How to Write a Thesis
For B.A. and M.A. Students (and maybe Ph.D. students, too)
by

Charles Lipson
University of Chicago

This draws on the advice in my book:
How to Write a B.A. Thesis: A Practical Guide from Your First Ideas to Your Finished Paper
(Chicago: University of Chicago Press, Spring 2005)

A quick overview of the main points:

Now, let's get down to brass tacks.

Researching and writing a thesis is a major project that takes several months. Doing it well requires self-discipline, but it really helps to love the subject matter. To sustain your interest and energy over the long haul, pick a topic that interests you deeply, one you are excited about. The best way to begin is to search for an interesting question. At this stage, you should hunt for a good question, not the answer to it. Finding an answer is the purpose of your subsequent research and writing.

A successful thesis poses an interesting question you can actually answer. Having found a topic that really interests you, you need to locate an empirical or theoretical puzzle that needs more exploration. Working with your advisor, narrow your topic to a manageable size.

A topic is manageable if you can

• master the relevant literature;
• collect and analyze the necessary data; and
• answer the key questions you have posed.

Some problems are simply too big and unwieldy to master within the time limits. Some are too small to interest you. This is a Goldilocks problem, and the solution is to select a well-defined topic that bears on some larger issue. You can begin either with a large issue or a well-defined topic, depending on your own interests.

From "Big Issue" to manageable thesis topic: You might start with a grand-scale issue, such as "Why has the U.S. fought so many wars since 1945?"  Working with your advisor, you could then zero in on a related but feasible research topic, such as "Why did the Johnson Administration choose to escalate in Vietnam?" As you answer this Vietnam question, however, you can (and should) return to the larger themes than interest you, namely what does the Vietnam escalation tell us about the global projection of U.S. military power since 1945.

From well-defined topic to the "Big Issue": Perhaps you are already interested in a well-defined and manageable topic such as the decision to create NAFTA. If so, then your task is to clarify which larger issues your paper will bear upon. The problem (and the opportunity for you) is that NAFTA bears on several larger topics. You need to pick one that captures your interest. Are you mainly interested in US decisionmaking? Or perhaps Mexican or Canadian decisionmaking? Multilateral negotiations between big and small countries? The role of public opinion? The role of business lobbies or trade unions? The NAFTA decision is related to all these larger issues and more. You cannot tackle all of them so you must choose your focus. Your choice will shape the kind of research you do on NAFTA, leading you to study the lobbying process, for instance, or US-Mexican negotiations. Either would be an interesting thesis about NAFTA, but they are different theses.

You can start with the big issue or the narrow topic. Either approach is fine. A good thesis will connect the two: the well-defined topic and the larger issue.

How can I find the "right" thesis topic? There is no magic answer, but here is a technique I find helpful. I often ask students to propose 3 topics—briefly, in writing, and in order of priority. Why three topics? Because, believe it or not, it is sometimes easier to jot down three ideas than to pick just one. When you try to generate the "single best idea," there is a lot of pressure to pick exactly the right one. After all, you will have to work on it for months. By contrast, writing down 2, 3, or 4 ideas lessens the pressure since you are not committed to any one of them. Equally important, you and your advisor can talk about your multiple ideas. Your advisor will learn about your major interests and, as you talk, together you may discover that seemingly different ideas have a common theme. Once you have identified this common theme, you and your advisor can then find a researchable topic, which may be slightly different from any of the 3 written ideas you presented.

Your main argument should be brief and crisp. No matter how complicated and subtle your overall paper, your basic thesis should be expressed in clear, pointed language. This requires some serious thinking to boil down your views and some intellectual bravery to state them clearly, without weasel words. If possible, your argument should be clearly differentiated from others. These alternative arguments should be identified with specific scholars. The emphasis, however, should be on developing your own position and evaluating it honestly and rigorously.

It takes weeks, sometimes months, to develop a compelling central argument. That can be frustrating. But remember, if you knew exactly what you were going to say before you started work, the whole project would be boring—to you and probably to your readers. Most of us begin with some general ideas and puzzling problems and gradually work our way toward a sharper definition of the topic, the argument, and the best ways to test it. Your advisor is there to help at each stage along the way.

To put this another way, you should pose a focused question and offer a coherent answer.

Take the elevator test: How do you know when you have finally developed a clear-cut argument? Take the elevator test. You should be able to tell a visiting professor your basic argument and the rationale for it as you ride from the lobby to the 4th floor. If you can do that, then you have a clear argument.

Learn what other scholars have said about your topic. You need to explain

• What are the major approaches to your question?
• Why are existing answers unsatisfactory? and
• Why is your answer better?

