Where papers actually lose points

The first college paper I got back had a lot of red on it, and almost none of the red was about sentences.

The comment at the top, which I have thought about many times since: "You've told me a great deal about this. I still don't know what you think."

I'd written eight pages summarizing four sources. Accurate, organized, cited, on time. There was no argument in it anywhere.

The rubric is not shaped the way your editing pass is

Look at almost any college paper rubric and the point distribution goes roughly like this.

Argument and thesis on top. Evidence and how you use it right behind. Structure after that. Mechanics — grammar, formatting, citation style — is the smallest slice.

Now think about how a paper actually gets finished. The last four hours are spent on typos, comma splices, and getting the citations to look right. All of that effort is aimed at the smallest bar on the chart, because it's the only bar you can still move at 1 a.m.

Which is exactly the point: the big bars have to be settled days earlier. Not because starting early is virtuous, but because they're structurally unfixable at the end.

A thesis is a claim someone could argue with

This is the thing American writing instruction hammers on, and it took me a while to understand why it matters rather than just being a rule.

"This paper examines the role of railroads in westward expansion" is a topic. It announces a subject. It commits to nothing.

"Railroads changed settlement patterns in the American West" is an observation. True, uncontroversial, and it gives the paper nothing to do.

"Railroad land grants, not the Homestead Act, determined where western towns survived" is a thesis, because someone who read the same sources could disagree with it. That disagreement is what the rest of the paper is for.

The test is one question: could a reasonable person, having read the same material, argue the opposite? If no, you don't have a thesis yet, and every paragraph you write will be summary because there's nothing to defend.

The corollary is that a thesis is a decision, and decisions are uncomfortable, which is why people stall at the topic stage. Summarizing is safe. Claiming something is not. But the grade is mostly for the claim.

Evidence has to do work, not just appear

The second-largest slice, and the pattern that costs the most points is the drop-in quote: a quotation appears, and the next paragraph starts a new idea.

Every piece of evidence needs three moves around it.

Set it up. Whose is this, and why should we believe them for this particular claim?

Present it. Quote or paraphrase, whichever serves better. Long block quotes are usually a sign you haven't decided which part matters.

Do something with it. This is the missing move. Say what it shows, why it supports your claim, and — the part that separates a B from an A — what it doesn't settle.

That third move is where the thinking is visible. Without it a paper reads as a well-organized collection of other people's sentences, which is exactly what mine was.

Structure that follows from the argument

If the thesis is a claim, the structure is the case for it. Each section should be a reason the claim holds, or an objection you're answering.

A structure that follows the sources instead ("here's what Smith says, here's what Chen says") is a signal that the paper is organized around your reading rather than around your argument. Same information, much lower grade.

Two practical habits:

• Topic sentences that make claims, not ones that announce subjects. "Land grants concentrated capital along the route" beats "Another factor was land grants."
• Write the introduction last. You don't know what the paper argues until it's written. Writing the intro first means writing a promise you then spend eight pages failing to keep.

The mechanics slice, which is small but free

It's the smallest bar and it's the easiest points on the page, so don't skip it — just don't mistake it for the work.

The single most common deduction here isn't picking the wrong citation style. It's using two in one paper, because half the citations were pasted in from a database that formats them differently. Pick one, check the syllabus for which, and make the whole document consistent.

Which style is a question about your field, not your taste — APA in psychology and the social sciences, MLA in English and languages, Chicago in history, IEEE in engineering. But the syllabus overrides all of it, always.

The timeline that actually prevents the disaster

Not "start early." Specifically:

As soon as the prompt is out: read it, and write down what kind of paper is being asked for. Argue a position? Compare two things? Analyze one text closely? Getting this wrong is the most expensive possible error and it's completely avoidable.

A week out: have a draft thesis, even a bad one. One sentence. Then read against it — you're now reading for evidence rather than reading in general, which is dramatically faster.

Four days out: an outline that is a list of claims, not a list of topics.

Two days out: a full draft, however rough. The gap between "no draft" and "bad draft" is enormous; the gap between "bad draft" and "good draft" is an afternoon.

One day out: revise the argument. Read only your topic sentences in order — if they don't tell a coherent story on their own, the structure is wrong, and this is the last moment that's fixable.

Last few hours: mechanics. Now it's the right time for the typo pass.

One thing worth doing that almost nobody does

Take your draft to office hours, or to the campus writing center, which exists at basically every US school and is free and consistently underused.

Bring a specific question. Not "is this good." Something like "does my thesis actually get supported by section three, or am I just describing things there?" Specific questions get useful answers; open ones get "looks fine."

A twenty-minute conversation about the argument, four days before the deadline, is worth more than any number of hours of polishing afterward. It's aimed at the biggest bar on the chart, at the last moment when the bar can still move.