record the limitation instead of hiding it.

Read strategically and take retrievable notes

Do not read every source from the first line to the last before deciding whether it is useful. First read the abstract, headings, conclusion, and figures or tables. Then inspect the method and limitations. Read closely only after confirming that the source addresses your question.

Every note should preserve its source and location. Record the author, year, page or section, exact wording when quoted, your paraphrase, the study's method, and the reason the evidence matters. Mark copied words as quotations immediately; otherwise, they may later be mistaken for your own prose.

Weak note: “Nguyen says phones ruin sleep. Good for body paragraph.” Strong note: “Nguyen, 2022, p. 14: survey of 412 first-year students; self-reported use after midnight was associated with lower sleep-quality scores. Observational design does not show causation. Use in section on measurement limits.” The stronger note preserves the finding, method, limitation, location, and planned use.

• Citation key: Use a short label such as Nguyen2022 so the note can be matched to the reference record.
• Complete source details: Save authors, date, article title, journal or book title, volume, issue, pages, DOI, and a stable link where applicable.
• Evidence: Separate exact quotation from paraphrase. Add a page number or named section, even when your citation style does not require a page for paraphrases.
• Method: Record the sample, data source, research design, measures, and analysis relevant to your evaluation.
• Limitations: Note issues such as a small sample, self-reporting, missing demographic information, or a short observation period.
• Your use: Write a sentence such as “contrasts with Lee's phone-log study” or “defines sleep quality.” This turns notes into material for synthesis.
• File management: Save a PDF as Nguyen_2022_NightUse.pdf rather than download(7).pdf, and keep it in the same folder or reference manager as its citation record.

Match the structure to your discipline and assignment

There is no single structure for every research paper. Use the assignment sheet, grading rubric, and model papers from the course or target discipline. Do not add a results section if you did not collect or analyze results, and do not invent a methods section merely to make a humanities paper look scientific.

For example, an empirical psychology report usually follows an introduction-methods-results-discussion pattern. A history paper may organize evidence chronologically or by competing explanations. A literature review may explain its search and selection process, then arrange findings by theme rather than report original experimental results.

• Empirical social science: Title, abstract if required, introduction and literature review, methods, results, discussion, conclusion if separate, and references.
• Laboratory science: Title, abstract, introduction, materials and methods, results, discussion, and references. Figures and tables may appear in designated locations.
• Literature review: Introduction, scope or search method when required, thematic or methodological synthesis sections, gaps and limitations, conclusion, and references.
• Humanities research paper: Introduction, necessary historical or critical context, sections organized around texts or evidence, engagement with scholarship throughout, conclusion, and works cited or bibliography.
• Policy or business research paper: Executive summary if requested, problem definition, evidence or analytical method, findings or options, evaluation, recommendations, and references.
• Course-specific format: A professor may combine the literature review with the introduction or require a separate limitations section. Follow that requirement even if published articles use another structure.

Give each research-paper section a distinct job

A section earns its place by doing a specific part of the research task. The introduction defines the problem and question; the literature review establishes what sources collectively show; the methods section explains how evidence or data was produced; the results report findings; and the discussion interprets those findings.

Methods must be precise enough for the reader to understand what was done. Weak literature-review method: “I researched articles online about college students and sleep and selected the most relevant ones.” Strong version: “I searched PsycINFO and ERIC on February 12, 2025, using combinations of undergraduates, social media, nighttime use, and sleep quality. I included peer-reviewed English-language studies published from 2019 through 2024 that reported undergraduate samples and excluded studies focused only on children.”

Keep results and interpretation separate when the format requires separate sections. Weak results sentence: “Students with more nighttime use slept less, proving that social media is harmful.” Strong results sentence: “The high-use group reported a mean of 34 fewer minutes of sleep than the low-use group.” The discussion can then assess whether the design supports a causal claim.

• Title: Name the specific relationship or problem. “Nighttime Social Media Use and Sleep Quality Among First-Year Students” is more informative than “Social Media Research Paper.”
• Abstract: Summarize the question, method or evidence base, central finding, and implication. Write it after the paper so it reflects the completed work.
• Introduction: Define the problem, give only the context needed to understand it, state the research question, and indicate the paper's approach.
• Literature review: Organize studies by pattern, debate, method, or gap. Do not turn it into a sequence of article summaries.
• Methods: For original research, describe participants or materials, procedures, measures, and analysis. For a structured review, describe databases, search terms, dates, and selection criteria when required.
• Results: Report what the analysis found, using tables or figures where appropriate. Do not introduce unrelated sources or exaggerate the finding.
• Discussion: Explain how the findings answer the question, compare them with earlier studies, address limitations, and state what the evidence can and cannot support.
• Conclusion: Give the answer at the level justified by the evidence. Do not introduce a new study or claim in the final paragraph.
• References: Include every cited source in the required style. Unless the assignment says otherwise, do not list material that the paper never cites.

Weak and stronger research moves

Each pair starts with what a first research paper usually does with sources, followed by the version that holds together.

Common mistakes and how to fix them

Choosing “social media and mental health” as the topic for an eight-page research paper.

Limit the question by population, platform, outcome, period, or evidence type. For example, examine recent studies of nighttime Instagram use and sleep quality among first-year college students.

Typing the complete research question into one database search and using only the first results.

Divide the question into concept groups, add synonyms with OR, connect concepts with AND, and test several versions in subject databases.

Treating every PDF, database result, or Google Scholar result as peer-reviewed research.

Verify the source type, journal, author information, method, references, publication status, and relevance to the exact research question.

Saving sources as browser tabs or bookmarks without full citation information.

Create a source record as soon as you save an item. Include complete publication details, DOI or stable link, page locations, and a clear file name.

Writing the literature review as one paragraph about each source.

Group sources by finding, method, concept, or disagreement. Compare multiple sources within each section.

Drafting beside open articles and changing a few words from source sentences.

Take attributed notes first, close the articles, and draft from the notes. Reopen each source afterward to verify accuracy and citation details.

Writing a methods section that says only that relevant articles were researched.

For a review, identify databases, search terms, search dates, limits, and inclusion criteria. For original research, explain participants or materials, procedures, measures, and analysis.

Related guides and tools

• Annotated bibliographies What goes in an annotation
• Citation generator APA, MLA, Chicago from a DOI or ISBN
• Word counter Words, pages, and reading time
• Research paper service Source-based work
• Research essays Turning a topic into a question

Frequently asked questions