To write a research paper: turn the assignment into one answerable question, gather and evaluate sources, stake a thesis that answers it, build an outline, draft section by section, then run three revision passes before formatting the citations. Work the phases in order and the paper assembles itself; skip one and it falls apart at 2 a.m. the night before.
What a research paper actually is (and isn't)
Strip away the anxiety and a research paper is a documented argument: you ask a question nobody at your table can settle from memory, you read what scholars have already found, and you take a position your sources support. The length, the citation style, the formal tone — those are packaging. The argument is the product.
Three things it is not:
- Not a summary. Stringing five article abstracts together with connective sentences is a report, and graders mark reports as C work. Every section should exist to move your thesis forward.
- Not an opinion piece. "I just feel that school uniforms are bad" carries no weight. The claim only counts when a cited source, a dataset or a worked example backs it.
- Not a mystery novel. Don't reveal what you found in the conclusion. State the thesis in the introduction and spend the paper proving it.
Throughout this walkthrough we'll follow one illustrative running example — a student writing about why some city streets stay cooler than others during summer heat waves. Every number or claim attached to that example here is hypothetical, labeled so, and exists only to show the mechanics. Swap in your own topic and the moves stay identical.
The seven phases, in order
Each phase has one job. Finish it before starting the next, because each phase feeds the one after it — a vague question in phase 1 becomes a doomed thesis in phase 3.
-
Pick a question you can actually answer
Take the broad topic from the brief and cut it until it fits the word count. "Urban heat" is a book; "why do streets with more tree cover feel cooler at street level" is a 2,500-word paper. Test: could a classmate argue the other side in one sentence? If yes, you have a question. If no, it's still a topic.
-
Gather and evaluate sources
Six to ten solid sources beat thirty skimmed ones. Run each candidate through the five-question source test below before it earns a place in your notes.
-
Stake a thesis
One sentence, debatable, answerable inside the word count. It names the position and — this is the part students skip — hints at the reasoning. The thesis statement generator drafts candidates in under a minute; you keep the best one.
-
Build the outline
Map every section, assign it a job sentence, and budget words. The research paper outline guide generates a skeleton with live word budgets for your paper type — argumentative, analytical or compare/contrast.
-
Draft ugly, draft fast
Write the body sections first (the introduction is easier once the body exists). Don't edit mid-sentence; a finished messy draft beats a polished half-page every time. Watch the budget: if your example section on canopy coverage runs 700 words of a 2,500-word paper, the tail sections starve.
-
Revise in three passes
Argument pass, evidence pass, sentence pass — in that order, because polishing sentences in a paragraph you're about to delete is wasted work. Details below.
-
Format and cite
Fix margins, headings, title page and references last, when nothing will move again. APA? Use the APA paper format guide. MLA? Use the MLA paper format guide. Generate entries with the APA citation generator or MLA citation generator and check them against your style manual.
Phase 2 in detail: the five-question source test
Before a source earns a line in your notes, it answers five questions. Graders call the underlying skill "source evaluation"; libraries teach it under many names, and the acronym some use is unrepeatable in polite company. The five questions:
- Currency — when was it published? In fast-moving fields, three years is old; in history, three centuries can be fine. The point is knowing, not rejecting.
- Relevance — does it speak to your question, or only to a neighboring one? "Urban heat" sources won't help a paper on rural farming.
- Authority — who wrote it, and what makes them credible? Peer-reviewed journals and university presses rank above anonymous blogs.
- Accuracy — can you trace the evidence? Sources that cite their own data or named studies beat sources that gesture at "experts say."
- Purpose — is it informing, selling or persuading? A think-tank brief can be useful, but you must know which side funds it.
Tap each test as you run it on a real source.
One more habit worth stealing: read the abstract and the conclusion before committing. Five minutes per source here saves hours later, because a source that only half-fits your question will fight you in every draft. Once the paper is drafted, the research paper abstract guide compresses it into six budgeted sentences.
Note-taking that doesn't backfire
Most citation panic at the end of a paper is note-taking debt from the middle. The fix is mechanical:
Three columns, always
For every note, record: the source (author, year, page), the idea (quote, paraphrase or summary — labeled as such) and the use (which outline section it supports). A note missing any column is a future dead end.
Cite at capture time
Build the citation entry the moment you take the note, not the night before submission. Draft it with the bibliography generator, park it in a running list, and formatting day becomes copy-paste day.
Quote sparingly: two or three quoted lines per section, paraphrase the rest, and always translate a source's point into your own sentence while you understand it — paraphrasing from memory a week later is where accidental plagiarism is born.
Thesis: weak vs strong
The thesis is one sentence and the whole paper hangs on it. Compare these two (illustrative, from our running heat-island example):
"There are many interesting factors that affect how hot city streets get in summer."
Why it fails: it announces a topic instead of a claim. Nothing here can be wrong, so nothing here can be argued. "Many interesting factors" also tells the grader you haven't picked among them yet.
"In the illustrative scenario above, shade from street trees explains street-level temperature gaps better than pavement color does, because canopy lowers surface absorption for most of the day while pale pavement only helps at noon."
