Empirical Application Paper Option

From the course syllabus:

“Empirical applications similarly (1) draw one or a few core arguments from a book and (2) demonstrate

its (their) operation in some empirical case(s) (country-time(s)) beyond those explored by the author. (2) is the key. That is, analyze some other cases with the analytical tools and theories of the author. As with the other type of paper, you must explain how one may evaluate empirically this explanation for events; how would events have evolved differently if the theories you are applying did not hold?”

The goal of this paper is for you to take an existing argument (or more than one) from the texts and apply it somewhere else outside of the author’s dataset. You should draw on substantive expertise developed in other courses to identify cases of interest. You may use cases from a different region of the world and/or a different temporal period; your paper should justify why you believe the theory/ies should operate in this situation, as well as giving an explanation of the theory/ies and a brief summary of the author’s findings from his/their initial tests.

Your argument may be that the author’s argument does explain your case. Identify the key variables of the theory, then present evidence that supports an argument that these variables took on the values the theory predicted. Depending on your choices, particularly if you cannot use the author’s original measurement strategy or believe that it is actually inappropriate, you may need to discuss how you would measure some of the variables (what will you look at as ‘evidence’ for a variable’s value?), and/or defend your measurement strategy. Then, think about the counterfactual. What might have happened if the theory had not held in this case—if your variables had taken on different values? What kinds of factors might have caused the variables to take on different values?

You may also argue that the author’s argument should but fails to explain your case. If you are discussing a failure, are reasons for failure evident, prominent, or obvious? What were the theory’s key variables? What values should these have had, and what values did they have in your case? (Again, you may need to discuss issues of measurement, and what you looked at as evidence about each variable’s value.) What variables mattered instead of the things the theory predicted? Is your failed case really an example of the theory generally working, but one variable omitted from the initial argument overriding the rest of the theory? (e.g., Watergate in 1974)

You may choose to conduct your empirical research in any of a variety of ways. Any type of research design is acceptable, so long as it allows you to create and support a clear and specific argument. I am happy to help you create a research design for the project; this is a complicated process and you should plan to allow at least two weeks to complete the design and execute it. I can also provide information on data sources and availability. If you feel you have sufficient skills, you are more than welcome to proceed without consultation.

One option is to conduct a case study of a specific instance or several specific instances, where you use books, newspapers, memoirs, government documents, magazines, etc., to trace the process and events of your case. You would be looking for evidence that the actors were (or were not) thinking and/or acting in the ways the theory expects, or that the variables the theory claimed as important or influential did (or failed to) operate as expected.

Another option, depending on the argument and the case(s) you choose, might be to do content analysis of various types. You could examine public speeches, newspaper accounts, editorials, etc., looking for references attributable to the theory’s variables. This is a different way of establishing support for your argument: rather than tracing a process of events to suggest that the causal sequence lines up to predict what the theory predicts, you would measure the relative importance of your variables by seeing how frequently they were talked about or addressed by the key actors of your theory.

A third major option, for those of you who’ve had STAT 350 or the equivalent, is to use statistical techniques to sort through evidence for and against your argument. This is a more complex choice, and almost certainly requires more than one theory to test. It does require multiple cases; in particular, you need at least twenty to begin getting statistically reliable results for most analytical methods, and thirty (or more) is really better. Some options, like those employed by Tufte, can be done with fewer cases, though your results will be less certain. If you’d like to do something using this, perhaps applying an argument wholesale to a number of years for a country outside the author’s sample or something of that nature, you should almost certainly consult with an instructor. Prof. Franzese is an excellent resource on these issues, and I am happy to provide what resources or expertise I can.

As always, you are not limited to English-language sources. In particular, you may wish to consult sources in the language of the country/ies under consideration. The usual standards for academic credibility of sources and appropriate citation apply. I am happy to recommend credible sources in French and German for countries using those languages; recommendations for other languages can be provided with a little advance notice.

No matter your research design or tools of analysis, what you should NOT do in this paper is simply provide a journalistic-type account of a case: who, what, when, where. Even adding ‘why’ and ‘how’ is insufficient for the type of essay you are being asked to create. We are looking for a specific argument about the role(s) of key independent variables – explanatory factors – in predicting, creating, or otherwise leading to a given outcome (the dependent variable).

Back to the General Guidelines.

Back to the 343 Home Page.