A senior thesis is a serious undertaking—a major research project that will require countless hours of research, writing, and revising.

By the end of the three-term process, you should expect to produce a significant piece of social science research of which you can be proud. This handout is designed to give you some insights into each phase of this process. These suggestions are not hard and fast rules, nor do they cover the official departmental requirements for the senior thesis. See the department’s official rules and requirements.

This information offers a framework and advice for those who have chosen to undertake an empirical senior thesis project. Those writing a thesis on political theory may find some parts of this handout useful (selecting an advisor and second reader, for example) and other parts less applicable (generating testable hypotheses).

For other questions on the thesis, and for advice on how you might adapt some of the suggestions here for your own project, consult your advisor.

Choosing an Advisor and Second Reader

Your first task is to secure a thesis advisor. Your advisor will meet with you regularly, give you feedback on written work, and answer crucial questions and concerns you may have about your research. The advisor’s role is different from that of a professor in the classroom. Whereas in the classroom, the professor is presumed to be the expert on both the subject matter and the political science that treats the subject matter, in a thesis setting you take on the role of expert in your particular subject. This is because your goal in writing a thesis is to gather original evidence and make an original argument. Accordingly, do not suppose that all is lost if you are not able to find a thesis advisor who is an expert in your particular area of research. If you are writing a thesis on India, for example, do not despair if you cannot find an India scholar in the political science department. 



An advisor can and should offer expertise on the conduct of political science research. All members of the political science department have been through what you are about to go through: they have designed an original research project and shepherded it through to its conclusion. One strategy for selecting an advisor, therefore, is to find someone who has done research in the manner in which you plan to do research. If you will be working with statistics, find someone with knowledge of statistics. If you will be conducting detailed case studies, find someone who has done case studies, and so on.



A second reader’s role is to provide a third-party “set of eyes” on your work. Because the second reader is usually not as closely connected with the project as the thesis advisor, his or her feedback can be very valuable as a check on whether your research makes sense to someone who is not intimately familiar with your thought process. In selecting a second reader, consider someone whose skill set complements, rather than duplicates, that of the thesis advisor.



Finally, do not assume that the thesis advisor and second reader are your only means of getting good advice about your project. This is a college full of great resources and experts who may be useful to you. Take advantage of these assets! Consult other professors in the political science department and other departments; make use of the writing center in the Center for Teaching, Learning, and Research; discuss your project with your peers and solicit their feedback; and make use of the library and all it has to offer (contact social science librarian Brenda Ellis with questions on library materials).

Components of a Good Thesis Project

Good senior theses are clear, well-organized, and contain original thinking and new evidence. The goal is to make a contribution to scholarship in the field by offering something fresh.

Take a look at some articles in major peer-reviewed political science journals (American Political Science ReviewAmerican Journal of Political ScienceWorld PoliticsJournal of Conflict Resolution, etc.), or at a major political science book or two that you have read in class. Note how these works are organized. Each work typically contains a beginning literature review section, a middle section in which it outlines the research design, a third section in which it describes the empirical analysis, and a conclusion that summarizes the broader implications of the work. Try to emulate this pattern. In fact, if you are concerned about how much space to expend on literature review vs. data analysis (for example), these articles and books can serve as a good model.



Embedded in this organizational pattern are four key elements of an empirical research project.

Introduction: The Research Question

What is the unexplored area that you are trying to shed light on?  What has the other research missed? What “puzzle” are you trying to solve? Tell the reader the research question as early as possible in the introduction.

This is not as easy a task as it sounds. Many of us begin our research projects because we are interested in a particular subject: nuclear proliferation, campaign contributions, non-governmental organizations in Africa, elections in France, etc. But a subject is not enough. Each subject can be associated with innumerable research questions. Suppose, for instance, that I’m interested in the 2012 election in France. Four questions that could generate productive research projects are: How has gender’s impact on French elections changed in the aftermath of the Dominique Strauss-Kahn scandal? How deep does Euroskepticism run and what is its impact on voting behavior? How do ethnic and religious animosities affect French election results? How has online social networking changed campaign styles in Europe?

Each one of these research questions could lead to an entirely different project. When you settle on precisely the question you want to ask, you have gone a long way toward narrowing and focusing your research. Even a thesis-length research project should have only one research question. 

