How to Write an Essay That Argues (Not Summarizes)

The thesis is a claim, not a topic

A topic is what the essay is about; a claim is what the essay says about it that someone could dispute. 'Remote work' is a topic. 'Remote work kept mid-career employees but quietly starved junior staff of the mentorship that gets them promoted' is a claim — it names a winner, a loser and a mechanism, and a reasonable person could argue back. That arguability is the point. Instructors grade argument, and a sentence nobody can disagree with gives them nothing to grade. The Thesis Statement Builder on this page forces the distinction by making you choose a claim type — cause, evaluation, policy or interpretation — before it assembles candidates.

Placement is convention, and the convention is useful: one sentence, 20-40 words, at the end of the introduction, where readers have been trained to look. Be skeptical of the three-pronged thesis you learned in high school. Three parallel points are training wheels that often flatten an argument into a list — if the only way to defend your thesis is to count to three, it is a category, not a claim. A claim with tension ('the policy worked, but for the wrong reasons') generates better paragraphs than three parallel buckets, because tension forces each paragraph to answer the one before it.

What students get wrong about essays

The dominant failure is summary wearing an essay's clothes. The student recounts the plot, the history, or the study's methods, accurately and in order, and never quite claims anything. Graders diagnose it fast: body paragraphs open with authors' names instead of the writer's assertions, and the conclusion repeats the summary rather than the stakes. On a standard rubric, 'analysis and argument' typically carries more weight than 'mechanics' — often twice as much — yet struggling writers spend their editing hours on commas. The fix is unglamorous: write the claim sentence first, in your own words, before any source is allowed into the paragraph.

The second cluster of errors is scope arithmetic. A 1,000-word essay is about four double-spaced pages and holds two body points comfortably; students routinely cram five, which leaves each point 130 words — enough to announce it, never to prove it. Word count is not a bureaucratic hurdle; it is the assignment telling you how big a claim it will pay for. Match the claim to the budget: two points, developed through evidence and analysis, beat five points at headline depth every time. When a professor assigns 1,500 words, they are buying depth, not breadth.

Paragraphs: one job each

Every body paragraph advances exactly one point, and the anatomy is predictable enough to build from. The claim sentence states the point in your words — never in a quotation's. Evidence comes next: one or two pieces, a quote, a statistic, or an example, introduced with context. The analysis is two or three sentences explaining how the evidence supports the point; this is where the grade lives, and it is the part most drafts skip. A final link ties the point back to the thesis. For a 1,500-word essay, that anatomy repeats five or six times at 150-220 words per paragraph.

• One-job test: if you cannot state the paragraph's point in a sentence, it contains two points — split it.
• Word count: 120-220 words is the healthy band; a 350-word paragraph is almost always two paragraphs in a trench coat.
• First-sentence audit: read only the opening sentences, top to bottom. They should read like a compressed version of the essay.
• Evidence budget: one or two pieces per paragraph. Five quotes in a row is a citation dump, not a paragraph.
• Analysis floor: the explanation should run at least as long as the quotation it interprets.

Notice what the anatomy quietly forbids. A paragraph that opens by announcing what a source says has made the source the protagonist, and your argument has become a footnote to it. Flip the order: your claim first, then the source as hired support. The difference sounds cosmetic and grades like it isn't — reversed paragraphs read as summary even when every sentence in them is technically accurate. The Essay Structure Builder scaffolds this anatomy for each body section, which makes the habit easier to keep than to break.

Evidence needs analysis, not applause

Dropping in a quotation is not using it. Compare two treatments of the same finding. Version one: 'Smith (2022) found that retention rose 14 percent among remote staff (p. 3). This shows remote work has benefits.' That second sentence is applause — it restates the quote and calls it proof. Version two: 'Smith's 14 percent retention gain among mid-career staff supports the first half of the claim, but her interviews complicate the rest: nine of twelve junior hires reported having no mentor since going remote.' Same evidence. The second version weighs it, connects it to the thesis, and notices the tension the raw number hides.

Two selection rules earn their keep. First, two strong pieces of evidence beat five weak ones; a single 2023 study from a named journal outweighs three undated listicles, and graders notice source quality within the first page. Second, match evidence age to claim shelf life — a 2012 statistic is fine for constitutional history and stale for anything involving AI tools, social platforms or post-pandemic work. And when the evidence cuts against you, use it anyway. A paragraph that acknowledges the strongest objection and answers it earns more credibility than three paragraphs of agreement ever will.