Present these alternatives seriously, thoughtfully, not as "straw men." Grapple with them intellectually. Most important of all, as your thesis unfolds, show that your answer is compelling and better than the alternatives.

Your answer probably relies on some major theory and applies it to your particular question. If so, then show that this theory actually applies well to your topic and leads you to a better answer than the alternatives, not only in the abstract but in this particular case.

The introductory section of the paper should do three things. It should

• entice the reader into the subject matter, probably with an interesting opening paragraph, perhaps with a compelling anecdote, concrete example, or real-life puzzle;
• explain the topic you are studying, the basic material you will cover, and your central argument or testable proposition; and, finally, at the end of the introductory section,
• orient your readers by giving them a "road map" for the overall paper, explaining briefly what each section does.

As the paper unfolds, you should introduce each new section briefly, saying why it is important to your overall argument. Most sections should conclude with a few summary remarks and a transition to the next section. Occasionally, it makes more sense to put the transition at the beginning of the new section. Wherever you put the transitional sentences, they should take the reader smoothly to the next topic. That means you should tell the reader why you are tackling the upcoming topic, how it matters to your overall argument, and why it logically comes next in your paper.

Some papers do not use case studies. They may simply present a logical model, usually in mathematical form. Or they may test their propositions by using large data sets (sometimes called large-n samples). But many papers use individual cases to show how the explanation works and to evaluate it in detail.

The cases chosen need not be typical. They can be striking or unusual. But they must illuminate the general problem under investigation. The reader needs to be told—in advance and in plain language—why you are using these particular cases.

The best cases to use are often the hardest ones. That is, they are cases where your own argument seems least likely to apply but, in your judgment, still does. These hard cases will be most convincing to readers because they show the power of your argument and its generality.

If, for example, you wish to show that bureaucrats have extensive power over policy outcomes, a "hard case" would be one where high-level elected officials really cared about the issue. (If politicians didn't care, then of course bureaucrats would control the outcome. What does that prove? Not much unless you could show that there were many such issues, all under the thumb of bureaucrats.) The hard case is much more interesting. The toughest case would be one where bureaucrats and politicians wanted different outcomes and where politicians cared deeply about the issue. If you could demonstrate that in such "hard cases" bureaucrats still profoundly affected the outcomes, then you would have strong supporting evidence for your general proposition. That is why you need to select cases carefully and explain how they help test your overall argument.

As you work out your argument, you may decide to formulate and test some generalizations. That is surely one of the major goals of social science, and it is a rewarding exercise in a thesis. If you intend to test some generalizations, it is crucial to think about what kinds of evidence bears on them. You should be particularly attentive to what kinds of evidence could actually refute them. If any kind of evidence is consistent with your argument, then you don't really have an argument at all. Back to the drawing table.

Beyond this essential "pass-fail" test, a thoughtful evaluation should ask,

a. "What conditions affect the impact of a particular generalization?"
b. "What are the limits of this generalization?" (Does my data allow me to say if the generalization applies broadly?)

As an example, take this important generalization in the field of international relations: "Democracies do not fight wars against other democracies." Your evaluation may conclude that this generalization fails (or succeeds) entirely. Or you may find that it applies frequently, but not always, and only under significant limiting conditions. What are these limiting conditions? Maybe you find it applies only to rich democracies, or well-established democracies, or Presidential systems (as opposed to Parliamentary ones). Such findings allow you to circumscribe the generalization, or at least propose some limits to it. You should also be aware of the limits imposed by the data you use. You may tentatively confirm the generalization about democracies and war. But if your evidence is drawn exclusively from the period after 1945 (or from a particular region), you may wish to add that we cannot be sure if the generalization applies to other time periods or other regions without further testing. Drawing such limits requires hard thinking about your topic and your data. That is precisely what is intellectually rewarding about doing a major project. Of course, these are major issues to discuss with your advisor.

• Write in the active voice.
• Use plain language.
• When in doubt, break long sentences into shorter ones, as long as they are not choppy.
• Write brief, coherent paragraphs, each with a single topic sentence.
• Rewrite any sentences that string together prepositions.
• Check to see if you are repeating yourself or using the same words too often.
• Use direct quotations sparingly and name the person being quoted.
• Double-check the paper's opening paragraphs. They should engage the reader.
• Introduce your key questions and central arguments early & clearly. Don't bury them.
• Edit, edit, and edit some more.

Although this may be elementary advice, it is still important to remember, and it is all too often ignored, especially in writing early drafts.

Well-organized paragraphs are the main building blocks of your paper. Through them, you develop your question, your answer, and your evidence in a well-ordered, sequential way. Each paragraph should be relatively short and focused, with a clear topic sentence that articulates the main point. Double check any paragraphs that run more than five or six sentences to see if you are cramming in too much.

