Five Steps to Writing Your Senior Thesis

Professor Marc L. Busch

1. What is my puzzle?

What is it about your topic that makes it theoretically and empirically interesting?

Typically, puzzles resonate with readers because, looking at the evidence, a trend or

development doesn’t match up with theoretical expectations. When groups or states act

in concert, for example, we’re surprised in light of what we know to be the obstacles to

collective action. The puzzle should be grounded in “real world” evidence and, in turn,

interpreted against a theoretical backdrop. This is what makes for an interesting set-up.

When your advisor encourages you to “motivate” your thesis, this is what they are talking

about.

2. What does the literature say about my puzzle?

This is the literature review section, but it is focused on your puzzle. The point is to

discuss what the literature has said about your puzzle, and to critically assess these

contributions. Sometimes existing theories don’t speak directly to your puzzle per se. In

this case, you have to refine them to help them get at your puzzle. You might derive

more specific hypotheses (“if, then” conjectures) so that they shed some light on your

puzzle. This is a very useful exercise, since you are likely to refine your own thoughts
about your own argument in the process. And pushing existing theories a little further

helps advance the literature, and helps bring out the value-added of your argument. What

you are preventing against is the charge that you’ve caricatured the literature and

underestimated its grasp of your puzzle. You conclude this section by discussing the

gaps in the literature, and hinting at what you need to do, through your argument, to fill in

these gaps.

3. My explanation for the puzzle

This is the “theory section” of the paper. Here, you reveal the assumptions that you’ll

work with and specify your independent and dependent variables. The assumptions

might include invoking a rational unitary state, or taking perfect information as given.

Your independent and dependent variables require a great deal of explicit attention. Most

basically, you need to make sure to clarify exactly what your variables are. Explain what

you mean by offering definitions of your variables and articulating how they vary. And,

of course, you need to explain how variation in your independent variables leads to

variation in your dependent variable. One of the challenges we all encounter in this

endeavor is simply delimiting our variables. The temptation is to try and explain

everything, and yet this is a barrier to building “good theory.” For example, while filings

with the WTO may well lead you to be interested in the verdicts handed down by the

WTO, you want to make sure that you do not conflate these two different dependent

variables. Indeed, it may well be the case that different theories explain these two

different dependent variables.

how your going to test your argument by offering measures of your variables, so that the

reader can see that variation in your independent variables does, in fact, lead to variation

in your dependent variable. This is more tricky than it sounds. Sometimes we have

certain independent variables in mind that, while theoretically interesting, are empirically

hard to work with. You need to spend some time thinking about how to make “testable”

your argument. How do I measure “social capital,” for example? How do I know a

democracy when I see one? What evidence would persuade me that pressure group

politics really mattered in explaining a trade policy outcome, as opposed to the whims of

policymakers and technocrats? In brief, you cannot conclude the theory section without

showing how the argument can be tested empirically. You won’t always find the best

empirical proxies for your variables, but a “good faith” effort is always better than no

effort at all.

4. Methodology

Generally, students debate whether to offer case studies of quantitative analysis. The

decision should be based on the puzzle at hand. Some topics lend themselves to

statistical techniques, given a large number of cases and clear empirical measures of the

relevant variables, for example. Even if you decide that you will do a qualitative study, it

is a good idea to think about how you would do your same project statistically, since the

same concerns (such as degrees of freedom, concern for outliers and biased results) loom

large regardless. If you do a quantitative study, be sure to explain the techniques (i.e., the

need for a probit model, etc.) and think about regression diagnostics. Most software
packages make it easy to do diagnostics, and you should always include diagnostics in

your work (i.e., report them fully). Moreover, case study research can compliment

quantitative analysis very nicely, especially if you can qualitatively probe an “outlier”

identified by your statistical work.

If you proceed with case studies, you need to explain why you chose the cases you did,

and why you think you can generalize from these cases, compare these cases, etc. In a

word, you are subject to the same criticisms as your colleagues doing quantitative work,

and so you need to be sensitive to the same concerns. Most importantly, you need to

write the cases with your variables in mind. It is useful, for example, to organize your

cases under the subheadings “independent” and “dependent” variables. This will ensure

that you don’t forget about your theory when you turn to the empirical material. This is a

common problem across the board, and not just among students. Especially in a research

paper, you should ask yourself, while writing each paragraph, whether you showed the

reader anything about your argument in writing what you did, or whether you simply

recounted an interesting episode, and expected the reader to figure out how it relates to

your argument. Case study work can be extremely useful if you walk the reader through

the cases with your argument in mind.

5. Conclusion

In the conclusion, you want to reflect on the introduction, recall why the puzzle was

interesting, and recount how your argument shed new light on the puzzle. This isn’t quite

the same thing as cutting and pasting your introduction, since there should be some
honest reflection here. Moreover, you might want to assess the limitations of your

argument, and think about the implications of your work for future research. For

example, if you invoked a rational unitary state in making your argument, you might

want to think about how you could enrich your study later by opening up the state and

bringing the demand side back in. You should also think about the policy implications of

your argument. What are you suggesting, in light of your argument, that policymakers

should do? If you are right, what should observers learn from your work? This is where

the project gets wrapped up, having touched all the bases.