How to Write a Thesis: A Step-by-Step Guide

A thesis is an extended piece of original research written to earn a degree. In the US, thesis usually means a master's or honors project and dissertation means a doctoral one; in the UK and many other systems the terms are reversed. The writing process is the same either way, so this guide uses "thesis" for both. If you are unsure which your program expects, see thesis vs. dissertation.

The word "thesis" also names the one-sentence central claim of an essay or research paper. That is a different, much smaller task, covered in the section on writing a thesis for a research paper or essay below.

Whatever the length, examiners judge a thesis on a handful of questions:

• Is there a clear question? The reader should be able to state your research question after reading the first few pages.
• Does it know the field? The literature review shows what is known and exactly where your contribution sits.
• Does the method fit the question? A causal question needs a design that can support causal claims; an exploratory question needs flexibility.
• Do the conclusions follow from the evidence? Overclaiming is the most common reason for major corrections.
• Is it readable? Consistent terminology, signposting and correct formatting signal care.

Keep these five questions in view from day one. Every chapter exists to answer one or more of them.

The sequence below is how most successful theses actually get written. It is not strictly linear; you will loop back, especially between steps 3 and 6.

1. Choose a topic you can sustain. Pick an area you care about, that your supervisor can support, and where data or sources are realistically available to you.
2. Narrow it to a research question. Move from "social media and teenagers" to "Is daily time on short-video apps associated with self-reported sleep duration among 14–17-year-olds?" A good question names the population, the concept and the relationship.
3. Do a scoping read. Spend two to four weeks reading recent review articles and the most-cited studies. You are checking that the question is open, not already settled.
4. Write a proposal. Even if your program does not require one, a two-to-five page proposal (question, rationale, method, timeline) forces decisions early. See writing a research proposal.
5. Build your outline and timeline. Use the chapter template below and plan backward from the submission date.
6. Get ethics approval and collect data. If you work with human participants, approval can take weeks; apply as soon as your method is fixed.
7. Analyze as you go. Run preliminary analyses on early data so problems surface while you can still fix them.
8. Draft the chapters out of order. Methodology and results first (they describe what you did), then the literature review and discussion, then the introduction, conclusion and abstract.
9. Revise in passes. First for argument and structure, then for paragraph logic, then for sentences, then for formatting and references.
10. Proofread, format and submit. Check your department's formatting guide line by line, build the front matter, and leave a buffer of at least a week before the deadline.

Most people stall at the start because they try to write the introduction first. The introduction is the hardest chapter to write well because it depends on everything else. Start somewhere easier.

Write a one-paragraph research summary

Before any chapter, write five sentences: the problem, what is already known, the gap, your question, and how you will answer it. For example:

Small municipal libraries are expanding digital lending, but little is known about who uses it. Existing studies focus on large urban systems. This leaves rural patrons, who may face weaker broadband access, largely unstudied. This thesis asks whether broadband availability predicts e-book borrowing rates across rural library districts. It answers the question using five years of circulation records linked to public broadband coverage data.

That paragraph becomes the spine of your introduction and the basis of your abstract. Revise it whenever your project changes.

Turn your reading notes into a working outline

Group the sources you have read by the question each one helps answer. Those groups become the subsections of your literature review, and often reveal the gap more clearly than any amount of thinking about it.

Set a small, daily output target

A target of 300–500 words or one completed subsection per working day is sustainable for most writers. Rough drafts are expected; the goal at this stage is material you can revise, not polished prose.

The table below is a standard thesis outline for empirical research in the social, health and natural sciences. Humanities theses are usually organized by argument or theme rather than by these headings, but they still need an introduction, a review of scholarship, a stated approach and a conclusion. Proportions are typical guides, not rules.

Common variations

• Results and discussion combined: common in qualitative theses, where interpretation is woven into the presentation of themes.
• Thesis by publication: an introduction and conclusion frame two to four article-style chapters, each with its own methods and results.
• Humanities thesis: an introduction with the literature and approach, three to five argument-driven chapters, and a conclusion.

For detailed guidance on two of the hardest chapters, see writing a methodology chapter.

Here is an example thesis outline for the rural library project described above, written at the level of detail a supervisor would want to see before you start drafting.

Title: Broadband Access and Digital Borrowing in Rural Public Libraries

1. Introduction
   1.1 Growth of digital lending in public libraries
   1.2 The rural access problem
   1.3 Research questions
       RQ1: Does broadband availability predict e-book borrowing per capita?
       RQ2: Does the relationship differ by patron age profile?
   1.4 Significance for library funding decisions
   1.5 Structure of the thesis

2. Literature Review
   2.1 Digital lending: adoption and patron profiles
   2.2 The digital divide and rural broadband
   2.3 Libraries as access points
   2.4 Gap: no district-level study linking coverage to borrowing

3. Methodology
   3.1 Design: longitudinal panel of library districts
   3.2 Data sources and linkage
   3.3 Variables and measurement
   3.4 Analysis: fixed-effects regression
   3.5 Ethics and data permissions
   3.6 Limitations of the design

4. Results
   4.1 Descriptive overview
   4.2 RQ1 findings
   4.3 RQ2 findings
   4.4 Robustness checks

5. Discussion
   5.1 Interpreting the main association
   5.2 Comparison with urban studies
   5.3 Implications for practice
   5.4 Limitations

6. Conclusion
   6.1 Answers to the research questions
   6.2 Contribution
   6.3 Future research

Notice that every subsection has a job, the research questions are numbered and reappear in the results, and the gap at 2.4 is stated in the outline itself. If you cannot write a heading like 2.4 yet, you are not ready to start the literature review in earnest.

