27  How to Write a Thesis

TipPrerequisites and see-also

Prerequisites (read first if unfamiliar): Chapter 26, Chapter 30.

See also: Chapter 25, Chapter 28, Chapter 31, Chapter 29.

Purpose

It’s the start of senior year, and the professor you finally worked up the nerve to email about an honors thesis writes back with one line: “Great, what do you want to work on?” Or you’re a master’s student with a data set, a deadline in May, and no idea what a thesis is supposed to look like. Or you’re a PhD student with three papers in various states of finish, and someone has just told you they need to become one dissertation. Each comes with the same quiet panic: you’ve written plenty of papers, and none of them prepared you for this.

That feeling is reasonable, because a thesis really is a different kind of object. It’s the longest single argument most people ever write, it’s yours in a way a co-authored paper isn’t, and it’s too big to hold in your head at once. So the hard parts are less about sentences and more about scale and people: sizing a topic you can finish, outlining at several levels, revising without losing the thread, working with an advisor and a committee who don’t always agree, and actually stopping.

This chapter covers that craft for an honors or senior thesis, a master’s thesis, or a PhD dissertation. It doesn’t cover the substance of your research; that’s the work of your advisor, your committee, and your reading. It describes common U.S. practice in general terms; rules differ between universities and even departments, so your program’s handbook and thesis guide have the final word. Read it alongside Chapter 26: a thesis is a bigger room with its own acoustics, and some of what works for a paper doesn’t scale.

Why read this chapter

• You’ve been told to pick a thesis topic by October, and every idea you have feels either too big to finish or too small to matter.
• Your advisor takes three weeks to answer an email, and you’re not sure whether you’re allowed to ask for more of their time.
• Two committee members gave you opposite advice on the same chapter, and you don’t know whose to follow.
• You’ve rewritten the first page of chapter one five times and still haven’t started the chapter the whole thesis depends on.
• You have three papers that are supposed to become one dissertation, and stapled together they read like three strangers.
• Your defense is on the calendar, and you have no idea what actually happens in the room, or what happens if it goes badly.
• People keep saying “prospectus,” “comps,” “quals,” and “deposit,” and nobody has told you which of them apply to you.

Running theme: a thesis is one argument, told once, at length, in your own voice

Even a dissertation built from three published papers isn’t a stack of papers: it’s one argument that the introduction names, the chapters develop, and the conclusion pulls together.

27.1 What a thesis is for

A thesis has three audiences, and they want different things. The committee (for many undergraduate theses, your advisor and a second reader) is the gatekeeper. They’ll read the whole thing, sign off on it, and sit in the room if there’s a defense, and they want to be persuaded that you’ve done a coherent piece of work you can stand behind. Your future self is the second audience: in five or ten years you’ll open this document to remember what you actually did and what you’d do differently, and it’s the only record of the project written from the inside. Your future readers are the third: colleagues, students, hiring committees, someone who found your thesis in a library search. Most of them will read the introduction and maybe one chapter.

So the committee wants completeness, your future self wants an honest record, and your readers want compression. A workable compromise is to write the introduction and conclusion for compression, the chapters for completeness, and the appendices for honesty (the failed analyses, the full survey instrument, the code).

27.2 Kinds of thesis

The vocabulary alone trips people up. At most U.S. universities you write a thesis for a bachelor’s or master’s degree and a dissertation for a doctorate; in the UK the words roughly flip, and Wikipedia’s thesis article collects the variations country by country. This chapter says “thesis” for all of them. Four forms cover most of what you’ll meet, and the page counts below are rough, since they vary a lot by field and program.

Honors or senior thesis. A focused project, usually written over one or two semesters of your final year with a faculty advisor and often a second reader. It usually isn’t held to a dissertation’s standard of a new contribution to the field (though the best ones make one), but it does have to ask a real question, use real sources or data, and argue carefully. It’s often the first time you’ll sustain one argument across dozens of pages, which is good practice if graduate school is on your mind.

Master’s thesis. Often fifty to a hundred pages, with one or two research questions and one coherent contribution. It’s often a stepping stone to a PhD or to research work in industry, and it’s shaped more like a long paper than a book, with a bigger literature review than a paper would carry.