Do not keep the question a secret from the reader – make it as clear and concise as possible very early in the thesis. One tactic is to announce it in the first sentence or two. For example: “In a political system where nearly every adult may vote, but where knowledge, wealth, social position, access to officials, and other resources are unequally distributed, who actually governs?”[1] Or: “The distribution of wealth is one of today’s most widely discussed and controversial issues. But what do we know about its evolution over the long term?”[2]

Literature Review: The Theories

Once you have posed the question, the next task is to propose possible answers to the question. This is where theories come in. A theory is “a reasoned and precise speculation about the answer to a research question, including a statement about why the proposed answer is correct.”[3] Whereas each research project should have one research question, you should examine multiple theories. There are several reasons for this rule. First, if you ignore other possible answers to your research question, you open yourself up to the criticism that you have unfairly biased your research in favor of your own preferred explanation. Second, if there is only one conceivable answer to your research question, it is a question with an obvious answer and therefore is uninteresting.

Science is a cumulative endeavor. Rarely, if ever, will a social science research project treat a question that has never been addressed before in any form. Previous research may have examined particular aspects of your research question, tested alternative theories, and come to divergent conclusions. In a literature review section, you should critically examine previous research, and draw on this research in developing the theories that you will test in your own work.

Note that the literature review’s purpose is to set the stage for your own project – it is not intended to be a summary of all that you know about your research subject. Instead, the literature review ought to be a focused argument that 1) explains why previous literature is insufficient to answer the research question, and 2) highlights what you consider to be the major potential theoretical answers to the question. As part of the literature review section, you may also propose an alternative theories or theories that you will test and that the previous literature has missed.

A common problem in doing background research is to be too narrow in the initial round of searches.  For example, suppose someone is interested in doing research about Argentina’s decision not to pursue a nuclear weapons program.  This person might begin by searching for articles that other scholars have written about Argentina’s lack of a nuclear weapons program, find only a few articles, and decide that little if anything has been written on this issue.  This would be a mistake.  Instead, the researcher ought to think of the broader category or categories into which this research falls: Nuclear proliferation, weapons of mass destruction, militarism in Latin America, etc. Existing research on these general topics ought to be relevant and plentiful.

After casting a wide net for relevant research, your task is to efficiently discuss it with reference to the key theories that you will be testing. Do not use the “boxcar method” of discussing each relevant article or book in turn. This would take forever and be tedious to read. Instead, group studies together into categories and give details about them only where they are relevant. Here is an example from the literature review section of a recent American Political Science Review article on the initiation of conflicts by authoritarian regimes:

To date, only a handful of studies have explored variation in the conflict behavior of autocracies. A series of early studies by Mark Peceny and colleagues (Peceny and Beer 2003; Peceny, Beer, and Sanchez-Terry 2002; Peceny and Butler 2004) concluded that personalist dictatorships, in which the leader depends on only a small coterie of supporters, are more likely to initiate conflicts than both democracies and other authoritarian regime types. Peceny and Butler (2004) attribute this pattern to Bueno de Mesquita et al.’s (2003) selectorate theory, arguing that personalist regimes are more likely to initiate conflicts than single-party regimes because of their small coalition size.[4]

Note that the author makes an argument that previous research has fallen short, she groups many similar studies together, and she highlights the theory tested by these studies (“selectorate theory”) rather than other less important details. Take a look at other literature review sections of articles and books for other examples.

By the end of your literature review section the reader should know why the previous literature has fallen short of answering your research question and should be aware of the theories from past research that you consider most important. You may develop an original theory as part of the literature review section, or you may describe it in a separate section immediately after the literature review. In any case, by this point in the thesis it should be clear to the reader which theories you will be testing.

Research Design: The Hypotheses

Each theory should generate at least one hypothesis.  A hypothesis is a theory expressed in a way that can be empirically measured.  For example, one theory might be that people don’t vote because they’re cynical about politicians.  One hypothesis based on this theory is as follows: if I examine survey data, those people who express low levels of trust in public officials are going to vote at lower rates than those who express higher levels of trust.  This is not the only possible hypothesis that can test this theory — on the contrary, most theories can be associated with many hypotheses. You should be prepared to test as many hypotheses as you believe will allow you to make a convincing case about which theories are supported and which theories are not supported by the evidence. This may mean you test one hypothesis per theory; this may mean you test several per theory.

In the research design section, you should provide an explanation of how you will gather the evidence needed to test your hypothesis, an account of how you will analyze this evidence, and a description of what type of evidence you will accept as confirming or disconfirming each hypothesis. Using the cynicism and voting example above, if I find that expressions of trust in public officials seems unrelated to voting rates, then my hypothesis is not confirmed.