Transitions that argue

Transitions are logic joints, not sequencers. 'Firstly, secondly, in conclusion' tells the reader where the paragraphs sit in line; it says nothing about how they connect. Compare 'Firstly, remote work saves money' with 'The savings argument, however, assumes turnover is the only cost that matters.' The first is a queue number. The second carries the argument's direction and quietly previews the objection the next paragraph will handle. The rule worth internalizing: the first sentence of a paragraph should need the last sentence of the previous one, and when it doesn't, the transition is decoration.

• Concession: 'Admittedly, the pilot program was small — yet the districts that copied it saw the same pattern.'
• Escalation: 'The retention numbers, though, tell only half the story.'
• Contrast: 'Where the 2019 policy failed quietly, the 2023 revision failed loudly.'
• Consequence: 'The result was a budget that funded signage but not staff.'
• Correction: 'That reading assumes the survey measured intent; the wording measured satisfaction.'

Run the deletion test before submitting. Delete every 'firstly,' 'secondly' and 'moreover' in the draft; if the essay still reads cleanly, your transitions were doing nothing and the paragraph openings are doing the real work — keep it that way. If the essay suddenly jumps, the joints were masking an order problem, and no transition phrase will fix an argument whose steps do not follow. That is a structural revision problem, and it belongs in pass one, not in the final hour.

Introductions last

Write the introduction after the body exists. An introduction is a promise, and a promise drafted before the argument does not yet know what it is promising — which is why introductions written first describe a different essay than the one that got written. The recipe for a course essay is short: one or two sentences of context that a classmate would recognize as relevant, one sentence naming the problem or debate, then the thesis. That is 100-150 words for a 1,500-word essay. Skip the dictionary-definition opener ('Webster's defines justice as...') and the 'since the dawn of time' opener; graders have read both several thousand times.

The conclusion deserves the same respect, and it is not a copy-paste of the introduction with synonyms. Two moves, roughly 120-150 words: restate the argument in fresh words — prove you still believe the version the body actually defended — and answer the 'so what' question with one concrete implication, limit, or next question. No new evidence belongs there; a quotation appearing for the first time on the last page reads as an argument that ran out of road. End on the stakes, not on a summary of the summary.

A self-editing checklist

Editing works when it is a sequence of small, boring checks rather than one mood called 'proofreading.' Budget real time for it: of the five or six hours a typical 1,500-word essay takes, one full hour belongs to editing, and skipping that hour is the most common source of lost mechanical points. Our three-pass editing guide covers the full method; the checklist below compresses it into the checks that catch the most frequent losses in course essays.

• Reverse-outline the draft: read the first sentence of every paragraph in order and confirm they argue, not summarize.
• One-claim test: split any paragraph holding two points, even if splitting feels wasteful.
• Quote ratio: analysis must run at least as long as the quotation it explains.
• Cut 10 percent: adverbs, 'very,' 'really,' and throat-clearers like 'it could be argued that' go first.
• Word Counter: stay within about 10 percent of the assigned length, and treat being 300 words under as under-argued, not under-typed.
• Citations: every in-text entry appears in the reference list and vice versa — format with the Citation Generator, then verify two entries by hand against MLA 9 or APA 7.

When the essay is due tomorrow

Panic narrows options; a triage order keeps the panic useful. Reread the prompt and underline the verbs — 'analyze,' 'compare' and 'evaluate' demand different essays, and misreading the verb is the most expensive mistake available at this stage. Pick two points, not five. Gather six to eight sources at most; beyond that, reading displaces writing. Then draft in the order that earns the most points per hour: thesis, two body paragraphs, conclusion, introduction last — the introduction is fast to write once the essay exists, and slow to write well before it does. At 300-500 words per focused hour, a 1,200-word draft is three to four hours.

Cut breadth, never integrity. A shorter essay that argues one point cleanly outscores a padded one that gestures at five — graders reward the rubric line that says 'develops a claim,' and padding fails it visibly. What you must not cut under time pressure is citation accuracy, because missing or invented citations are the one error that turns a B- paper into an academic-integrity case. Use the Word Counter to size the draft honestly and the Readability Checker to untangle the worst sentences, then submit at 97 percent quality rather than never. A finished B- beats an imaginary A+.