Monograph dissertation. Often a hundred and fifty to three hundred pages of sustained argument, a monograph in all but name, and common in the humanities and much of the social sciences. Its strength is depth: one big idea developed across chapters. The risk is that no chapter ever becomes a standalone publication, and years of work get read by almost no one outside your committee.

Three-paper (or article-based) dissertation. Several publishable papers (most often three in U.S. programs) plus an introduction and conclusion that frame them as one contribution. It’s the standard thesis as a collection of articles in some countries and fields, and common in information science, HCI, and computational social science. You finish with papers already out or under review; the price is coherence, because if the papers are too different, the introduction has to work very hard to make them one project.

Three large empirical studies fit the paper-based form; an argument that builds chapter by chapter fits the monograph. Settle the form with your advisor early, within your program’s rules, because switching halfway through is painful.

27.3 Choosing a topic you can finish

Picking a topic is where many theses stall before they start. Every idea seems to land in one of two bins: so big it would take a career, or so small it can’t possibly be worth a year. Almost everyone bounces between those bins for a while. The way out is to stop asking “is this important?” and start asking “can I answer this with the time and data I have?”

A thesis topic is really a research question sized to a calendar, and a good test is whether you can state it in one sentence that names what you’ll study and how. “Social media and democracy” is an area. “Did one large subreddit’s 2023 moderation policy change reduce reported harassment over the next six months?” is a question. It sounds smaller, and that’s the point: a finished narrow question is worth far more than an unfinished broad one.

Workable topics usually grow from what you already have: a course paper you wished you could have taken further, a data set you already know how to clean, or the data, methods, and access your advisor’s group already has. Talk to two or three faculty members before you commit; they know which questions are live and which will eat a year of data collection before you write a word. Expect the topic to narrow as you go, which is healthy. The opposite, scope creep, is what keeps a project from ending, so when a promising tangent shows up, put it in a “future work” file and keep going.

27.4 The proposal

Most U.S. PhD programs, many master’s programs, and some honors programs ask for a written proposal (sometimes called a prospectus) before the main work begins. It can run from a few pages for an honors thesis to thirty, sixty, or more for a dissertation, and it usually covers a literature review, the research questions, the methods, a timeline, an ethics plan (such as approval from your university’s institutional review board if you’ll study people), and the contribution you expect.

Think of the proposal as your introduction, written before you’ve done the work. Plenty of its sentences won’t survive, but writing it forces you to say out loud what you don’t know yet, and if you can’t state your research question in one sentence at proposal time, the next year of work will drift.

Pat Thomson’s blog has thoughtful posts on proposals as a kind of writing.¹ Joan Bolker’s Writing Your Dissertation in Fifteen Minutes a Day is the standard companion on the writing side,² and Wendy Belcher’s twelve-week framework (cited in Chapter 26) carries over to thesis-scale writing once you scale the units up.

¹ For example, Pat Thomson, “the research proposal as writing ‘work’” (2014).

² Bolker’s book is listed in Further reading at the end of this chapter.

27.5 Your advisor and your committee

A thesis is the first writing project where you manage people as much as prose, and nobody teaches you how. Your advisor (also called a supervisor, chair, or doctoral advisor) is the person you meet with most, who reads your drafts most closely, steers the project through each milestone, and backs you up when the committee disagrees. The other committee members are often two to four more faculty for a doctorate, fewer for a master’s, and for many honors theses a single second reader. Some programs require one member from outside your department, to check that the work makes sense beyond your field. Their attention comes in bursts, mostly at the proposal and near the defense.

When your advisor seems to disappear. Perhaps the most common thesis frustration is an advisor who is kind, brilliant, and nearly impossible to reach. It usually isn’t about you: faculty juggle teaching, grants, and a dozen other students, and your thesis gets what’s left. You can make that attention easier to give. Ask for a standing meeting (every two weeks is common) instead of scheduling each one from scratch. Bring a short agenda and a specific question (“Should chapter 3 come before chapter 4?” rather than “What do you think?”). Send a summary email after each meeting listing what you agreed (there’s a template below), give each draft a date you’d like comments by, and find other readers too: a postdoc, another faculty member, or a writing group. If the relationship is truly broken, not just busy, talk to your program’s graduate or honors director; many universities also have an ombuds office for confidential advice.

