Lesson 15 of 30 · Structure the paper editors accept

Build the skeleton before the prose

Editors and reviewers read structure before they read sentences. A paper that is well written but badly built still gets rejected, because the argument does not hold.

The fix is to construct the skeleton first, in a fixed order that mirrors how the reader evaluates you: the one figure the paper is built around, then the results that figure supports, then the methods that produced them, then an introduction that funnels from the field to your specific question, then a discussion that answers "so what", and the abstract and title last, once you know what the paper actually says. Writing in that order (results-first, intro-and-abstract-last) is the most reliable way to avoid the most common structural failure: an introduction promising a paper the results do not deliver.

A skeleton is a page of headings with one sentence under each. If a section cannot hold its one sentence, the problem is the science or the framing, and it is far cheaper to find that here than in a full draft.

Your move

Write the skeleton of your paper: a heading for each section (or planned section), and one sentence under each stating what it will establish. Start with the central figure and the results; leave the intro and abstract sentences for last. Mark any section whose one sentence you cannot yet write.

Why: The structural failures that sink papers are visible in a one-page skeleton and invisible in a full draft you are too close to. Building it now is the cheapest structural edit you will ever make.

Worked example

Outline the paper as a list of claims, one per section, before writing a sentence of prose. If the claims do not form an argument in the outline, no amount of good writing will make them one on the page.

The pattern that loses papers here

Writing paragraphs before the argument exists, then trying to reverse-engineer a structure from prose you are now attached to. The seams always show.

Check it against your draft

Each section maps to one claim
The claims read as an argument in order
Results lead; methods support, not the reverse

Quick check

2 questions, to catch the misconceptions this lesson exists to fix.

1.You should build the paper:
2.A section whose one-sentence summary you cannot write is a problem to fix now.

Write your answer and keep it

Reading is free and always will be. An account is what lets you write your own answer to this move, get it read back to you, and keep everything you build as you go.