Formatting is set by your institution's graduate school or department, not by a universal standard, so download the official thesis template or formatting guide before you type a word. Many universities supply a Word or LaTeX thesis template; using it from the start saves days of reformatting at the end.

If you have no template, these settings are common defaults and are safe until you learn otherwise:

Practical typing tips

• Use styles, not manual formatting. Apply heading styles so the table of contents generates automatically.
• Use a reference manager from the first source. Retyping references at the end is slow and error-prone.
• Insert captions with the caption tool so lists of tables and figures update themselves.
• Keep one master file or a clear folder of chapter files with version dates in the file names, and back them up in two places.

Unsure which citation style applies? See choosing a citation style.

Timelines vary widely by degree and discipline. The plan below assumes a nine-month master's thesis with empirical data collection; stretch each phase proportionally for a longer project or a doctorate. Plan backward from your submission date, and treat the revision and buffer phases as fixed.

Making the timeline hold

• Put supervisor meetings on the calendar now, with a draft due a week before each one.
• Allow for waiting. Ethics review, participant recruitment and supervisor feedback all take longer than you expect.
• Budget about a quarter of the total time for revision. A first draft is rarely close to a submittable one.
• Track progress weekly against the deliverables column, not against hours spent.

Introduction

Move from the broad problem to your specific question in a few pages. End with explicit research questions or aims and a short roadmap of the chapters. A strong thesis introduction answers "why does this matter, and why now?" without promising more than the study delivers. See how to write a research paper introduction for paragraph-level patterns.

Literature review

Organize by theme or debate, not by author. Each subsection should end by saying what the studies together show and what they leave unresolved. The final subsection states your gap precisely.

Methodology

Give enough detail that another researcher could repeat the study, and justify each choice. "We used a survey" is description; "a survey allowed us to reach dispersed rural patrons at low cost, at the expense of depth" is justification.

Results

Report findings in the order of your research questions. Let tables carry the numbers and use the text to point out what matters. Save interpretation for the discussion.

Discussion and conclusion

Explain what the findings mean, compare them with the studies in your literature review, and say honestly where your design limits the claims you can make. The conclusion answers each research question in plain words and states your contribution in one or two sentences.

Abstract

Write it last, in 150–300 words: problem, method, main findings, implication. Draft it from your one-paragraph research summary, updated with what you actually found.

If you searched for how to write a thesis because you need a thesis statement, this section is for you. A thesis statement is one or two sentences, usually at the end of the introduction, that state the paper's main claim and, often, the reasons that support it. A strong thesis statement is specific, arguable and limited to what the paper can actually show.

A simple formula

[Specific claim] because [reason 1], [reason 2] and [reason 3]. Use the reasons as the backbone of your outline: one body section per reason.

Argumentative thesis

Take a position someone could reasonably dispute. Weak: "Homework has pros and cons." Stronger: "Middle schools should cap nightly homework at one hour, because longer assignments crowd out sleep, show diminishing learning returns and widen gaps between students with and without help at home."

Rhetorical analysis thesis template

Name the author, the text, the strategies and their effect: "In [text], [author] uses [strategy 1], [strategy 2] and [strategy 3] to [purpose] for [audience]." For example: "In her commencement speech, the author uses personal anecdote, direct address and repetition to persuade graduating nurses that small acts of attention matter more than credentials."

Compare-and-contrast thesis

Go beyond "A and B are similar and different." State what the comparison reveals: "Although both cities adopted congestion pricing, the one that reinvested revenue in buses saw stronger public support, suggesting that visible use of funds matters more than the price itself."

Research paper thesis

In a research paper, the thesis often takes the form of the answer your evidence supports, or a clear statement of the question and approach. For more patterns, see thesis statement examples.

• A topic, not a question. Chapters drift because nothing defines what counts as relevant.
• A literature review that summarizes instead of synthesizes. One paragraph per study, with no comparison, signals that you have not yet found your gap.
• Misalignment. The research question asks "why," but the method can only describe "what." Check that each question has a matching analysis and a matching results section.
• Overclaiming. Correlational data reported as causal, or a small sample treated as representative. Use cautious verbs: "suggests," "is associated with," "is consistent with."
• Results without interpretation, or interpretation without results. Keep the jobs of chapters 4 and 5 distinct, then connect them.
• Missing or token limitations. A specific, honest limitations section increases an examiner's confidence in the rest of the work.
• Inconsistent terminology. Pick one term for each key concept and use it throughout.
• Reference errors and formatting slips. Mismatched in-text citations and reference list entries suggest carelessness elsewhere.

A final read-through with fresh eyes catches many of these. If you want a second reader, we coach, edit and explain; you stay the author of every word and decision.