When feedback stings, read it charitably, and not right away if you can help it. A comment that feels harsh on a Tuesday morning often reads as entirely reasonable on Wednesday afternoon.

When two committee members disagree, say one asks you to cut the section another asked you to add, don’t quietly pick one and ignore the other; that surfaces at the defense, in public. Take both comments to your chair, in writing and side by side, and ask how to resolve them. Mediating is part of the chair’s job.

27.6 Comprehensive and qualifying exams

Most U.S. PhD programs put some kind of exam between coursework and the proposal, called comprehensive exams (“comps”), qualifying exams (“quals”), or preliminary exams (“prelims”). Undergraduate and most master’s theses don’t have one. Formats vary enormously (timed written exams, oral exams, paper portfolios, reading lists with synthesis essays), so ask your program coordinator for the local rules. The advice that travels: start earlier than you think you need to, treat the reading list as a literature map (see Chapter 25), and ask students a year or two ahead of you for their actual study materials. Most will happily share.

27.7 Outlining and drafting at thesis scale

Here’s where many people freeze: you’ve read a lot, you have results, and when you sit down to write, the thesis is too big to see all at once. The fix is to outline at three levels and keep all three alive.

At the chapter level, keep a one-page table of contents with each chapter’s title, one sentence on what it does for the argument, and a rough target length. It’s your map, and you’ll redraw it many times.

At the section level, outline each chapter’s sections in order. This is where a chapter’s argument becomes visible, and where you discover, sometimes painfully, that two sections are saying the same thing.

At the paragraph level, for the chapter you’re drafting now, list the paragraphs as topic-sentence stubs. A good writing day turns a few stubs into paragraphs.

Then draft ugly. Anne Lamott’s famous case for terrible first drafts, in her book Bird by Bird, applies here more than anywhere. The most common way a thesis stalls is polishing the opening of chapter one for two months and never writing chapter four; the introduction is the last thing to finish, because you don’t know what it introduces until the rest exists. The second is writing in bursts and then vanishing for weeks. Steady output beats heroic effort (the whole argument of Bolker’s and Belcher’s books), so set a small daily target, even Bolker’s fifteen minutes, and block it on your calendar like a meeting you don’t move.

Two habits save grief later. Keep every source in a reference manager such as Zotero from the first week, so a two-hundred-page bibliography builds itself (Chapter 25 shows the setup). And keep the thesis under version control, one chapter per file (Chapter 31).

27.8 Integrating papers into a dissertation

For a paper-based dissertation, integration is the work that turns three papers into one thesis, and it’s easy to underestimate.

The introduction does the framing. It shouldn’t say “here are three papers I wrote”; it should say “here is one research question, and these three studies approach it from three angles.” It’s what most readers will read, and it’s what makes the thesis feel like a thesis, so spend real time on it.

Each paper chapter usually gets a preface of a paragraph or two: where it was published (or is under review), how it relates to the other chapters, and what the reader should know coming in. The body is usually the paper itself, with figure numbers, cross-references, and notation made consistent across the thesis. Check two rules early, because they catch people late: many programs ask you to state your own contribution to any co-authored paper, and for a published paper you should check the copyright agreement you signed (many publishers explicitly allow reuse in a thesis, sometimes with a required citation).

The conclusion synthesizes. What did you learn that no single paper could show? What does the dissertation contribute as a whole? What comes next? A monograph handles all this differently: its chapters run continuously, with no paper-shaped seams to hide, though its introduction and conclusion still do the framing.

27.9 Revision and feedback cycles

The path from draft to defense has a typical shape: your advisor comments on chapters and you revise; the committee reads the proposal and later the full draft, and you revise again; you defend; you revise one last time. Each cycle tends to be shorter than the last, from months of back-and-forth on your first chapter to days or weeks of corrections after the defense. Plan for that, and assume everything will take longer than your plan says; the planning fallacy is a documented bias, not a personal failing.