The particular method you use to analyze the evidence ought to be dictated by the types of evidence that are available to test each hypothesis. If the most viable hypotheses concern evidence that is best observed through case studies (e.g. marginal elite movements play a central role in social revolutions[5]), then case studies are appropriate. If the hypothesis seems well suited for statistical analysis (high mortality rates among Europeans led to the creation of extractive colonial institutions[6]), then statistical analysis may be appropriate. The theories and hypotheses should precede the choice of method of analysis. 

Empirical Evidence: The Data

The centerpiece of your thesis should be your analysis of the evidence, or “data,” broadly construed. There are many different ways to present information effectively. Again, look to political science articles and books – as well as past Middlebury political science theses – for good ideas. The analysis chapters should appear in a logical order and should be clearly related to the theories and hypotheses you discussed earlier. Whatever your approach, the reader should always have a good idea of why you are doing what you are doing, how you are doing it, and how each part of the thesis relates to the other parts. Do not be afraid of some repetition – it can be a good idea to remind the reader of the overall outline of the project in the introduction of each chapter, for example. Headings and subheadings also reinforce a clear organizational scheme.

By the end of the analysis section, the reader should know how closely the evidence you have gathered fits your hypotheses, and by implication, which theory (i.e. which answer to your research question) is most likely to be correct.  If one theory is supported and other theories are not, discuss the implications of this finding.  If several theories are supported, it is probably a good idea to propose a research design that could potentially determine which theory is actually correct.  Or, if you think that several theories are correct to a certain extent, propose how they could be combined into a single broader theory.  For example, two possible theories of non-voting are 1) non-voters are cynical, and 2) non-voters are people who don’t have the time to figure out how to register.  One way to combine these two theories might be to say:  For those 25 years of age and under, non-voting is probably the result of not knowing how to register, since young people are more mobile and can’t figure out a new system of voter registration every time they move.  Older people, on the other hand, are more settled, and therefore their non-voting is probably less the result of registration difficulties and more the result of cynicism. (Note: the above theory is not necessarily true – it is just an example of how you can combine two theories into a coherent whole.)

Do not be upset if the data fail to show exactly what you thought they would show. Remember, the reason that you asked your research question in the first place was that the question did not have an obvious answer. It should be no surprise, therefore, that the results were not predictable in advance. As a scholar, you are not judged on how well you anticipated the precise result, but on whether you considered the most relevant theories, whether you tested hypotheses with appropriate evidence, and whether your interpretation of this evidence is convincing.

Conclusion

In a concluding section, give the reader a concise summary of all you have done. Explain the implications of your research both for your topic area and perhaps for related similar topic areas. This is your last chance to drive home the importance of your work. The points you make here about the broader implications of your research can form the basis for future research, and – more practically – may generate significant discussion during your oral defense of your thesis.  

This guide necessarily contains only the broadest overview of how to write a successful empirical senior thesis. You will have other questions and concerns. Consult your advisor, thesis coordinator Amy Yuen, other members of the political science department, and one or more of the works listed below.

Recommended Reading

Johnson, Janet Buttolph & H.T. Reynolds. Political Science Research Methods (7th Ed.). Washington, DC:CQ Press, 2012.

Lipson, Charles. How to Write a B.A. Thesis: A Practical Guide from Your First Ideas to Your Finished Paper Chicago: University of Chicago Press, 2005.

King, Gary, Robert Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research Princeton, NJ: Princeton University Press, 1994.

Van Evera, Stephen. Guide to Methods for Students of Political Science Ithaca, NY: Cornell University Press, 1997.


[1] Robert A. Dahl, Who Governs? Democracy and Power in an American City (New Haven: Yale University Press, 1961), p. 1.

[2] Thomas Piketty, Capital in the Twenty-First Century (Cambridge, MA: Harvard University Press, 2014), p. 1.

[3] Gary King, Robert O. Keohane, and Sidney Verba, Designing Social Inquiry: Scientific Inference in Qualitative Research (Princeton, NJ: Princeton University Press, 1994), p. 19.

[4] Jessica L. Weeks, “Strongmen and Straw Men: Authoritarian Regimes and the Initiation of International Conflict,” American Political Science Review 106:2 (May 2012), pp. 326-347.

[5] Theda Skocpol, States and Social Revolutions: A Comparative Analysis of France, Russia, and China (New York: Cambridge University Press, 1979).

[6] Daron Acemoglu, Simon Johnson, and James T. Robinson, “The Colonial Origins of Comparative Development: An Empirical Investigation,” The American Economic Review 91:5 (December 2001), pp. 1369-1401.