Why it works: it takes a side (canopy beats pavement color), previews the reasoning (duration of shade vs noon-only effect), and is falsifiable — a source could disagree, which is exactly what makes it a thesis.
Notice the pattern: claim + because + preview. If your sentence lacks a "because," it's a topic, not a thesis.
The outline you can copy
Here is the skeleton in plain text — the same structure the outline builder generates with word budgets. Copy it, paste it into your editor, and replace the bracketed parts:
Budget rule of thumb: if a section's share of the outline can't justify its share of the word count, either the outline is lying or the section is padding. Fix it now, in outline form, where fixes are free.
Section anatomy: what each part is for
| Section | Job in one sentence | Failure mode to avoid |
|---|---|---|
| Introduction | Move the reader from "so what?" to "tell me more" in three moves: hook, context, thesis. | Starting with "Since the dawn of time…" |
| Background / lit review | Show the conversation your paper joins and the gap it fills. | Summarizing source after source with no thread of comparison. |
| Body sections | One claim each, proved with evidence, tied back to the thesis. | Evidence dropped in with no "so what" sentence after it. |
| Counter-argument | State the best objection fairly, then answer it. | Attacking a straw man nobody holds. |
| Conclusion | Restate, imply, and stop — no new evidence, no new citations. | Copy-pasting the introduction with synonyms. |
Word counts vary by assignment, but the proportions hold from 1,500-word essays to 5,000-word term papers: about ten percent for the introduction, half to two-thirds for body sections, and whatever is left split between counter-argument and conclusion. For the last section, the research paper conclusion guide shows how to answer your research question and land the implication without padding.
Revision: three passes, three different brains
Do not revise on the same day you finish drafting — you'll read what you meant, not what you wrote. Then run these passes separately:
Argument pass (read the first sentences only)
Read just the first sentence of every paragraph, in order. If that sentence-chain alone tells the whole argument, the skeleton holds. If not, reorder sections before touching anything else.
Evidence pass (audit every claim)
Highlight every factual claim. Each one needs a citation or a flag that it's your reasoning. Claims with neither get sourced or cut — this is the pass that keeps you out of plagiarism trouble.
Sentence pass (read aloud)
Read the paper aloud, or feed it to the readability checker. Flag every sentence you stumble on; shorten the ones over 30 words; delete every "very" that survives.
Then run the mechanical checks: word count against the brief (word counter), citations matched both ways, and formatting per your style guide.
Phase checklist: where are you now?
Tap each phase as you complete it. Honest taps only — the bar doesn't lie to your grader either.
0 of 7 phases
- Phase 1 — my question fits the word count and has a debatable other side
- Phase 2 — every source has passed the five-question test
- Phase 3 — my thesis states claim + because + preview
- Phase 4 — my outline assigns every section a job and a word budget
- Phase 5 — a full ugly draft exists, body sections first
- Phase 6 — all three revision passes are done on separate reads
- Phase 7 — formatting and references match the required style
Quiz: five questions before you draft
1. Which of these is a workable research paper question?
2. A source is a 2022 advocacy brief funded by a pavement company, arguing pale pavement beats trees. What do you do?
3. Which thesis earns marks?
4. Your introduction is done. When should you write it again?
5. In the argument revision pass, what do you actually read?
Stuck on phase 5 with the deadline tomorrow?
If the draft won't cooperate, a quoted, deadline-bound writer can take the paper off your hands while you keep the outline and sources. Get a free price quote and see the terms first.
Frequently asked questions
How long does it take to write a research paper?
It depends on length and how much you already know, but a workable rule is: half the total time on phases 1–4 (question, sources, thesis, outline) and half on phases 5–7. Students who skip the first half always pay it back with interest in the second, because a weak outline makes every paragraph twice as hard.
How many sources does a research paper need?
Follow your brief first. As a floor, think in terms of claims: every body section usually needs one to three sources behind it, so a six-section paper rarely survives on fewer than six solid sources. Quality beats quantity — six sources you actually read outrank twenty you skimmed.
Can I change my thesis after I start writing?
Yes, and experienced writers often do — drafting is where you discover what the evidence actually supports. When the thesis changes, update the outline the same day, or the body will argue two different papers. The thesis serves the evidence, never the reverse.
Should I write the introduction first?
Write it last, or at least rewrite it last. The introduction must promise exactly what the body delivers, and you only know that delivery precisely once the body exists. A placeholder intro keeps you moving; the real intro comes in the revision passes.
What's the difference between a research paper and a case study?
A research paper argues a general claim from scholarly sources; a case study dissects one specific situation and often ends in recommendations. If your assignment centers on a named company or scenario, the case study writing guide is the better starting point; if it centers on a question, you're in the right place. And when the case is the research vehicle itself — a named firm studied with a research instrument behind it — the business research case study page walks that hybrid end to end.
How do I avoid plagiarism when paraphrasing?
Close the source, then write the idea from memory in your own words, then reopen and check for accuracy — and cite it anyway. Paraphrasing changes the wording, not the ownership of the idea. If three words in a row match the original, either requote or reword.
Was this guide useful?