How to Write a Good Research Paper

This note keeps the original lecture structure and content, while lightly polishing the language and layout to make it easier to read as a compact academic-writing guide.

A good paper is rarely the result of strong results alone. More often, it succeeds because the author knows how to identify a central idea, present it as a coherent story, support it with evidence, and make the argument easy to follow. The following seven suggestions summarize that discipline.

Seven simple, actionable suggestions

1. Do not wait: write

• Not: your idea -> do research -> write paper
• Better: your idea -> write paper -> do research
• Writing forces you to become clear, focused, and concrete about what you do not yet understand.
• Writing opens a path to dialogue with others: criticism, reality checks, disagreement, and collaboration.
• Writing is not merely the final report of research. It is one of the primary ways research gets done.

2. Identify the core idea

• A worthwhile paper usually revolves around one idea that is both interesting and defensible, even if it appears small at first.
• Writing a paper is one of the earliest ways to make an idea real, so the paper should center on a single, sharp point.
• You may not know exactly when to begin writing, but you must know when to stop. If you truly have many ideas, write many papers.
• Many papers contain good material but fail to distill the main point. The reader should never have to guess what the paper is really saying.

3. Tell a story

• Imagine explaining the work at a whiteboard: here is the problem, here is why it matters, here is what remains unsolved, here is my idea, here is why it works, and here is how it compares with alternative approaches.

A useful conference-paper structure often looks like this:

• Title (1000 readers)
• Abstract (4 sentences, 100 readers)
• Introduction (1 page, 100 readers)
• The problem (1 page, 10 readers)
• My idea (2 pages, 10 readers)
• The details (5 pages, 3 readers)
• Related work (1-2 pages, 10 readers)
• Conclusions and further work (0.5 pages)

4. Nail your contributions to the mast

Make the introduction do real work

• State the problem plainly.
• Explain why it matters.
• Make the central contribution visible early.

State the contribution explicitly

Do not make the reader infer your contribution from scattered hints. List the key contributions clearly and organize the paper around proving them.

A useful contrast between weaker and stronger framing is shown below.

Support each claim with evidence

• The introduction should state the claims.
• The body should provide evidence for each claim.
• Review every claim in the introduction and ask: where is the evidence?
• Evidence may take the form of analysis, comparison, theorem, measurement, or case study.

Avoid the empty roadmap paragraph

• Avoid: “The rest of this paper is structured as follows. Section 2 introduces the problem. Section 3 … Finally, Section 8 concludes.”
• Prefer: forward references that are embedded naturally in the introduction’s argument.

The introduction should already survey the whole paper, and it should do so as part of the narrative rather than as a mechanical table of contents.

5. Put related work in the right place

Prefer this order

Avoid this structure

• Abstract (4 sentences)
• Introduction (1 page)
• Related work
• The problem (1 page)
• My idea (2 pages)
• The details (5 pages)
• Conclusions and further work (0.5 pages)

Prefer this structure

• Abstract (4 sentences)
• Introduction (1 page)
• The problem (1 page)
• My idea (2 pages)
• The details (5 pages)
• Related work (1-2 pages)
• Conclusions and further work (0.5 pages)

No related work yet!

This earlier placement causes two immediate problems:

1. The reader does not yet understand the problem, so compressed discussion of technical trade-offs becomes almost impossible to follow.
2. Alternative approaches are placed between the reader and your own idea.

Give credit generously

Fallacy: To make my work look good, I must make other people’s work look bad.

Truth: Credit is not a scarce resource.

• Warmly acknowledge people who helped you.
• Be generous to competing work. A fair sentence such as “Foogle’s paper lays an inspiring foundation, which we extend in the following ways” is stronger than petty dismissal.
• Acknowledge the weaknesses of your own approach when necessary.
• Giving credit to others does not reduce the credit your own paper receives.

6. Put the reader first

A paper usually moves through three broad stages:

• The problem (1 page)
• My idea (2 pages)
• The details (5 pages)

A technically impressive explanation is still a failure if it puts the reader to sleep or makes the reader feel foolish.

Present the intuition before the machinery

• Explain the idea as if you were speaking to someone at a whiteboard.
• Intuition is primary, not secondary.
• Once the reader has the intuition, the details become followable.
• If the reader skips the details, they should still carry away something valuable.

Introduce the problem and the idea through examples first. Only then move to the general case.

Choose the shortest path to understanding

• Do not recapitulate your personal journey of discovery.
• The fact that the road was painful for you is not, by itself, interesting to the reader.
• Choose the most direct route to the idea.

7. Listen to your readers

Get the paper read early

• Ask as many friendly readers as possible to read the draft.
• Experts are valuable.
• Non-experts are also valuable.
• Every reader can only read your paper for the first time once, so use that first reading carefully.
• Ask for the right kind of feedback. “I got lost here” matters more than “you misspelled a word here.”

Ask experts for help

• A useful strategy is to send the draft to a competitor and ask whether you have represented their work fairly.
• Often they respond with criticism that is both serious and useful.
• They may well be your reviewers anyway, so early criticism is a gift.

Read criticism as diagnostic data

• Read every criticism as a suggestion about what could be explained more clearly.
• Do not respond with: “You misunderstood me. I meant X.”
• Instead, revise the paper so that X becomes obvious even to the least prepared reader.
• Thank reviewers warmly. They gave their time to your work.

Summary

• Write early.
• Identify the key idea.
• Tell a story.
• State the contribution explicitly.
• Put related work later, not sooner.
• Put the reader first, especially through examples and intuition.
• Listen carefully to readers and reviewers.

Language and style

Basic discipline

• Submit by the deadline.
• Respect the length restrictions. Do not narrow the margins. Do not shrink the font. If necessary, place supporting material such as proofs or experimental detail in an appendix.
• Always use a spell checker.

Build visible structure

• Give the paper a strong visual structure through sections, subsections, bullets, italics, and well-laid-out code.
• Learn to draw figures, and use them wherever a picture makes the argument clearer.

Use the active voice

The passive voice can sound respectable, but it often drains energy from the prose. Avoid it whenever possible.

Use simple, direct language

Use consistent terminology

Write with clear logic

Respect de facto standards