Readers will ask “what changed since the version I read?”, so keep track. If your committee works in Word, Track Changes is the common language. If you write in LaTeX (see Chapter 29), latexdiff compares two versions of a .tex file and writes a third that marks every deletion and addition; compile it as usual to get a marked-up PDF:

latexdiff ch3-sent.tex ch3-now.tex > ch3-diff.tex

If the thesis is in Git, tag each version you send out so you can compare against it later; --word-diff shows changes word by word, which suits prose far better than Git’s usual line-by-line view:

git tag -a sent-to-committee -m "Full draft sent to the committee"
# ...weeks of revision later...
git diff --word-diff sent-to-committee -- chapters/ch3.md

Address most comments without argument. For the few you push back on, reply in writing, even if the conversation happened in person, so there’s a record of what you decided and why.

The draft that won’t end. At some point many theses stop converging: each round of revision turns up two more things to fix, and “almost done” stretches across a semester. Often nobody has said what “done” means. Ask your advisor directly, “What has to be true before I can schedule the defense?”, write the answer down as a short list, and work the list. New ideas go in the future-work file, not the draft.

27.10 The defense

Many students picture the defense as a trial. In most U.S. programs it’s closer to a structured conversation about work the committee has already read, and your advisor usually won’t let you schedule it until they believe the thesis is ready. The thesis examination looks different elsewhere: in the UK and Ireland it’s the viva, usually a private exam with two examiners (one from another university) that can last a few hours. And plenty of honors theses have no oral defense at all, or end with a short presentation to the advisor and second reader, so ask early which kind yours is.

A U.S. graduate defense usually opens with a talk, often thirty to forty-five minutes (Chapter 28 covers defense talks among its genres). The questions that follow are the real substance and often run longer than the talk. Prepare for the questions you’d ask if you were on the committee, and especially for the one you most fear, because it usually gets asked. “I don’t know, but here’s how I’d find out” beats bluffing.

Afterward, the committee sends you out, talks briefly, and calls you back with the outcome: a pass with no changes, a pass with minor or major revisions, or (rarely) a requirement to revise and defend again. Passing with revisions is very common, so don’t hear “revisions” as failure. They usually come with a deadline and one person, often your advisor, who checks them.

27.11 Finishing

“Done is better than perfect” is the hardest advice in this chapter to take. The thesis isn’t the last thing you’ll write; finishing it is a milestone, not a verdict on your worth. Near the end, many people hit a stretch where the whole project suddenly looks obvious, boring, or wrong. Inger Mewburn of the Thesis Whisperer calls it a valley (her post’s title is blunter), and her point is that you’re passing through it, not stuck there.³ The sense that you’re the only one faking it has a name, impostor syndrome, and saying so to someone you trust helps.

³ Inger Mewburn, “The Valley of Shit”, The Thesis Whisperer (2012).

⁴ https://www.proquest.com/products-services/dissertations/

Then there’s the paperwork, which bites people who leave it to the last week. Your graduate school or honors program will have rules for margins, fonts, and front matter, and many run a format review before accepting the final file. If yours publishes a Word or LaTeX template, write in it from the start (Overleaf keeps a gallery of university thesis templates, but your own school’s version is the one that counts). Doctoral dissertations at many U.S. universities also go to ProQuest’s dissertation database,⁴ and most universities post theses in their institutional repository; if you’re publishing chapters as papers or a book, ask your advisor about an embargo that delays public release. None of this is intellectual work, and all of it has deadlines, so read your formatting guide the first time long before you need it.

27.12 Stakes and politics

Open a few theses in your department’s repository and read the acknowledgments. You’ll often find a partner thanked for covering the rent in a final, unfunded year, a parent for watching the kids on writing days, or a fellowship for years of support. The acknowledgments are the one page of a thesis that admits it ran on time, and that someone paid for the time.

A funded PhD is a part-time wage for full-time scholarly work, set by the institution rather than by the cost of living in its city. Students with caregiving responsibilities, debt, visa rules that limit outside work, or chronic illness face a different calculation from those without, so “just take an extra year” isn’t neutral advice. The undergraduate version is quieter: a year of unpaid research fits more easily around a life without twenty hours a week of paid work. Advising is also a power relationship. Your advisor signs off on every milestone, writes your letters, and shapes the network you leave with; when the relationship fails through neglect, mismatch, or harassment, the remedy is often starting over, which is part of why “all but dissertation” is a familiar phrase. Surveys report anxiety and depression among graduate students at rates well above the general population (see Further reading). Even your citations take a position: whose work you build on, and whom you thank, shape who counts as the field (see Chapter 26).