Editing (and re-editing) your early drafts is the key to making your thesis sharper, deeper, and more readable. Don't be afraid to cut extraneous material, even if it took you a long time to write. Remember, you are not being paid by the hour. What matters is the quality of the final product. It should be taut, clear, and polished. It is painful to cut your own hard-wrought prose. I know, believe me, I know. But your paper will be much better for it. To lessen the pain, save these cuts in a "scrap file." That gives you the chance to reinsert sentences or paragraphs if you really need them.

Quotations can be a source of writing trouble. Do not overuse them. When you are simply presenting data or well-known opinions, rephrase the quote in your own words and footnote the source.

So, when are quotes really useful? In at least two instances.
• First, use quotes when you want to capture the speaker's striking, memorable phrase. For example:

"When you come to a fork in the road, take it," was Yogi Berra's sage advice.

• Second, a few, well-chosen quotes can illustrate the viewpoint of a scholar, policymaker, or participant. For example:

In President John Kennedy's ambitious phrase, "We will pay any price, bear any burden."  Or

Senator Joseph McCarthy did more than call his opponents misguided, he damned them as "communists" or "fellow travelers."

• In a few cases, you may wish to use longer quotations, running several sentences. They should be used very sparingly and indented in your text. Introduce these longer quotes with your own summarizing sentence so they make sense even if the reader skips over them, which sometimes happens. For example, to introduce a longer quote:

Churchill's Iron Curtain speech argued emphatically that the Soviet Union threatened not only the Western security but Western values: followed by indented quote from speech.

• Finally, do not put quotation marks around "ordinary words" unless you specifically wish to call attention to a word's use or misuse, and you make your purpose and viewpoint clear. Do not use quotation marks to be snide or ironic. For your readers, that wears thin very fast. Here is an example of quotation marks used properly:

What is often called "collateral damage" is really the killing and maiming of innocent civilians, cloaked in deadening, bureaucratic language.

The first page of your paper: The poorest writing in a thesis is often on the first page, when you are striving to say something terribly BIG and IMPORTANT. However worthy the goal, the danger is that you will begin with a vague platitude rather than a crisp, compelling introduction to your work. Concentrate on introducing your main question and saying, in a concrete way, why it has larger significance.

One common problem is that these opening paragraphs are written quite late in the game, after you have finished the other writing and polished it. You haven't really had time to re-read and edit the first page closely, as you have the rest of the paper. It is perfectly fine to write these paragraphs last, but be sure to edit them carefully. The goal is to raise your main question and get to the heart of your argument quickly, certainly in the first couple of pages. Too much introduction can bury the main point of your paper. One useful technique: see if you can simply chop off the first few paragraphs of your draft paper.

Concluding section of paper: Your paper should have a concluding section, usually a succinct one. It should summarize your findings, not retrace everything you have done. Remember, it is a concluding section, not a summary section. The main thrust should be the interpretation of your findings. Hit the high points, and then say what they mean. What are your chief findings? Why are they significant (that is, how do they matter for policy, theory, moral action, or whatever)? What are the limitations of your findings? Now is the time to reintroduce the larger questions that animate you and say how your findings bear on them. Make it a high priority to discuss these conclusions with your advisor. In my experience, the main danger here is that you finally get to this concluding section with only a week or two left before the due date. The solution: begin discussing your conclusions with your advisor when you are still writing the heart of the paper, when your conclusions are still tentative.

Read aloud as you edit: These good things will not all happen on your first draft, or even the second. You need to re-read and edit, time and again. One of the best ways to do that and to improve your writing is to read it aloud to yourself. If you are a practiced reader, you will have a good ear and will be able to hear when your own prose does not sound quite right.

Guides to good writing: Are there any good books on writing nonfiction like this? You bet. The two most helpful are William Zinsser's On Writing Well and John Trimble's Writing with Style. Unlike most books on writing, they are not only readable, they are pleasurable. Buy either one, or, better yet, buy both. Zinsser's book is not specifically about academic writing; it is about writing good nonfiction. It is enjoyable, wise, and filled with practical advice. So, too, is Trimble's slim book, which includes examples from his undergraduate writing classes. Your thesis will be better if you read Trimble or Zinsser before writing and editing. You may also wish to review Strunk and White's "little book," Elements of Style. It is a classic for a reason.

It is up to you to propose the schedule. Take my word for it, your faculty advisor won't do it. Leave plenty of time for faculty to read your drafts and then for you to revise them. After spending months on research and writing, you will need time to polish the results. Nothing will improve your work more than successive drafts.