See Chapter 8 for the broader framework. The concrete prompt to carry forward: when you plan a thesis, plan also for the conditions that make finishing possible (money, time, and people), and notice that those conditions aren’t handed out evenly.

27.13 Worked examples

A three-paper dissertation outline

Here’s a hypothetical dissertation, “Governance on Volunteer-Run Online Platforms,” built from three papers plus a framing introduction and conclusion.

Introduction (8,000–10,000 words). Frames the broad question: how do volunteer moderators on platforms like Reddit, Wikipedia, and Discord govern user behavior, and what changes when a platform shifts policy from above? Reviews the literature on online governance, content moderation, and volunteer labor, sets up the three angles, and names the three contributions.

Chapter 2 (Paper 1, published at CSCW 2025). A natural-experiment study of one subreddit before and after a moderation policy change, using difference-in-differences with a matched comparison community.

Chapter 3 (Paper 2, under review at New Media & Society). An interview study of twenty volunteer moderators across five platforms, using semi-structured interviews and reflexive thematic analysis.

Chapter 4 (Paper 3, in preparation). A computational analysis of the rules of a thousand subreddits, asking how rule complexity relates to community size and moderator turnover.

Conclusion (3,000–5,000 words). Argues that volunteer moderation shows a pattern of escalation and burnout that quantitative analysis alone can document but not explain, and interviews alone can describe but not generalize, so the three approaches together produce an account none could alone. Then future work.

Each chapter keeps its standalone form, with a one-paragraph preface noting its publication status; the introduction and conclusion are the connective tissue that makes them one argument.

A 12-month proposal-to-defense timeline

Proposals often show the timeline as a Gantt chart; a table by month does the same job:

Month	Milestone
1	Draft proposal outline; share with advisor.
2	Draft proposal §1 (intro/literature).
3	Draft proposal §2 (method/timeline).
4	Circulate full proposal to committee; schedule proposal defense.
5	Proposal defense; revise per feedback. IRB submission if needed.
6–9	Data collection / analysis / Paper 3 work.
10	Draft chapter 4 (Paper 3); update introduction.
11	Circulate full dissertation to committee; schedule defense.
12	Defense; final revisions; deposit.

This is aggressive; most students take longer, and that’s normal. The shape is what matters: proposal early, data work in the long middle, defense and deposit at the end. A two-semester honors thesis has the same shape in miniature, so work backward from your program’s published deadline, leaving room for the revisions your advisor will ask for.

An advisor-feedback log entry

Your advisor has commented on your methods section: “I’m not convinced the interview sample is balanced. Why only women moderators?” You disagree: the focus is deliberate and theoretically motivated. Your job isn’t to win the argument; it’s to record the decision and your reasons, calmly, in writing.

Advisor feedback (2026-03-12): “Why only women moderators?”

My response: The all-women sample is intentional, for two reasons. First, the studies of volunteer moderation I build on in §2.3 drew their samples mostly from men, and this study is designed to address that gap. Second, my interviews surfaced gendered patterns of harassment and emotional labor that are central to the contribution; a mixed sample would have diluted that analysis. I will revise §3.2 to make this rationale explicit. Discussed with advisor 2026-03-15: agreed on revised framing.

When the committee, or a reviewer of the published paper, raises the same question later, the answer is ready, and you won’t have to reconstruct it from memory.

27.14 Templates

A proposal outline:

# {Working dissertation title}

**Advisor:** {name}
**Committee:** {names}
**Target proposal defense:** {YYYY-MM-DD}
**Target final defense:** {YYYY-MM-DD}

## §1 Introduction (~3000 words)
- The phenomenon
- The gap
- The research questions
- Roadmap

## §2 Literature review (~6000 words)
- Theme A: ...
- Theme B: ...
- Theme C: ...
- Synthesis: the gap

## §3 Method (~4000 words)
- Study 1: ...
- Study 2: ...
- Study 3: ...

## §4 Timeline (~500 words + Gantt table)
...

## §5 Expected contribution and limitations (~1000 words)
...

A chapter outline at three levels (paste at the top of chapter-N.md):