Cite your sources. If you use the exact words of another author, put them in quotation marks and cite them, too.

Citation can be done in several ways. I explain them and show how to avoid plagiarism in a short book:
Charles Lipson, Doing Honest Work in College: How to Prepare Citations, Avoid Plagiarism, and Achieve Real Academic Success (Chicago: University of Chicago Press, 2004), Part II has chapters on each citation style, with lots of examples. It should cover everything you need in your thesis. Description at University of Chicago Press Site  |  Amazon  |  Barnes & Noble

Give your advisor several items in writing as early as possible:

a. main background readings on your general topic; your advisor can be very helpful with this bibliography;
b. your relevant academic preparation; it may help your advisor to know if you read French or do advanced statistics, depending on the topic you choose;
c. your general topic of interest, which you will later refine into a
d. one-page summary of your basic topic and argument; if you cannot state the point so concisely, you probably do not have a coherent thesis at all;
e. a statement of plausible alternative explanations, and how they compare with yours; eventually, you may wish to explain why your explanation is better, why your interpretation is more compelling;
f. a list of the most pertinent evidence, including the cases or data you are likely to use; your advisor can be a valuable guide in point you to relevant evidence;
g. the basic outline for your paper, along with the main arguments in each section.

Of course, no one can do all these at the beginning, but do as much as you can and look for feedback. You will probably start with a general topic and some basic readings and then narrow your topic and focus your research and writing. The idea is to mark off a manageable (but still significant) topic so that you can probe it in depth.

My advice: try to do a, b, and c in your first two weeks. That is, in your first sessions with your advisor, discuss your general topic, your academic background, and the initial readings you plan to do. Put these key points in writing quickly. Then, over the next two months, work toward a clearer, more manageable topic, a more focused set of readings, a statement about what data you need to collect, and a statement about the major alternative explanations. Again, put them in writing. Some may be short essays, such as a 3-4 page annotated paper on alternative explanations. After that, you can write an outline for your paper and begin to write individual sections and tables. As you complete these tasks, discuss them with your advisor and set your next task. Frequent short meetings are best. They will give you the most feedback and keep you on deadline.

a. your name,
b. phone,
c. e-mail,
d. today's date, and
e. paper title (even if it is tentative).

Pages should be numbered, and the paper should be stapled, not paper-clipped. With this information on the paper itself, your advisor can speak with you about specific issues on specific pages and then keep a copy of your latest version.

If your written work is only a page or two, then your advisor can read it at your meeting and discuss it with you right there. If your written work is longer, turn in a copy before the meeting (but always bring two copies to the meeting itself). Don't bother putting your paper in a fancy binder. Nobody cares. Ask your advisor if you should provide this advance text as a hard copy, an e-mail attachment, or both.

For heavens sake, proofread everything you turn in to your advisor. Nothing says "I can't be bothered about this project" like a few missspelllings. Obviously, you will run spellcheck. Do it every time before you turn in a draft to your advisor. In addition, you must re-read the paper carefully, looking for errors the computer missed, such as using "there" instead of "their" or inadvertently leaving out a word because of editing. We all make these mistakes. That's why you have to proofread each time. If you want your advisor to read your work with care, then you must do the same.

Tables, graphs, and figures are often the clearest way to present your data. A simple table may also be the best way to lay out your argument and compare it to others. Think about these presentational issues and talk with your advisor about them. If your paper has tables or figures, make sure that there is not a page break in the middle of any table. You will need to recheck this with each new version.

Save a backup copy of your research and writing on your school computer or somewhere else. Safety first.

Your advisor will undoubtedly have specific issues to raise, but so should you. Think about

a. which issues you need help on, and
b. which topics you need additional readings for.

Ask your advisor whether there are additional perspectives you may have overlooked or need to explore further.

Also, remember to go over your central thesis with your advisor in several meetings, as you develop that thesis. You may begin with two or three vague arguments, but you will hone them down as you do research, writing, and discussion.

The clearer you make your own agenda in meetings with your advisor, the more productive those meetings will be. Learning how to make such meetings fruitful is an important part of the thesis project and a useful step toward managing large projects on your own.

Don't leave without setting a tentative date for your next session with your advisor.
 

Some suggestions here are most useful when you begin thinking about the thesis project, others when you start writing, and still others when you are polishing your final draft. It really does help to reread this advice as your thesis project develops.

I'd love to include your own advice and tips for thesis writers. Please send them to c-lipson@uchicago.edu
Thanks to
Prof. Florence Riffe of Ohio University for suggesting the most important item, #1: "Pick a topic you really care about"
Prof. Robert Pape for five questions every thesis writer should ask

For my advice on getting a good recommendation for graduate school or a job, click here.