# Chapter N: {title}

## Section 1: {section title}
- Paragraph 1: {topic-sentence stub}
- Paragraph 2: {topic-sentence stub}
- ...

## Section 2: {section title}
- Paragraph 1: {topic-sentence stub}
- ...

A committee-feedback tracker:

| Date | Source | Comment | Decision | Status |
|---|---|---|---|---|
| 2026-03-12 | Advisor | Sample skew? | Defended in §3.2 | Resolved |
| 2026-03-19 | R. Smith | Add formal analysis | Adding Appendix B | In progress |
| ... | ... | ... | ... | ... |

A meeting summary to send your advisor the same day:

Subject: Summary of our meeting, {date}

Thanks for meeting today. Here's what I have down:

- Decided: {what we agreed, e.g., chapter 3 moves before chapter 4}
- I'll do next: {tasks}, by {date}
- Open question for next time: {question}
- Next meeting: {date}

Please let me know if I've misremembered anything.

27.15 Exercises

1. Write a 250-word “contribution statement” for a thesis project of your own (real or hypothetical): the question, the approach, and the contribution you expect.
2. Take the broadest topic you care about and narrow it, in three steps, to a one-sentence research question you could answer in the time you have.
3. Find a recent thesis or dissertation in your subfield in your university’s institutional repository and reverse-engineer its outline at chapter and section level. Which kind of thesis is it? How does the introduction frame the chapters?
4. Draft a short email inviting a faculty member to be on your committee or to be your second reader: one paragraph on the project, one on why you’re asking them.
5. Build a timeline for your own (real or hypothetical) thesis, working backward from your program’s deadline. Mark the proposal, the data-collection windows, the full draft, and the defense.
6. Write the five questions you’d most dread hearing at your defense, and draft a two-sentence answer to each.

27.16 One-page checklist

• Did you choose your thesis’s form deliberately, with your advisor, and within your program’s rules?
• Can you state your research question in one sentence?
• Do tangents go in a “future work” file instead of the draft?
• Has each committee member (or your second reader) confirmed they can serve?
• Do you have a standing meeting with your advisor, and send a summary after each one?
• Are you outlining at three levels (chapter, section, paragraph)?
• Is your daily writing time blocked and protected?
• Are your sources in a reference manager and your chapters under version control (see Chapter 31)?
• Are committee comments and your decisions tracked in writing?
• Do you and your advisor agree, in writing, on what “ready to defend” means?
• Did you read your program’s formatting guide and deadlines before the final-week scramble?

27.17 Quick reference: thesis forms at a glance

Page counts are rough and vary a lot by field and program.

Form	Typical shape	Strength and risk
Honors or senior thesis	One focused project over one or two semesters; length set by your program	Strength: a real research experience. Risk: a topic too big for the time you have.
Master’s thesis	Often 50–100 pages; one project	Strength: a stepping stone. Risk: compressed scope.
Monograph dissertation	Often 150–300 pages; one sustained argument	Strength: depth. Risk: chapters that never get published.
Paper-based dissertation	Often 150–250 pages; three papers plus framing	Strength: publications along the way. Risk: coherence.

Note📚 Further reading

• Joan Bolker, Writing Your Dissertation in Fifteen Minutes a Day (Henry Holt, 1998) — short, durable advice on building a daily writing habit; the title is the method.
• Peg Boyle Single, Demystifying Dissertation Writing (Stylus, 2009) — a research-backed method for choosing a topic, outlining, drafting, and revising at thesis scale.
• Pat Thomson, patter, thesis category — long-running blog of practical advice from a PhD supervisor; especially good on revision and working with supervisors.
• Inger Mewburn, The Thesis Whisperer — the most candid public voice on the emotional reality of doing, and finishing, a research degree.
• Wendy Laura Belcher, Writing Your Journal Article in Twelve Weeks (2nd ed., 2019) — a companion for the paper-based route, since each paper goes through the same workbook.
• Karen Kelsky, The Professor Is In — practical advice on the academic job market, advising relationships, and finishing; useful when “what comes after the thesis?” needs planning.
• Teresa Evans et al., Evidence for a mental health crisis in graduate education (Nature Biotechnology, 2018) — a widely cited survey finding high rates of anxiety and depression among graduate students; the evidence behind the “Stakes and politics” section above.